-
Locked
OPA627 VS Burson V5 VS Burson V6 Classic op-amps perceived quality of musical reproduction
OPA627 VS Burson V5 VS Burson V6 Classic op-amps perceived quality of musical reproduction
Images show OPA627, Burson V5 and Buson V6 Classic op-amps and their relative size and Cyenne Dac 3100 with OPA627 and two V6 Classic op-amps in sockets.
In exchange for two Burson Classic dual op-amps I writing my honest review of their performance in comparison to OPA627 and the Burson V5s I already own. My system is composed of an Asus Gaming Computer running JRiver media center 28 which I modified to up samples all files to 96 KHz which are stored on a Passport external hard drive. I have an external Musiland external soundcard I use with optical out through a glass optical interconnect to a Cyenne 3100 DAC with three replaceable op-amp sockets. The DAC output goes to a Cambridge Audio CXA60 amp. I have Mordaunt-Short Aviano 6 speakers and a Wharfedale subwoofer. The Mordaunt-Short speakers are very neutral and I can hear differences in op-amps in the DAC typically. Lm4562 DAC chips came with the Cyenne Dac originally, but I found that despite their low THD, they sounded clean, but lean in expression. I previously owned a Maverick D2 DAC with 3 OPA627 op-amps and found them more textured and open. So, I am reviewing the OPA627 as a baseline vs the two Burson op-amps to see any difference. My dedicated music room is highly acoustically treated and has a fairly straight equalization curve along the x axis, according to Room EQ Wizard, without any equalization required in the JRiver application to achieve this.
Above example of ceiling room treatment in dedicated music room.
I plan to look at the factors Buson uses to describe their op-amps: transparency, details, color and texture, and dynamics and soundstage when reviewing the op-amps previously noted.
Definitions:
Transparency: the sound is reproduced accurately without adding much extra.
Detail: hearing the small, fine sounds of a recording; the most delicate parts of the original sound.
Color and texture: Tone color or texture of a sound is the quality that allows you to separate two instruments playing the same instrument. It is hearing the different layers and harmonies of music.
Dynamics and soundstage: variation in loudness between notes in the spatial sound image presented to the listener.
The Cyenne Dac has 3 removal socketed op-amps. One serves as the combiner op-amp for the DAC. The other 2 are gain op-amps. The gain sockets will vary from OPA627 to the Burson V5 to V6 Classic. I will listen to a playlist of Flac files with a minimum 44.1 KHz and make note of the above features of transparency, etc. for any sound characteristics noticeable for each op-amp.
The flowing listening set began with OPA627 in all three op-amp sockets followed by the Burson V5 in the gain op-amp sockets and OPA627 in the combiner op-amp socket.
Just Like A woman, live 1966 by Bob Dylan.
This is a beautiful 6:45 version of his studio song which only totals 4.42. With three OPA627 op-amps the song sounded less detailed than I remembered. I choose my music in this evaluation for its familiarity. Reference material I believe has to be used so elements like detail and transparency can be assessed as different from a typical presentation. Dylan’s guitar and voice had a sadly muted color and texture. It is actually a beautiful version of a song in which we hear Dylan during a grueling tour in 1966 play an emotionally different version of his studio album while leaning over his stool and expressing longingness and compassion during the song, not the derision found in his studio album. But the OPA627 did not convince me I was listening to a live performance of a wasted Dylan generously sharing his feelings of tenderness with the audience.
The Burson V5 turned the concert and song live. We hear the echo in the room and his harmonica notes are more present. The detail is enhanced. The whole texture and color of the song in his strained voice emoting his regret, not his signature emotional sneer. I was enveloped in the large soundstage during the recitation.
Visions of Johanna by Dylan, the studio track from Blonde on Blonde
The song sounded less textured as well with OPA627. Details were diminished. The notes were compressed and the soundstage was narrow. The music was colored, but with all the instruments the same color.
The Burson V5 again brought out more realism and detail. The bass is strong and Dylan sounds real, unfortunately the soundstage is medium but with clear textured sound separation with sharp details and accuracy.
Hotel California by The Eagles, MTV Live.
With the OPA627 texture is diminished in the guitar work and soundstage is small. Details seem muted. Henley seems to be singing through a cloth. There is a lack of presence.
The acoustic guitar beginning Hotel California with V5 sounds beautiful in its richness of color and texture. Its presence jumps out. The bass is stronger and more defined than with the OPA627 and the music seems more accurate. The instruments keep increasing as the music plays, but textural space is created for them. The vocal could be a bit more forward. But a great performance.
Maggot Brain by Funkadelic
In Maggot Brain with the OPA627 the echo in the left speaker of this beautiful guitar solo brings me right into the studio. The OPA627 shines when outputting the distorted gritty sounds displayed on this cut. I would like a little more transparency here but otherwise great performance.
With the Burson V5 the vocal is more authoritative at the beginning. The drum slap and echo is more detailed and colored. The solo is more in our face but cleaner than with the OPA627, although I like the OPA627 distortion guitar solo sound better. But the sound separation and soundstage are excellent with the Burson V5.
Early in The Morning by Harry Nilsson
Nilsson’s bouncy bluesy keyboard and infectious voice and lyric wants to be heard. But despite some spatial openness and slight texture separating the notes and voice, it was a boring presentation with OPA627.
The Burson V5 again has more detail and bass, but still the song lacks a natural and intimate soundstage. I want it to feel live, but it doesn’t.
The Burson V5 performed better to my ears and with my equipment. It was less fatiguing than the OPA627. The frequency curves on the OPA627 and Burson V5 were identical with REW. The great details, texture, dynamics, presence, etc. with the Burson V5 made no change in frequency distribution. They were both flat.
Burson V6 Classic performance
The Burson V6 Classic has been reviewed frequently and most reviewers indicate it is a good fit for folk and jazz, but not necessarily for rock. I found this to be false in my evaluation. It is clear it is more evolved in design than the earlier produced V5 I already owned, particularly the better ventilation qualities on the V6. I chose the V6 Classic for review since the V6 vivid is based on the Burson V5 I already own and they seem to share many characteristics, if the published qualities in the bar graph are accurate. The V6 Classic seemed different enough from the V5 to interest me in hearing its advertised sonic qualities. Since the OPA627 was found lacking in previous listening tests, I filled the integrator op-amp socket of my Cyenne DAC with a Burson V5 and the gain op-amp sockets with the two V6 Classic op-amps I received from Burson.
Joni Mitchell The Asylum Albums CD 1 and CD 4
I could hear Joni’s deliberate pace in the music that reveals a more realistic music experience with a musical texture which was beautiful and a clarity that was a joy to listen to. Joni in her live performances on CD 4 sounded present in my listening room with a sweetness to her voice which was big and clear, yet textured with a beautiful timbre to her voice I had not previously heard; while the instruments encompassed me. The bass was strong amongst rounded vocals that echoed slightly, revealing the music hall acoustics. I had never heard such a beautiful voice, although I had heard Joni many times. Her phrasing was accomplished.
Roger Water, Amused to Death
The first five tracks on Roger Waters album, known for its soundstage and imagery, were listened to with the Burson V6 Classic op-amps. The spacing and imagery of the sound was revealing. The texture and dynamics of the music, with pounding drums, electric solo and the chorus singing was three dimensional. The best I have heard it. The presence of Rogers is real. I had oddly read that V6 Classic op-amps were not good for rock, as previously noted, but I had a different experience, the realism the Classics added to my system was nothing short of fantastic. The sound wrapped around the back of my head. Roger seemed to be singing in my face, he is so close.
Early in The Morning by Harry Nilsson
I was so disappointed in the OPA627 and Burson V5 presentation of Early in the Morning I wanted to give the V6 Classics a chance to properly present the song. With the Burson V6 Classics in the gain op-amp sockets, finally the song sounds real. The keys of the piano can be heard vibrating and Harry is singing next to me. There is a brilliant texture revealed by the Classic op-amps positioning the struck keys, voice and distant bass. The realism of musical timbre is heard with an energy in the sound not previously heard.
Kill City by Iggy Pop
Kill City with Iggy Pop followed, with Williamson on guitar. It is a textually dense and energetic album. Iggy is in your face while the listener is enveloped by this forceful performance. The songs are catchy and floor tapping. This is a vinyl quality performance in sound by the Burson V5 and Burson V6 Classic combination. There is an absence of distortion, only the authentic presentation of instruments and voice. A total musical presentation which envelopes you with its presence, but maintains its clarity and details. The tone of each note is cleanly offered, while grandly performed by Iggy in his musical prime.
Avenue B By Iggy Pop
Avenue B by Iggy is live and his spoken introduction echoed as he spoke in the theater. His guitar strings vibrate in front of me. His voice is huge. The texture and details of the sound is so real I believe I am in the theater. Details in his performance are brilliantly revealed in this live performance which I had not previously heard. It was obvious now the sounds were always present, but only revealed now with the V6 Classics by Burson in the Cyenne 3100 DAC.
Clearly many things affect my perception of the V5 combined with the V6. The V6 Classics are so much more realistic and texturally beautiful than the Burson V5. I am sure though they complement each other with their revealing musical presentation and clear sonic image and details previously never heard on my system. I would pay for the V6 Classic op-amps if they were not offered free for a review.
I do not know how Burson was able to design the V6 Classic to produce such a musically beautiful presentation, suitable for rock or folk in my view. I am convinced from my experience that op-amps can change the sound signature of a DAC for the better. The best op-amps I have found to my taste have been made by Burson. I realize an op-amp interacts with the DAC circuitry and another DAC might produce a difference in sound quality, but in my system, the sound was superb and these op-amps will remain in place.
The value of discrete op-amps is denigrated by many, however many such as PS Audio co-founder Paul McGowan mention the intolerance of heat of IC op-amps due to size limitations; while discrete op-amps are not limited in this respect. He further states that discrete op-amps can be designed with a certain sound in mind and the use of the highest performing circuits and endless designer choices. Of course, the success of the project depends on the ability of the designer.
Specifications can not alone explain an op-amps appeal. Claims that discrete op-amps have more distortion than IC op-amps is usually used by critics to support the claim discrete op-amps are not worth buying or using. Gaskell’s 2016 dissertation at McGill university mentions that many designers and critical listeners believe that low level distortion may impart a beneficial tonal, timbral, or transient quality. Small amounts of distortion in a signal affect the “warmth” of the sound. He lists several studies showing that distortion is preferred by many listeners and his dissertation’s conclusion is that low level capacitor and op-amp distortion was found pleasing to many listeners.
Metrics can only replace subjective testing if they correlate well with a jury response. Measurement equipment in audio has to correlate with what people find pleasing in music, not what measurements such as THD say is best. Gaslighting someone who says something sounds good when it has poorer specifications is not a way to improve the science of psychoacoustics in the development of measures which quantity what consumers find pleasing.
Finally, Harman’s “How to Listen” is a computer-based, self-guided software program which teaches listeners how to identify, classify and rate the quality of recorded and reproduced sounds according to their timbral, spatial, dynamic and nonlinear distortion attributes. Other critical listening programs are available and can be used to improve the evaluation and improvement of audio systems through trained listener evaluations of products. Many audio companies do this to improve speakers, DACs, Amps, etc. Improvement of listening skills may improve the opportunity for agreement between critics as to what sounds good. One study found audio salesmen had the best listening skills, better than music critics.
I would be happy to review any discrete op-amps that manufacturers wish to have honestly evaluated.
Images show OPA627, Burson V5 and Buson V6 Classic op-amps and their relative size and Cyenne Dac 3100 with OPA627 and two V6 Classic op-amps in sockets.
In exchange for two Burson Classic dual op-amps I writing my honest review of their performance in comparison to OPA627 and the Burson V5s I already own. My system is composed of an Asus Gaming Computer running JRiver media center 28 which I modified to up samples all files to 96 KHz which are stored on a Passport external hard drive. I have an external Musiland external soundcard I use with optical out through a glass optical interconnect to a Cyenne 3100 DAC with three replaceable op-amp sockets. The DAC output goes to a Cambridge Audio CXA60 amp. I have Mordaunt-Short Aviano 6 speakers and a Wharfedale subwoofer. The Mordaunt-Short speakers are very neutral and I can hear differences in op-amps in the DAC typically. Lm4562 DAC chips came with the Cyenne Dac originally, but I found that despite their low THD, they sounded clean, but lean in expression. I previously owned a Maverick D2 DAC with 3 OPA627 op-amps and found them more textured and open. So, I am reviewing the OPA627 as a baseline vs the two Burson op-amps to see any difference. My dedicated music room is highly acoustically treated and has a fairly straight equalization curve along the x axis, according to Room EQ Wizard, without any equalization required in the JRiver application to achieve this.
Above example of ceiling room treatment in dedicated music room.
I plan to look at the factors Buson uses to describe their op-amps: transparency, details, color and texture, and dynamics and soundstage when reviewing the op-amps previously noted.
Definitions:
Transparency: the sound is reproduced accurately without adding much extra.
Detail: hearing the small, fine sounds of a recording; the most delicate parts of the original sound.
Color and texture: Tone color or texture of a sound is the quality that allows you to separate two instruments playing the same instrument. It is hearing the different layers and harmonies of music.
Dynamics and soundstage: variation in loudness between notes in the spatial sound image presented to the listener.
The Cyenne Dac has 3 removal socketed op-amps. One serves as the combiner op-amp for the DAC. The other 2 are gain op-amps. The gain sockets will vary from OPA627 to the Burson V5 to V6 Classic. I will listen to a playlist of Flac files with a minimum 44.1 KHz and make note of the above features of transparency, etc. for any sound characteristics noticeable for each op-amp.
The flowing listening set began with OPA627 in all three op-amp sockets followed by the Burson V5 in the gain op-amp sockets and OPA627 in the combiner op-amp socket.
Just Like A woman, live 1966 by Bob Dylan.
This is a beautiful 6:45 version of his studio song which only totals 4.42. With three OPA627 op-amps the song sounded less detailed than I remembered. I choose my music in this evaluation for its familiarity. Reference material I believe has to be used so elements like detail and transparency can be assessed as different from a typical presentation. Dylan’s guitar and voice had a sadly muted color and texture. It is actually a beautiful version of a song in which we hear Dylan during a grueling tour in 1966 play an emotionally different version of his studio album while leaning over his stool and expressing longingness and compassion during the song, not the derision found in his studio album. But the OPA627 did not convince me I was listening to a live performance of a wasted Dylan generously sharing his feelings of tenderness with the audience.
The Burson V5 turned the concert and song live. We hear the echo in the room and his harmonica notes are more present. The detail is enhanced. The whole texture and color of the song in his strained voice emoting his regret, not his signature emotional sneer. I was enveloped in the large soundstage during the recitation.
Visions of Johanna by Dylan, the studio track from Blonde on Blonde
The song sounded less textured as well with OPA627. Details were diminished. The notes were compressed and the soundstage was narrow. The music was colored, but with all the instruments the same color.
The Burson V5 again brought out more realism and detail. The bass is strong and Dylan sounds real, unfortunately the soundstage is medium but with clear textured sound separation with sharp details and accuracy.
Hotel California by The Eagles, MTV Live.
With the OPA627 texture is diminished in the guitar work and soundstage is small. Details seem muted. Henley seems to be singing through a cloth. There is a lack of presence.
The acoustic guitar beginning Hotel California with V5 sounds beautiful in its richness of color and texture. Its presence jumps out. The bass is stronger and more defined than with the OPA627 and the music seems more accurate. The instruments keep increasing as the music plays, but textural space is created for them. The vocal could be a bit more forward. But a great performance.
Maggot Brain by Funkadelic
In Maggot Brain with the OPA627 the echo in the left speaker of this beautiful guitar solo brings me right into the studio. The OPA627 shines when outputting the distorted gritty sounds displayed on this cut. I would like a little more transparency here but otherwise great performance.
With the Burson V5 the vocal is more authoritative at the beginning. The drum slap and echo is more detailed and colored. The solo is more in our face but cleaner than with the OPA627, although I like the OPA627 distortion guitar solo sound better. But the sound separation and soundstage are excellent with the Burson V5.
Early in The Morning by Harry Nilsson
Nilsson’s bouncy bluesy keyboard and infectious voice and lyric wants to be heard. But despite some spatial openness and slight texture separating the notes and voice, it was a boring presentation with OPA627.
The Burson V5 again has more detail and bass, but still the song lacks a natural and intimate soundstage. I want it to feel live, but it doesn’t.
The Burson V5 performed better to my ears and with my equipment. It was less fatiguing than the OPA627. The frequency curves on the OPA627 and Burson V5 were identical with REW. The great details, texture, dynamics, presence, etc. with the Burson V5 made no change in frequency distribution. They were both flat.
Burson V6 Classic performance
The Burson V6 Classic has been reviewed frequently and most reviewers indicate it is a good fit for folk and jazz, but not necessarily for rock. I found this to be false in my evaluation. It is clear it is more evolved in design than the earlier produced V5 I already owned, particularly the better ventilation qualities on the V6. I chose the V6 Classic for review since the V6 vivid is based on the Burson V5 I already own and they seem to share many characteristics, if the published qualities in the bar graph are accurate. The V6 Classic seemed different enough from the V5 to interest me in hearing its advertised sonic qualities. Since the OPA627 was found lacking in previous listening tests, I filled the integrator op-amp socket of my Cyenne DAC with a Burson V5 and the gain op-amp sockets with the two V6 Classic op-amps I received from Burson.
Joni Mitchell The Asylum Albums CD 1 and CD 4
I could hear Joni’s deliberate pace in the music that reveals a more realistic music experience with a musical texture which was beautiful and a clarity that was a joy to listen to. Joni in her live performances on CD 4 sounded present in my listening room with a sweetness to her voice which was big and clear, yet textured with a beautiful timbre to her voice I had not previously heard; while the instruments encompassed me. The bass was strong amongst rounded vocals that echoed slightly, revealing the music hall acoustics. I had never heard such a beautiful voice, although I had heard Joni many times. Her phrasing was accomplished.
Roger Water, Amused to Death
The first five tracks on Roger Waters album, known for its soundstage and imagery, were listened to with the Burson V6 Classic op-amps. The spacing and imagery of the sound was revealing. The texture and dynamics of the music, with pounding drums, electric solo and the chorus singing was three dimensional. The best I have heard it. The presence of Rogers is real. I had oddly read that V6 Classic op-amps were not good for rock, as previously noted, but I had a different experience, the realism the Classics added to my system was nothing short of fantastic. The sound wrapped around the back of my head. Roger seemed to be singing in my face, he is so close.
Early in The Morning by Harry Nilsson
I was so disappointed in the OPA627 and Burson V5 presentation of Early in the Morning I wanted to give the V6 Classics a chance to properly present the song. With the Burson V6 Classics in the gain op-amp sockets, finally the song sounds real. The keys of the piano can be heard vibrating and Harry is singing next to me. There is a brilliant texture revealed by the Classic op-amps positioning the struck keys, voice and distant bass. The realism of musical timbre is heard with an energy in the sound not previously heard.
Kill City by Iggy Pop
Kill City with Iggy Pop followed, with Williamson on guitar. It is a textually dense and energetic album. Iggy is in your face while the listener is enveloped by this forceful performance. The songs are catchy and floor tapping. This is a vinyl quality performance in sound by the Burson V5 and Burson V6 Classic combination. There is an absence of distortion, only the authentic presentation of instruments and voice. A total musical presentation which envelopes you with its presence, but maintains its clarity and details. The tone of each note is cleanly offered, while grandly performed by Iggy in his musical prime.
Avenue B By Iggy Pop
Avenue B by Iggy is live and his spoken introduction echoed as he spoke in the theater. His guitar strings vibrate in front of me. His voice is huge. The texture and details of the sound is so real I believe I am in the theater. Details in his performance are brilliantly revealed in this live performance which I had not previously heard. It was obvious now the sounds were always present, but only revealed now with the V6 Classics by Burson in the Cyenne 3100 DAC.
Clearly many things affect my perception of the V5 combined with the V6. The V6 Classics are so much more realistic and texturally beautiful than the Burson V5. I am sure though they complement each other with their revealing musical presentation and clear sonic image and details previously never heard on my system. I would pay for the V6 Classic op-amps if they were not offered free for a review.
I do not know how Burson was able to design the V6 Classic to produce such a musically beautiful presentation, suitable for rock or folk in my view. I am convinced from my experience that op-amps can change the sound signature of a DAC for the better. The best op-amps I have found to my taste have been made by Burson. I realize an op-amp interacts with the DAC circuitry and another DAC might produce a difference in sound quality, but in my system, the sound was superb and these op-amps will remain in place.
The value of discrete op-amps is denigrated by many, however many such as PS Audio co-founder Paul McGowan mention the intolerance of heat of IC op-amps due to size limitations; while discrete op-amps are not limited in this respect. He further states that discrete op-amps can be designed with a certain sound in mind and the use of the highest performing circuits and endless designer choices. Of course, the success of the project depends on the ability of the designer.
Specifications can not alone explain an op-amps appeal. Claims that discrete op-amps have more distortion than IC op-amps is usually used by critics to support the claim discrete op-amps are not worth buying or using. Gaskell’s 2016 dissertation at McGill university mentions that many designers and critical listeners believe that low level distortion may impart a beneficial tonal, timbral, or transient quality. Small amounts of distortion in a signal affect the “warmth” of the sound. He lists several studies showing that distortion is preferred by many listeners and his dissertation’s conclusion is that low level capacitor and op-amp distortion was found pleasing to many listeners.
Metrics can only replace subjective testing if they correlate well with a jury response. Measurement equipment in audio has to correlate with what people find pleasing in music, not what measurements such as THD say is best. Gaslighting someone who says something sounds good when it has poorer specifications is not a way to improve the science of psychoacoustics in the development of measures which quantity what consumers find pleasing.
Finally, Harman’s “How to Listen” is a computer-based, self-guided software program which teaches listeners how to identify, classify and rate the quality of recorded and reproduced sounds according to their timbral, spatial, dynamic and nonlinear distortion attributes. Other critical listening programs are available and can be used to improve the evaluation and improvement of audio systems through trained listener evaluations of products. Many audio companies do this to improve speakers, DACs, Amps, etc. Improvement of listening skills may improve the opportunity for agreement between critics as to what sounds good. One study found audio salesmen had the best listening skills, better than music critics.
I would be happy to review any discrete op-amps that manufacturers wish to have honestly evaluated.
XT25 + Visaton WG 148 R with pictures
Hello everyone,
I just mounted the Vifa/Peerless XT25TG30-04 on a Visaton 148 R waveguide and thought that I'd post some pics of how I did it in case someone is interested.
First here is the couple:
Due to the slightly curved front plate of the XT25 a small ~2 mm gap remains between the tweeter and the wg which I thought might not do any good to the sound:
A cheap solution was found from blutack or what ever this sticky stuff is called in English. I made a thin donut and placed it to the very edge of the opening in the waveguide:
I also put three small blobs of blutack on the tweeter face while I aligned the wg to make sure that it stays in place while I mark the holes.
I don't have any pics of the alignment process itself, but it's of course very straightforward. I used a 4 mm wood drill bit to push small marks on the wg wg back plate as the bit fit the holes nicely and was thus centered well enough automatically.
Some of the holes go close to the original mounting holes with thread inserts on the wg, but the original holes are not at the same distance from the center point as the XT25 mounting holes so they can't be used.
I first put some white paint around the mark to make the mark more visible for drilling and used a normal 2,5 mm metal drill bit for the hole.
3,5x13 DIN 7981 screws and washers were used to mounting:
And here is the result from the front:
I think they came out quite nice.
No idea how the sound is yet, though.😱
I just mounted the Vifa/Peerless XT25TG30-04 on a Visaton 148 R waveguide and thought that I'd post some pics of how I did it in case someone is interested.
First here is the couple:
Due to the slightly curved front plate of the XT25 a small ~2 mm gap remains between the tweeter and the wg which I thought might not do any good to the sound:
A cheap solution was found from blutack or what ever this sticky stuff is called in English. I made a thin donut and placed it to the very edge of the opening in the waveguide:
I also put three small blobs of blutack on the tweeter face while I aligned the wg to make sure that it stays in place while I mark the holes.
I don't have any pics of the alignment process itself, but it's of course very straightforward. I used a 4 mm wood drill bit to push small marks on the wg wg back plate as the bit fit the holes nicely and was thus centered well enough automatically.
Some of the holes go close to the original mounting holes with thread inserts on the wg, but the original holes are not at the same distance from the center point as the XT25 mounting holes so they can't be used.
I first put some white paint around the mark to make the mark more visible for drilling and used a normal 2,5 mm metal drill bit for the hole.
3,5x13 DIN 7981 screws and washers were used to mounting:
And here is the result from the front:
I think they came out quite nice.
No idea how the sound is yet, though.😱
Genuine Tripath evaluation boards RB-TK2350;EB-TA0103;RB-TA0105
- By brunomiguel
- Swap Meet
- 4 Replies
Attachments
Has anyone tried an Icosahedron (20 sided with triangular facets.)
- By jgaupel
- Full Range
- 42 Replies
I share my ~shop with my brother and he is making these for fun.
https://en.wikipedia.org/wiki/Icosahedron#Convex_regular_icosahedron
What would anyone think the downfalls of using this shape for a cabinet be?
I,m thinking that a four inch bezel, 3 to 3.5 inch cone full range might be fun. To have a facet big enough to mount a four inch the inside edge (using 1/2 inch material) is 8.5 inches giving a volume of 0.7753665740741 cubic feet.
(I love to leave all those digits as if my measuring devices are N.A.S.A approved and calibrated.)
Any thoughts??
Jeremy
https://en.wikipedia.org/wiki/Icosahedron#Convex_regular_icosahedron
What would anyone think the downfalls of using this shape for a cabinet be?
I,m thinking that a four inch bezel, 3 to 3.5 inch cone full range might be fun. To have a facet big enough to mount a four inch the inside edge (using 1/2 inch material) is 8.5 inches giving a volume of 0.7753665740741 cubic feet.
(I love to leave all those digits as if my measuring devices are N.A.S.A approved and calibrated.)
Any thoughts??
Jeremy
Horn Loaded NEO 8 test results (modified/stretched Joseph Crowe NEO 3 horn).
- By Gigapunk
- Planars & Exotics
- 3 Replies
I'm building three line arrays with 6x GRS clones of the BG NEO 8 per speaker. I bought the Joseph Crowe Neo3 tweeter horn CAD file and elongated the design to fit the NEO 8. I then 3D printed the design in two different widths. There has been some discussion that blocking the 2 outermost slots in the NEO3/NEO8 speakers improves off-axis dispersion, so one of the designs only allows the center two of the four slots of the speaker to be exposed. I also tested two different sizes of rear chambers. The idea for my line array is to make very large (but shallow) speakers that will fit behind a 100" acoustically transparent projector screen. My modifications of Joseph's original design eliminates the nice round overs that curve to the back, so that the horns would fit flush in a large infinite baffle style speaker.
Here are all of the measurements together.
The complete Line Array would be 6x NEO8s stacked like below.
Here's the two designs, One Narrow/2-Slots Wide, one normal width/4-Slots.
These were rather rough prints with 1.0 x 0.3 mm layers. I printed in ABS because it sands and glues well. I spent minimal effort on sanding.
Here's the flat baffle I tested along with the large rear chamber (Old 6x9 Car audio box), rear chamber (green 3D print), and the two horns (2 and 4 slot).
The small chamber had a weird resonance around 2.4k The difference in level may not be real, as I forgot to test the small chamber, and had to re-set everything up to get the measurement. All of the other measurements below were done with the large chamber.
Small (blue) vs Large (red) chamber:
Large Chamber Distortion:
Small Chamber Distortion:
The flat baffle with flush-mounted GRS NEO8:
Straight-on measurement (0 degrees) of the flat baffle (red), narrow horn (green), wide/normal horn (purple).
15 Degrees. flat baffle (green), narrow horn (orange), wide/normal horn (red).
30 degrees:
Interesting that the narrow slot horn did help off-axis dispersion, but only above 16k.
45 degrees:
Flat baffle (no horn) 0, 15, 30, 45 degrees
Narrow Slot Horn at 0, 15, 30, 45 degrees
Normal Width Slot Horn at 0, 15, 30, 45 degrees
Final thoughts:
The normal-width horn did not have a very smooth output which was surprising, but it did have the same increase in output from 1k to 3k as the narrow horn.
I was also surprised that the off-axis levels diverged lower (1k vs 4k) with the horns.
Interesting that the narrow slot horn did help off-axis dispersion, but only above 16k; also interesting that the narrow slot was actually the loudest everywhere.
I don't think that I'm going to end up using the horns in my line array. The more gently upsloping response of the flat baffle should play well with low end lift that you get from a line array. And the lack of the sudden increase in response around 1k (that the horns have) should allow me to cross over the flat baffle mounted speakers lower (500-700 Hz).
Joseph already posted about a horn he actually made specifically for the NEO8. His wooden horn was much larger, and helps extend the usefull range. My 3D printed horn raised output down to ~1200 Hz, but Josephs's larger horn increased output down to 300 Hz! He posted about it here: https://www.diyaudio.com/community/threads/grs-pt6825-8-planar-test-results-300hz-horn.379684/
The fact that a horn-loaded NEO 8 can play down to 300 HZ is amazing. I'm convinced now that there is really no point to the NEO3. The NEO8 plays the upper range just as well, but allows a 2-way speaker with a lower crossover and probably more headroom as well.
Would anybody else choose one of the horns over the flat baffle in a line array? What do you guys think? Big thank you to Joseph Crowe for making his CAD files available, makes this silliness possible.
Here are all of the measurements together.
The complete Line Array would be 6x NEO8s stacked like below.
Here's the two designs, One Narrow/2-Slots Wide, one normal width/4-Slots.
These were rather rough prints with 1.0 x 0.3 mm layers. I printed in ABS because it sands and glues well. I spent minimal effort on sanding.
Here's the flat baffle I tested along with the large rear chamber (Old 6x9 Car audio box), rear chamber (green 3D print), and the two horns (2 and 4 slot).
The small chamber had a weird resonance around 2.4k The difference in level may not be real, as I forgot to test the small chamber, and had to re-set everything up to get the measurement. All of the other measurements below were done with the large chamber.
Small (blue) vs Large (red) chamber:
Large Chamber Distortion:
Small Chamber Distortion:
The flat baffle with flush-mounted GRS NEO8:
Straight-on measurement (0 degrees) of the flat baffle (red), narrow horn (green), wide/normal horn (purple).
15 Degrees. flat baffle (green), narrow horn (orange), wide/normal horn (red).
30 degrees:
Interesting that the narrow slot horn did help off-axis dispersion, but only above 16k.
45 degrees:
Flat baffle (no horn) 0, 15, 30, 45 degrees
Narrow Slot Horn at 0, 15, 30, 45 degrees
Normal Width Slot Horn at 0, 15, 30, 45 degrees
Final thoughts:
The normal-width horn did not have a very smooth output which was surprising, but it did have the same increase in output from 1k to 3k as the narrow horn.
I was also surprised that the off-axis levels diverged lower (1k vs 4k) with the horns.
Interesting that the narrow slot horn did help off-axis dispersion, but only above 16k; also interesting that the narrow slot was actually the loudest everywhere.
I don't think that I'm going to end up using the horns in my line array. The more gently upsloping response of the flat baffle should play well with low end lift that you get from a line array. And the lack of the sudden increase in response around 1k (that the horns have) should allow me to cross over the flat baffle mounted speakers lower (500-700 Hz).
Joseph already posted about a horn he actually made specifically for the NEO8. His wooden horn was much larger, and helps extend the usefull range. My 3D printed horn raised output down to ~1200 Hz, but Josephs's larger horn increased output down to 300 Hz! He posted about it here: https://www.diyaudio.com/community/threads/grs-pt6825-8-planar-test-results-300hz-horn.379684/
The fact that a horn-loaded NEO 8 can play down to 300 HZ is amazing. I'm convinced now that there is really no point to the NEO3. The NEO8 plays the upper range just as well, but allows a 2-way speaker with a lower crossover and probably more headroom as well.
Would anybody else choose one of the horns over the flat baffle in a line array? What do you guys think? Big thank you to Joseph Crowe for making his CAD files available, makes this silliness possible.
Can anyone handle the idea of a CD that costs $1,200 each? I can... .
- The Lounge
- 50 Replies
Attachments
EV 1824M/8HD horn mids and T35A Tweeters
- Swap Meet
- 4 Replies
TWEETERS ARE SOLD EV T-35A tweeters $90 per pair plus shipping. Taken from my home stereo. Resistance measurements in pictures. I have 2 pairs, and they are very similar. Take both pair for $150. Shipping will be 10x8x6 inches and 4 pounds, shipped from 28601.
MIDS STILL AVAILABLE EV 1824M mid driver with 8hd horns. One pair (A and B) are in excellent condition for $200 per pair plus shipping the other pair (C and D) have small toddler finger indentations on the screen but work and sound fine, $150. Taken from my home stereo. Resistance measurements in pictures. I have 2 pairs, and they are very similar. Take both pair for $300. Shipping will be 16x12x12 and 20 pounds per pair, shipped from 28601.
View attachment 1010061View attachment 1010062View attachment 1010063
MIDS STILL AVAILABLE EV 1824M mid driver with 8hd horns. One pair (A and B) are in excellent condition for $200 per pair plus shipping the other pair (C and D) have small toddler finger indentations on the screen but work and sound fine, $150. Taken from my home stereo. Resistance measurements in pictures. I have 2 pairs, and they are very similar. Take both pair for $300. Shipping will be 16x12x12 and 20 pounds per pair, shipped from 28601.
View attachment 1010061View attachment 1010062View attachment 1010063
Bookshelf speaker with decent bass?
Hello everybody, newbie here:
I'm looking for a pair of bookshelf speakers for my PC (not for monitoring, just play games and listen to music), something better than the typical 2.1 Logitech kit, but in a 2.0 fashion trying not to miss the bass so much.
Searching on internet the DIY project "DINAS - Do I need a sub" was pretty promising:
Login to view embedded media
Until I came across these comments about the amp:
I'm not sure about the TPA3116 as alternative due I'm currently using one on a Bluetooth chinese board from Aliexpress and it has noticeable background noise and a 'hiss' noise when a bluetooth device is connected (maybe is another component and not the amp itself):
On top of that i'm from EU and the Lepai LP210PA ($63 on Parts Express) costs 105€ at soundimports.eu. Almost double the price 😵
So my questions are:
Thanks!
I'm looking for a pair of bookshelf speakers for my PC (not for monitoring, just play games and listen to music), something better than the typical 2.1 Logitech kit, but in a 2.0 fashion trying not to miss the bass so much.
Searching on internet the DIY project "DINAS - Do I need a sub" was pretty promising:
Login to view embedded media
- Amp: Lepai LP210PA
- Subwoofer: Tang Band W5-1138SMF
- Woofer/mids: Dayton Audio RS100-4
- Tweeter: tn28b
Until I came across these comments about the amp:
- "The Lepai LP210PA is based on TPA3118 chips without heatsinks so don't use with 4ohm loads! use TPA3116 instead"
- "The power supply that usually comes in the kit is completely insufficient to give full power"
I'm not sure about the TPA3116 as alternative due I'm currently using one on a Bluetooth chinese board from Aliexpress and it has noticeable background noise and a 'hiss' noise when a bluetooth device is connected (maybe is another component and not the amp itself):
On top of that i'm from EU and the Lepai LP210PA ($63 on Parts Express) costs 105€ at soundimports.eu. Almost double the price 😵
So my questions are:
- Suggestions?
- Alternative /better components?
- Is there a more convenient way to sound my PC instead of buying 2 amps + 2 good power supply + DAC with volume knob? I've seen some monitors that have all integrated in a single plate:
Thanks!
For Sale Vifa/PAE Shielded BC25SG18-04 Tweeters, New Pair
- By GeoffMillar
- Swap Meet
- 0 Replies
Attachments
3" or 3.5" full-range driver that works with infinite-baffle ?
- By CZ Eddie
- Full Range
- 18 Replies
I know this is an odd request. So, I'm putting it in the Full-range section where oddball speakers are the name of the game.
I'm replacing the incredibly cheap factory 3.5" speakers that came preinstalled in the top of my enclosed shower unit.
The back of the speaker is infinite baffle, using my entire bathroom as its enclosure.
While the front of the speaker fires into a medium-closet sized shower enclosure.
I know I won't get any bass. But can you think of any drivers that would excel more than something designed for a tiny speaker box and would work from I guess, 120hz-12Khz (or better)?
I'll need something that isn't un-treated paper.
Because it won't get wet but there will be lots of moisture in the air.
Original speakers:
Per my digital calipers, the cutout diameter is a tad over 3.3".
I'm replacing the incredibly cheap factory 3.5" speakers that came preinstalled in the top of my enclosed shower unit.
The back of the speaker is infinite baffle, using my entire bathroom as its enclosure.
While the front of the speaker fires into a medium-closet sized shower enclosure.
I know I won't get any bass. But can you think of any drivers that would excel more than something designed for a tiny speaker box and would work from I guess, 120hz-12Khz (or better)?
I'll need something that isn't un-treated paper.
Because it won't get wet but there will be lots of moisture in the air.
Original speakers:
Per my digital calipers, the cutout diameter is a tad over 3.3".
Anyone know what these transformers are?
- By djn
- Tubes / Valves
- 14 Replies
I can’t tell if these transformers would be useful for audio builds.
Any help???
Any help???
Steeplejack ..... or how F8 met Mighty Babelfish M25 (& Schade)
It started here : https://www.diyaudio.com/community/threads/scryer-or-how-f8-met-mighty-sissysit.388550/
After that one, couldn't avoid my own greediness to try this combo too
Though, clever as I am, God's ways are his own and Buddha is everywhere........ meaning - better name for this one would be Mad Hatter, you'll see from 4R load behavior ........
Ok, for starters usual set of graphic files
Schematic is aggregate, find separate ones for pure Mos and for Schade later
After that one, couldn't avoid my own greediness to try this combo too
Though, clever as I am, God's ways are his own and Buddha is everywhere........ meaning - better name for this one would be Mad Hatter, you'll see from 4R load behavior ........
Ok, for starters usual set of graphic files
Schematic is aggregate, find separate ones for pure Mos and for Schade later
Attachments
Denafrips ARES II board repair
- By mattworld
- Digital Line Level
- 4 Replies
I just got a spare Denafrips ARES II board. I bought it used ("in working condition") but I doubt it is the case. The initial plan was to use it in a DIY DAC perhaps with a lithium battery power supply. But first I need to power-up this board, repair it and test it. Can you help me with that ?
I am looking for information about the original transformer and how it is wire on the board (PIN 1 to 8) ?
Here is the top of the board:
I already see two problems on the board :
I am looking for information about the original transformer and how it is wire on the board (PIN 1 to 8) ?
Here is the top of the board:
I already see two problems on the board :
The two center PIN of the chip (red arrow) were connected. Any clue of the chip reference ? I think it is a LDO fix the gate voltage the transistor. | On the LT1763 LDO seems defected (yellow, too much heat) |
Attachments
Woden Empress of India, BVR for Alpair 12pw
- By planet10
- Planet 10 hifi
- 11 Replies
Big Vent Relex for the Mark Audio Alpair 12pw. Comment from builder (he may chip in)
Contach me or Scott about getting acess to the full planset.
dave
Contach me or Scott about getting acess to the full planset.
dave
Digital level meter for R-R tape deck
- By Leoniv
- Analogue Source
- 5 Replies
The analog level meter has been designed for use as a recording level meter for a reel to reel tape recorder. But it can also be used as a standalone device. The LED bar can contain up to 50 elements per channel. To reduce the level of interference, static indication is used. A feature of the meter is fully digital signal processing. That allows you to make the meter metrologically accurate, with an exact match of the dynamic characteristics. All parameters are programmed using special software on a computer.
Each meter channel contains two signal processing branches, the readings of the first one are displayed as a bar, the second - as a dot. This can be used to display two values, such as average + quasi-peak, or quasi-peak + true peak. For each of the branches, you can individually set the gain, integration time, response time, decay time, hold time, fall time, peak hold time, peak fall time. The scale type is set by assigning an individual threshold level to each LED with an accuracy of 0.01 dB. This allows you to build a linear, logarithmic, S-shaped with a stretch near 0 dB and other scales. A set of all parameters, including the scale type, is stored in EEPROM as a preset. There can be 4 different presets in total. They can be selected with the control button. The second button can turn on the indication of statistics of maximum levels, as well as reset it.
Here we will talk about how simple things can be made difficult 🙂
Block diagram
The block diagram of the level meter is shown in the figure below. This diagram shows one stereo channel of the meter. The input signal is fed to a differential amplifier implemented on an op-amp (Meter50_v4_sch.pdf). It performs two tasks: it shifts the signal by half the ADC scale and allows you to implement a balanced input. The signal shift is necessary because the built-in ADC is unipolar, its input range is from 0 to the reference voltage. A balanced input is desirable even if the signal source is single ended. It allows you to take the signal relative to "analogue" ground and eliminate the interference present on the ground of the meter circuit. If the level meter is made as a separate device, then balanced inputs with XLR connectors can be implemented.
All signal processing is done digitally by the STM32F100/103 microcontroller. To avoid the use of an anti-alias filter at the ADC input, as well as to be able to register short signal peaks, the sampling frequency was chosen quite high - 96 kHz. With DMA, ADC samples are stored in a buffer in RAM. When half of the array is ready, an interrupt occurs, the data is read and processed.
DC removal filter
The first operation on digital data is to remove the DC component from the signal. This must be done before the signal is sent to the detector. With a meter dynamic range of about 60 dB, even such a small DC component as 0.1% of the ADC scale can be equal in magnitude to the useful signal and distort it.
Detector
Next, a full-wave rectification of the signal is performed. This is the easiest part of processing. When the DC component is removed from the signal, the task is to calculate the absolute value.
Filters
Next are several filters that define the dynamic characteristics of the meter.
Integration time
To obtain a given integration time, the detector output must be smoothed with a filter. The smoothing filter is not a simple low-pass filter, but a more complex filter with different charge and discharge time constants. The charge must be fast to provide the specified meter integration time. And the discharge must occur slowly to provide a given return time, which is much longer. Therefore, the integration time filter is turned on only when the signal rises; when the signal decreases, it is kept at the same level. A separate filter is used for signal decay.
The difference filter equation has the following form:
y( n ) = y(n-1) + A*(x( n ) - y(n-1)), where A = 0.0083 for Fs = 96 kHz
A 32-bit accumulator is used to obtain the required calculation accuracy. Filter coefficients are 16-bit unsigned. They are defined as A = 65536 / 96000 / tau.
Each meter channel contains two signal processing branches, the readings of the first one are displayed as a bar, the second - as a dot. This can be used to display two values, such as average + quasi-peak, or quasi-peak + true peak. For each of the branches, you can individually set the gain, integration time, response time, decay time, hold time, fall time, peak hold time, peak fall time. The scale type is set by assigning an individual threshold level to each LED with an accuracy of 0.01 dB. This allows you to build a linear, logarithmic, S-shaped with a stretch near 0 dB and other scales. A set of all parameters, including the scale type, is stored in EEPROM as a preset. There can be 4 different presets in total. They can be selected with the control button. The second button can turn on the indication of statistics of maximum levels, as well as reset it.
Here we will talk about how simple things can be made difficult 🙂
Block diagram
The block diagram of the level meter is shown in the figure below. This diagram shows one stereo channel of the meter. The input signal is fed to a differential amplifier implemented on an op-amp (Meter50_v4_sch.pdf). It performs two tasks: it shifts the signal by half the ADC scale and allows you to implement a balanced input. The signal shift is necessary because the built-in ADC is unipolar, its input range is from 0 to the reference voltage. A balanced input is desirable even if the signal source is single ended. It allows you to take the signal relative to "analogue" ground and eliminate the interference present on the ground of the meter circuit. If the level meter is made as a separate device, then balanced inputs with XLR connectors can be implemented.
All signal processing is done digitally by the STM32F100/103 microcontroller. To avoid the use of an anti-alias filter at the ADC input, as well as to be able to register short signal peaks, the sampling frequency was chosen quite high - 96 kHz. With DMA, ADC samples are stored in a buffer in RAM. When half of the array is ready, an interrupt occurs, the data is read and processed.
DC removal filter
The first operation on digital data is to remove the DC component from the signal. This must be done before the signal is sent to the detector. With a meter dynamic range of about 60 dB, even such a small DC component as 0.1% of the ADC scale can be equal in magnitude to the useful signal and distort it.
The simplest filter that suppresses the DC component is the differentiator. It has one zero at zero frequency. The DC gain is zero (as required), and as the frequency increases, the gain increases. For a discrete implementation of the differentiator, its difference equation has the following form:
y( n ) = x( n ) – x(n-1)
In fact, this is the simplest FIR filter. When implementing, it must be taken into account that the output value must have a width of at least 1 bit more than the input value. Since the input signal in theory in one cycle can change to the full scale. Increasing the gain with increasing frequency is not satisfactory, since the gain will change in the operating frequency range. Instead, you need to get a linear frequency response above a certain cutoff frequency. Obviously, this requires adding a pole at the desired cutoff frequency. This can be done by connecting the differentiator and the 1st order low-pass filter in series. The overall frequency response will have the desired form. Such a 1st-order low-pass filter in such applications is often called a "leaky integrator". Because in the analog version, such an integrator is implemented by adding a resistor in parallel with the capacitance, through which it will slowly discharge. For a discrete implementation, the difference equation of an ideal integrator is given below:
y( n ) = y(n-1) + x( n )
To add a leakage, you need to multiply the accumulated value by some coefficient less than one:
y( n ) = A*y(n-1) + x( n ), 0 < A < 1
This is the simplest IIR low pass filter. The name "leaky integrator" is used to indicate that this filter is applied in a somewhat unusual way - the operating frequency band lies in the stopband.
The cutoff frequency here should be chosen low compared to the sampling rate, which means that the A coefficient should be close to unity. The cutoff frequency can be calculated using the equation:
f = (1 - A)*Fs/2*pi, or A = 1 – 2*pi*f/Fs, where Fs is the sampling frequency.
The general difference equation for the filter will look like this:
y( n ) = x( n ) – x(n-1) + A*y(n-1)
In fact, this is a first-order IIR high-pass filter. It is equivalent to a differentiating RC chain. Such a filter has one zero and one pole. If you show them on the z-plane, then zero will lie at the point z=1, and a little to the left of it on the axis there will be a pole.
The audio bandwidth starts at 20 Hz. But it is desirable to make the cutoff frequency lower so that a noticeable phase shift does not appear in the audio band. This filter is not phase-linear, so signal components of different frequencies will be delayed for different times. The result of their addition will give a different waveform, the peak level will be incorrect. There are phase-linear FIR high-pass filters, but they require a lot of computing resources to obtain good linearity of the frequency response in the passband at a low cutoff frequency. There is a filter option for removing DC component based on MAF (moving average filter). It doesn't use much computational resources, but it does require a lot of memory to get a high sample rate to cutoff frequency ratio.
In this case, the requirements for the filter are not very strict. It is necessary to remove the DC voltage, which can only change very slowly (for example, due to the temperature drift of the op-amp). A significant part of the compensated bias is generally constant, as it is associated with the error of the bias voltage divider. Therefore, you can simply make the cutoff frequency lower, while minimizing the phase shift. A cutoff frequency of about 5 Hz would be a good choice. The phase shift at a frequency of 20 Hz will be about 17 degrees, which can be considered acceptable.
When implementing a filter in integer arithmetic, one must take care of the ranges of numbers at all stages of the calculation so that overflow does not occur. The leaky integrator and differentiator can be connected in any order. But to avoid overflow, the differentiator should be included first.
Another problem, more complex, is related to rounding errors. For a filter cutoff frequency of 5 Hz at a sampling rate of 96 kHz, the value of the coefficient A = 0.999673. To carry out calculations with such a coefficient with good accuracy, it is required to increase the bit depth. The reverse transition will be a quantization with a larger step, which leads to the appearance of an error, or quantization noise. As a result of such an error in the filter, a parasitic constant component may appear at the output, which is even greater than the one that is suppressed. This error may cause the filter to fail. As a practical test has shown, the filter does not actually work without requantization error correction.
In this case, the ADC is 12-bit, the samples are placed in 16-bit signed integers. Some margin is needed here to protect against overflow. Multiplication is done in 32-bit format. Then, when moving from a 32-bit intermediate result to a 16-bit result, a quantization error will occur. In one of the publications it is proposed to add a quantization noise spectrum shaper to the filter by introducing error feedback. The implementation of this method is very simple. I made my own implementation which uses the same error correction principle.
static const double POLE = 0.999673;
static const int32_t A = (int32_t)(UINT16_MAX * POLE);
static int32_t acc = 0;
static int16_t xx = 0;
static int16_t yy = 0;
x = Input;
acc = LO_W(acc) + A * yy;
yy = x - xx + HI_W(acc);
xx = x;
Output = yy;
Quantization occurs by discarding the lower half of the 32-bit number. The discarded value is the quantization error. The original number is signed, but the error is always negative. For example, if the least significant bits of a positive number are replaced with zeros, the number will decrease, and some positive number will have to be added to correct the error. If the least significant bits of a negative number are replaced by zeros, the number will become large in absolute value, while remaining negative. This means that in order to correct the error, you will again need to add some positive number. Therefore, the error can be extracted from the original 32-bit number by simply zeroing its high half along with the sign bit.
y( n ) = x( n ) – x(n-1)
In fact, this is the simplest FIR filter. When implementing, it must be taken into account that the output value must have a width of at least 1 bit more than the input value. Since the input signal in theory in one cycle can change to the full scale. Increasing the gain with increasing frequency is not satisfactory, since the gain will change in the operating frequency range. Instead, you need to get a linear frequency response above a certain cutoff frequency. Obviously, this requires adding a pole at the desired cutoff frequency. This can be done by connecting the differentiator and the 1st order low-pass filter in series. The overall frequency response will have the desired form. Such a 1st-order low-pass filter in such applications is often called a "leaky integrator". Because in the analog version, such an integrator is implemented by adding a resistor in parallel with the capacitance, through which it will slowly discharge. For a discrete implementation, the difference equation of an ideal integrator is given below:
y( n ) = y(n-1) + x( n )
To add a leakage, you need to multiply the accumulated value by some coefficient less than one:
y( n ) = A*y(n-1) + x( n ), 0 < A < 1
This is the simplest IIR low pass filter. The name "leaky integrator" is used to indicate that this filter is applied in a somewhat unusual way - the operating frequency band lies in the stopband.
The cutoff frequency here should be chosen low compared to the sampling rate, which means that the A coefficient should be close to unity. The cutoff frequency can be calculated using the equation:
f = (1 - A)*Fs/2*pi, or A = 1 – 2*pi*f/Fs, where Fs is the sampling frequency.
The general difference equation for the filter will look like this:
y( n ) = x( n ) – x(n-1) + A*y(n-1)
In fact, this is a first-order IIR high-pass filter. It is equivalent to a differentiating RC chain. Such a filter has one zero and one pole. If you show them on the z-plane, then zero will lie at the point z=1, and a little to the left of it on the axis there will be a pole.
The audio bandwidth starts at 20 Hz. But it is desirable to make the cutoff frequency lower so that a noticeable phase shift does not appear in the audio band. This filter is not phase-linear, so signal components of different frequencies will be delayed for different times. The result of their addition will give a different waveform, the peak level will be incorrect. There are phase-linear FIR high-pass filters, but they require a lot of computing resources to obtain good linearity of the frequency response in the passband at a low cutoff frequency. There is a filter option for removing DC component based on MAF (moving average filter). It doesn't use much computational resources, but it does require a lot of memory to get a high sample rate to cutoff frequency ratio.
In this case, the requirements for the filter are not very strict. It is necessary to remove the DC voltage, which can only change very slowly (for example, due to the temperature drift of the op-amp). A significant part of the compensated bias is generally constant, as it is associated with the error of the bias voltage divider. Therefore, you can simply make the cutoff frequency lower, while minimizing the phase shift. A cutoff frequency of about 5 Hz would be a good choice. The phase shift at a frequency of 20 Hz will be about 17 degrees, which can be considered acceptable.
When implementing a filter in integer arithmetic, one must take care of the ranges of numbers at all stages of the calculation so that overflow does not occur. The leaky integrator and differentiator can be connected in any order. But to avoid overflow, the differentiator should be included first.
Another problem, more complex, is related to rounding errors. For a filter cutoff frequency of 5 Hz at a sampling rate of 96 kHz, the value of the coefficient A = 0.999673. To carry out calculations with such a coefficient with good accuracy, it is required to increase the bit depth. The reverse transition will be a quantization with a larger step, which leads to the appearance of an error, or quantization noise. As a result of such an error in the filter, a parasitic constant component may appear at the output, which is even greater than the one that is suppressed. This error may cause the filter to fail. As a practical test has shown, the filter does not actually work without requantization error correction.
In this case, the ADC is 12-bit, the samples are placed in 16-bit signed integers. Some margin is needed here to protect against overflow. Multiplication is done in 32-bit format. Then, when moving from a 32-bit intermediate result to a 16-bit result, a quantization error will occur. In one of the publications it is proposed to add a quantization noise spectrum shaper to the filter by introducing error feedback. The implementation of this method is very simple. I made my own implementation which uses the same error correction principle.
static const double POLE = 0.999673;
static const int32_t A = (int32_t)(UINT16_MAX * POLE);
static int32_t acc = 0;
static int16_t xx = 0;
static int16_t yy = 0;
x = Input;
acc = LO_W(acc) + A * yy;
yy = x - xx + HI_W(acc);
xx = x;
Output = yy;
Quantization occurs by discarding the lower half of the 32-bit number. The discarded value is the quantization error. The original number is signed, but the error is always negative. For example, if the least significant bits of a positive number are replaced with zeros, the number will decrease, and some positive number will have to be added to correct the error. If the least significant bits of a negative number are replaced by zeros, the number will become large in absolute value, while remaining negative. This means that in order to correct the error, you will again need to add some positive number. Therefore, the error can be extracted from the original 32-bit number by simply zeroing its high half along with the sign bit.
Detector
Next, a full-wave rectification of the signal is performed. This is the easiest part of processing. When the DC component is removed from the signal, the task is to calculate the absolute value.
Filters
Next are several filters that define the dynamic characteristics of the meter.
The main characteristics of level meters are integration time and return time. The integration time determines how quickly the meter reacts to changes in signal level. If the integration time is large (about 300 ms), the so-called VU-meter is obtained. Such a meter will not respond to short signal peaks that can overload the recording path. When pointer instruments were used as the indication device, the integration time could not be made small due to the inertia of the moving system. Therefore, often the VU-meter was combined with a faster LED peak indicator. The use of high-speed information display devices, such as gas discharge tubes or LEDs, removed the problem of obtaining short integration times and made it possible to build peak or quasi-peak meters.
Most often, quasi-peak level meters are used to measure the signal level in audio paths. Unlike true peak meters, which in theory have zero integration time, for quasi-peak meters this time is defined by the standards. It would seem that it is necessary to achieve the minimum possible integration time so that the indicator can register the shortest signal peaks without overloading the path. This approach is used in digital paths, where even a short-term overload leads to undesirable consequences. For analog magnetic recording, a short overload may not be audible at all. If you want to completely eliminate the overload at the peaks of the signal, you will have to reduce the average recording level, which will lead to a more audible problem - a decrease in the signal-to-noise ratio. Therefore, for analog magnetic recording, it makes sense to choose some optimal value for the integration time of the level meter so that it allows short-term overloads.
The integration time for quasi-peak meters is defined by IEC 60268-10 as "...the duration of a burst of sinusoidal voltage of 5000 Hz at reference level which results in an indication 2 dB below reference indication". This standard defines an integration time value of 5 ms for quasi-peak meters.
The integration time is not numerically equal to the charging time constant of the smoothing RC circuit. If an RC circuit with a charge time constant of 5 ms is installed at the output of the peak detector, then the level of -2 dB (about 0.8) will be reached in about 20 ms. To reach -2dB in 5ms, the charging time constant of the RC circuit would need to be around 1.25ms. Thus, the integration time is approximately 4 tau RC circuit.
In addition to the integration time, quasi-peak meters have a return time that is much longer. This time determines how fast the reading will fall after the signal is removed. If this time is made small (for example, equal to the integration time), then the indicator readings will change too quickly, making them difficult to read. For VU meters, such a problem did not arise due to their low speed, one time constant could be dispensed with. In high-speed quasi-peak meters, it is necessary to artificially slow down the fall in readings so that the operator can read the information.
The return time determines how quickly the meter reading will fall by 20 dB after the signal is removed. It is also not equal to the discharge time constant of the RC circuit, but is approximately 2.3 tau. The return time value is also set by the standards. It differs for indicators of different purposes. For indicators of the first type, which serve to check the signal level during its operational adjustment (this is just the case of adjusting the recording level in a tape recorder), the return time should be 1.7±0.3 sec. Accordingly, the discharge time constant of the RC circuit should be approximately 740 ms.
Level meters have another dynamic characteristic - response time. It may not be entirely clear. For inertialess devices, ballistics is artificially formed, similar to pointer devices. Increasing the response time in this case should not increase the integration time of the meter. The short signal pulses should still display correctly. To do this, the detector must "wait" until the reading reaches the measured value, and only then start the fall. In amateur designs of quasi-peak meters, with a fast increase in the signal level, the bar immediately jumps from one length to another. This is not the case with professional quasi-peak meters, because the increase in response time has a positive effect on the perception of readings by the operator. In VU meters, this is also not the case, there the smoothness of movement is obtained due to the large integration time. The response time is defined as the interval from the moment the test tone burst begins until the reading reaches -1 dB.
Most often, quasi-peak level meters are used to measure the signal level in audio paths. Unlike true peak meters, which in theory have zero integration time, for quasi-peak meters this time is defined by the standards. It would seem that it is necessary to achieve the minimum possible integration time so that the indicator can register the shortest signal peaks without overloading the path. This approach is used in digital paths, where even a short-term overload leads to undesirable consequences. For analog magnetic recording, a short overload may not be audible at all. If you want to completely eliminate the overload at the peaks of the signal, you will have to reduce the average recording level, which will lead to a more audible problem - a decrease in the signal-to-noise ratio. Therefore, for analog magnetic recording, it makes sense to choose some optimal value for the integration time of the level meter so that it allows short-term overloads.
The integration time for quasi-peak meters is defined by IEC 60268-10 as "...the duration of a burst of sinusoidal voltage of 5000 Hz at reference level which results in an indication 2 dB below reference indication". This standard defines an integration time value of 5 ms for quasi-peak meters.
The integration time is not numerically equal to the charging time constant of the smoothing RC circuit. If an RC circuit with a charge time constant of 5 ms is installed at the output of the peak detector, then the level of -2 dB (about 0.8) will be reached in about 20 ms. To reach -2dB in 5ms, the charging time constant of the RC circuit would need to be around 1.25ms. Thus, the integration time is approximately 4 tau RC circuit.
In addition to the integration time, quasi-peak meters have a return time that is much longer. This time determines how fast the reading will fall after the signal is removed. If this time is made small (for example, equal to the integration time), then the indicator readings will change too quickly, making them difficult to read. For VU meters, such a problem did not arise due to their low speed, one time constant could be dispensed with. In high-speed quasi-peak meters, it is necessary to artificially slow down the fall in readings so that the operator can read the information.
The return time determines how quickly the meter reading will fall by 20 dB after the signal is removed. It is also not equal to the discharge time constant of the RC circuit, but is approximately 2.3 tau. The return time value is also set by the standards. It differs for indicators of different purposes. For indicators of the first type, which serve to check the signal level during its operational adjustment (this is just the case of adjusting the recording level in a tape recorder), the return time should be 1.7±0.3 sec. Accordingly, the discharge time constant of the RC circuit should be approximately 740 ms.
Level meters have another dynamic characteristic - response time. It may not be entirely clear. For inertialess devices, ballistics is artificially formed, similar to pointer devices. Increasing the response time in this case should not increase the integration time of the meter. The short signal pulses should still display correctly. To do this, the detector must "wait" until the reading reaches the measured value, and only then start the fall. In amateur designs of quasi-peak meters, with a fast increase in the signal level, the bar immediately jumps from one length to another. This is not the case with professional quasi-peak meters, because the increase in response time has a positive effect on the perception of readings by the operator. In VU meters, this is also not the case, there the smoothness of movement is obtained due to the large integration time. The response time is defined as the interval from the moment the test tone burst begins until the reading reaches -1 dB.
Integration time
To obtain a given integration time, the detector output must be smoothed with a filter. The smoothing filter is not a simple low-pass filter, but a more complex filter with different charge and discharge time constants. The charge must be fast to provide the specified meter integration time. And the discharge must occur slowly to provide a given return time, which is much longer. Therefore, the integration time filter is turned on only when the signal rises; when the signal decreases, it is kept at the same level. A separate filter is used for signal decay.
The difference filter equation has the following form:
y( n ) = y(n-1) + A*(x( n ) - y(n-1)), where A = 0.0083 for Fs = 96 kHz
A 32-bit accumulator is used to obtain the required calculation accuracy. Filter coefficients are 16-bit unsigned. They are defined as A = 65536 / 96000 / tau.
Attachments
Are you really fine with IC voltage regulators ?
- By ginetto61
- Power Supplies
- 591 Replies
Attachments
PeeCeeBee V4 GB!
- By shaan
- Group Buys
- 1099 Replies
Hello guys!
A fresh new PeeCeeBee is here, the V4! In this thread group-buys for the printed circuit boards are arranged.
It's a Lateral MOSFET based Class-AB amplifier with symmetrical "current feedback", evolved from from the diverse PeeCeeBee series, and now with excellent sonics coupled with excellent looks. 🙂
The new design is electrically and sonically a step forward from the "stubbornly" simple but very good sounding designs released in the original PeeCeeBee thread. I say a step forward partly because it incorporates some extra well-known techniques to handle some of the problem areas in the old designs and allows a quicker and safer mode of set-up and operation, but mostly because while doing so it didn't become a fatigue-inducing clinical sounding amplifier. The layout is designed by myself and the amplifier is very stable, both thermally and electrically. The amplifier has similar sonic signature to the previous peeceebees in the mid and high frequency band, but the low-frequency performance has been substantially improved so that it can be used as a full-range amplifier with effortless bass as well as beautiful midrange and highs.
It uses a simple but effective diode-clamp based soft-clipping so that occasional overloads don't bother listening enjoyment. The two jumpers, when open, remove the chance of overloading Q9/Q10 while setting-up. The constant current sources keep the input/VAS bias much less affected by power supply variation during loud playback and/or mains fluctuation.
The following two screenshots show the design:
This is how the PCB looks in real world:
A fully assembled channel:
Some specs (measured with generic components/passives and basic unregulated linear PSU):
Maximum Power - 50WRMS/90WRMS into 8R/4R load with +/-35V PSU
Maximum allowed PSU voltage for 4R operation - +/-35V (with default clamp diodes and BD BJTs for Q9/Q10) and +/-42V (with higher voltage rated clamp diodes and MJE BJTs for Q9/Q10)
Frequency response(simulated) - ~2Hz to ~800KHz (-3dB)
AC gain - 23 (27dB)
DC gain - Unity
Input level for 50WRMS into 8R - 1VRMS
THD 100Hz 10W - 0.0005%, 1KHz 10W - 0.001%, 10KHz 10W - 0.002%
Slew Rate - 100V/uS
SNR - >100dB
Offset variation - +/-10mV
Screenshot of 1KHz THD spectrum (10W into 8R resistive load, 0.002% source THD):
Screenshot of 1KHz + 5.5KHz IMD spectrum (1.7W into 8R load):
__________________________________________
Group Buy Information:
The PCBs are 92mmx64mm 2-Layer and I plan to batch-print 50, 75 or 100 PCBs depending on interest and how many people join in the GB list. The boards are made with 2.4mm FR4, 60micron copper, HASL finish, white silkscreen and black solder mask, as shown in the above pictures.
> The PCBs cost US$10 a piece.
> After the GB list is full, each user will be contacted and sent the first invoice for PCB cost, and the rest (including shipping charge, packing and processing charge and PayPal charge) is to be paid before shipping the boards, for which the second invoice will be sent.
> International payments are accepted through PayPal and payments from within India are accepted through bank account transfer/deposit, NEFT etc.
> Delay between placing the batch order to PCB plant and delivery of the boards to me will be 18-20 days.
Shipping Information:
Two forms of shipping available. International Air Parcel and Registered Post. Former takes about 15-20 days to deliver and the latter takes about 30-40 days. Both support tracking.
_____________________________________________________
GB1 (April 26 2017)
--------------------------------
Update: Pending members and anyone wishing to join, Please notice:
PM/invoices will be sent up-to 5th May.
Last day of confirming order is 7th May.
PCBs will be ordered on 8th May.
Update(May 9 2017): Total 120 boards ordered. There are still 10 spare PCBs so join in if interested in a couple.
Update(May 11 2017): GB list full. No new entries are being taken until June 1 2017.
GB1 Ended.
--------------------------------
GB2 (June 15 2017)
-------------------------------
Second group buy has started.
GB2 Ended.
-------------------------------
GB3 (Sept 12 2017)
-------------------------------
Third group buy is now open for entry. Join in if interested!
GB3 Ended
-------------------------------
GB4 (Dec 23 2017)
-------------------------------
Fourth group buy has started.
GB4 Ended.
-------------------------------
GB5 (Aug 17 2018)
-------------------------------
Fifth Group Buy arranged. Boards available.
_____________________________________________________
Important Links:
Download Latest BOM for PeeCeeBee V4 >Here<
Download Setup Guide And Component Indication >Here< (Follow this to populate and start the PCBs)
Heat Sink Drill Template >Here<
Thanks.
shaan
🙂
A fresh new PeeCeeBee is here, the V4! In this thread group-buys for the printed circuit boards are arranged.
It's a Lateral MOSFET based Class-AB amplifier with symmetrical "current feedback", evolved from from the diverse PeeCeeBee series, and now with excellent sonics coupled with excellent looks. 🙂
The new design is electrically and sonically a step forward from the "stubbornly" simple but very good sounding designs released in the original PeeCeeBee thread. I say a step forward partly because it incorporates some extra well-known techniques to handle some of the problem areas in the old designs and allows a quicker and safer mode of set-up and operation, but mostly because while doing so it didn't become a fatigue-inducing clinical sounding amplifier. The layout is designed by myself and the amplifier is very stable, both thermally and electrically. The amplifier has similar sonic signature to the previous peeceebees in the mid and high frequency band, but the low-frequency performance has been substantially improved so that it can be used as a full-range amplifier with effortless bass as well as beautiful midrange and highs.
It uses a simple but effective diode-clamp based soft-clipping so that occasional overloads don't bother listening enjoyment. The two jumpers, when open, remove the chance of overloading Q9/Q10 while setting-up. The constant current sources keep the input/VAS bias much less affected by power supply variation during loud playback and/or mains fluctuation.
The following two screenshots show the design:
This is how the PCB looks in real world:
A fully assembled channel:
Some specs (measured with generic components/passives and basic unregulated linear PSU):
Maximum Power - 50WRMS/90WRMS into 8R/4R load with +/-35V PSU
Maximum allowed PSU voltage for 4R operation - +/-35V (with default clamp diodes and BD BJTs for Q9/Q10) and +/-42V (with higher voltage rated clamp diodes and MJE BJTs for Q9/Q10)
Frequency response(simulated) - ~2Hz to ~800KHz (-3dB)
AC gain - 23 (27dB)
DC gain - Unity
Input level for 50WRMS into 8R - 1VRMS
THD 100Hz 10W - 0.0005%, 1KHz 10W - 0.001%, 10KHz 10W - 0.002%
Slew Rate - 100V/uS
SNR - >100dB
Offset variation - +/-10mV
Screenshot of 1KHz THD spectrum (10W into 8R resistive load, 0.002% source THD):
__________________________________________
Group Buy Information:
The PCBs are 92mmx64mm 2-Layer and I plan to batch-print 50, 75 or 100 PCBs depending on interest and how many people join in the GB list. The boards are made with 2.4mm FR4, 60micron copper, HASL finish, white silkscreen and black solder mask, as shown in the above pictures.
> The PCBs cost US$10 a piece.
> After the GB list is full, each user will be contacted and sent the first invoice for PCB cost, and the rest (including shipping charge, packing and processing charge and PayPal charge) is to be paid before shipping the boards, for which the second invoice will be sent.
> International payments are accepted through PayPal and payments from within India are accepted through bank account transfer/deposit, NEFT etc.
> Delay between placing the batch order to PCB plant and delivery of the boards to me will be 18-20 days.
Shipping Information:
Two forms of shipping available. International Air Parcel and Registered Post. Former takes about 15-20 days to deliver and the latter takes about 30-40 days. Both support tracking.
_____________________________________________________
GB1 (April 26 2017)
--------------------------------
Update: Pending members and anyone wishing to join, Please notice:
PM/invoices will be sent up-to 5th May.
Last day of confirming order is 7th May.
PCBs will be ordered on 8th May.
Update(May 9 2017): Total 120 boards ordered. There are still 10 spare PCBs so join in if interested in a couple.
Update(May 11 2017): GB list full. No new entries are being taken until June 1 2017.
GB1 Ended.
--------------------------------
GB2 (June 15 2017)
-------------------------------
Second group buy has started.
GB2 Ended.
-------------------------------
GB3 (Sept 12 2017)
-------------------------------
Third group buy is now open for entry. Join in if interested!
GB3 Ended
-------------------------------
GB4 (Dec 23 2017)
-------------------------------
Fourth group buy has started.
GB4 Ended.
-------------------------------
GB5 (Aug 17 2018)
-------------------------------
Fifth Group Buy arranged. Boards available.
_____________________________________________________
Important Links:
Download Latest BOM for PeeCeeBee V4 >Here<
Download Setup Guide And Component Indication >Here< (Follow this to populate and start the PCBs)
Heat Sink Drill Template >Here<
Thanks.
shaan
🙂
Attachments
For Sale FS Scak Speak Sonus Faber Cremona units in KIT
UPDATED SALE.
Here for sale some Sonus Faber Cremona units.
Some units came from my Cremonas (damaged when my son pushed the dust cap of a woofer with his finger). A pair of woofers came from SF SAT. I will explain:
2x SF Cremona Tweeters. Perfect condition. No damages. They are used and in perfect working and cosmetic condition.
Asking 100€ +pp fees + shipping for the pair
2x SF Cremona Scan Speak 18W8531/G03 Woofers from my Original Cremonas.
One was repaired because the dust cap was damaged by my son. Speakers leads changed as well. Everything was repaired in a Professional service. I can provide invoice.
They measure the same. Perfect working condition. Some minor cosmetic issues (7/10).
Asking 190€ +pp fees + shipping for the pair
2x SF Cremona Scan Speak 18W8531/G03 Woofers
This pair of woofers came from one friend from SF SAT.
The owner coated them. They measure the same than original ones.
Perfect working condition. Average looking condition due to coating (5/10)
Asking 160€ +pp fees + shipping for the pair
I can make lots at discounted price and I can include for free a pair of front baffle for woofers
Each (mm) 758H x 210W x 19D Capable for 2x Scan Speak 18W8531 woofers and a jantzen Ø100mm Bass Refflex port (Ø138,50mm hole)
Here for sale some Sonus Faber Cremona units.
Some units came from my Cremonas (damaged when my son pushed the dust cap of a woofer with his finger). A pair of woofers came from SF SAT. I will explain:
2x SF Cremona Tweeters. Perfect condition. No damages. They are used and in perfect working and cosmetic condition.
Asking 100€ +pp fees + shipping for the pair
2x SF Cremona Scan Speak 18W8531/G03 Woofers from my Original Cremonas.
One was repaired because the dust cap was damaged by my son. Speakers leads changed as well. Everything was repaired in a Professional service. I can provide invoice.
They measure the same. Perfect working condition. Some minor cosmetic issues (7/10).
Asking 190€ +pp fees + shipping for the pair
2x SF Cremona Scan Speak 18W8531/G03 Woofers
This pair of woofers came from one friend from SF SAT.
The owner coated them. They measure the same than original ones.
Perfect working condition. Average looking condition due to coating (5/10)
Asking 160€ +pp fees + shipping for the pair
I can make lots at discounted price and I can include for free a pair of front baffle for woofers
Each (mm) 758H x 210W x 19D Capable for 2x Scan Speak 18W8531 woofers and a jantzen Ø100mm Bass Refflex port (Ø138,50mm hole)
TDA1541A DAC Project Discussion
- Digital Line Level
- 7 Replies
Hi,
I am sharing the file I got from my friend, it's a TDA1541A DAC with multiple input interface including I2S.
Also, for convenience there is a InterActive BOM that helps a lot while soldering components.
The attached .Zip file contains
A) GERBER files in Zip, ready for fabrication.
B) BOM in HTML.
Please let me know if there is problems.
Regards,
J
I am sharing the file I got from my friend, it's a TDA1541A DAC with multiple input interface including I2S.
Also, for convenience there is a InterActive BOM that helps a lot while soldering components.
The attached .Zip file contains
A) GERBER files in Zip, ready for fabrication.
B) BOM in HTML.
Please let me know if there is problems.
Regards,
J
Sonos/IKEA Symfonisk hack
- By Sigrm
- Construction Tips
- 5 Replies
Hi there,
I read this article about hacking the Sonos / IKEA Symfonisk, and thought I'd try it with a nice old radio I have in my kitchen. It earns its keep as a decoration alone, but it sure would be nice to get it playing as well.
Now, the first issue that I've run in to (I'm sure there will be more), is that the radio only has one single woofer, while the Sonos looks like some sort of bi-amp setup. I'm not familiar with such. Would it be possible to wire both sets of cables to the same woofer?
From makezine.com:
I read this article about hacking the Sonos / IKEA Symfonisk, and thought I'd try it with a nice old radio I have in my kitchen. It earns its keep as a decoration alone, but it sure would be nice to get it playing as well.
Now, the first issue that I've run in to (I'm sure there will be more), is that the radio only has one single woofer, while the Sonos looks like some sort of bi-amp setup. I'm not familiar with such. Would it be possible to wire both sets of cables to the same woofer?
From makezine.com:
Project Madness
Hello all
Everyone who know me understand that I am a big Audio Physic fan, so when I see this one launched at hi end München show, I am getting really stunned... at every point. There are many things that tick boxes here
https://www.audiophysic.com/en/medeos/
This is a really complex thing to do, 5 way with 3 separate cabinets each side.... Call me absolutely insanely crazy but I am thinking, is this a project I would like to consider?, maybe 😱
It could be a setup with Bliesma beryllium tweeters and hi mids, Purify mids and woofers, several possibilities with subwoofers. But to have any chance it must be active with digital crossovers, this would easily be a €20K to €30K project, so there is pretty high risk here.... I have a working title of "project madness" 🤣
I have not been doing any simulations or calculations on cabinets sizes and/or setup yet.....
Any thoughts guys
Login to view embedded media
Everyone who know me understand that I am a big Audio Physic fan, so when I see this one launched at hi end München show, I am getting really stunned... at every point. There are many things that tick boxes here
https://www.audiophysic.com/en/medeos/
- Point source, or really close to point source, the whole passband down to around 300 Hz is managed by the central 3 drivers really close together, the other drivers symmetrically around this point.
- Lot of cone area in the mid-bass up to 300 Hz, must leave us with unreal dynamics, and there is nothing like lots of cone area
- Woofers for midbass located low and high which means it triggers different room modes and so should be more even mid-bass and less room influence with mid-bass, the same gows for the lower bass
- Integrated subwoofers, an all in one solution that makes it feasible to place even in a medium size flat (like mine)
With four large / tall towers my place would look like Stonehenge, I don´t want that.... this is also going into the main living room - Really really really cool design, the way it looks, IMHO 😍
This is a really complex thing to do, 5 way with 3 separate cabinets each side.... Call me absolutely insanely crazy but I am thinking, is this a project I would like to consider?, maybe 😱
It could be a setup with Bliesma beryllium tweeters and hi mids, Purify mids and woofers, several possibilities with subwoofers. But to have any chance it must be active with digital crossovers, this would easily be a €20K to €30K project, so there is pretty high risk here.... I have a working title of "project madness" 🤣
- Bliesma T34B beryllium tweeter
- Bliesma M74-B beryllium upper midrange
- Purify 5.25" midrange down to 300 Hz
- 4 * Purify 6.5" Midbass 100 til 300 Hz
- 8 * Purify 8" sub-bass or many other choices from Seas, Peerless, ScanSpeak, SB Acoustics....
- Marani LPP480A FIR Speaker management system (only supports up to 2 x 4 channels though)
I have not been doing any simulations or calculations on cabinets sizes and/or setup yet.....
Any thoughts guys
Login to view embedded media
Passive crossover 2 way - 4 inductors ideal layout, please advise
Hello all,
here is my DCX464 crossover layout - it contains 4 coils. The bottom part is low pass for the mid part (LCL) and the upper part has a highpass (LC) on the left and a notch filter inductor on the right. The board is now 240 x 240 mm. The C-C distance of the inductors as shown is 170 mm along the edges. The squares are other components and terminals. Can I get the board smaller by rotating the inductors better? I could follow the orientation rules for two or three, but what would be the preference when having 4 with these functions? I think I can leave the boards this large, maybe shift the smaller inductors further away a bit closer to the edges if there is nothing much to be gained from different orientation of the inductors. Any opinion is welcome.
here is my DCX464 crossover layout - it contains 4 coils. The bottom part is low pass for the mid part (LCL) and the upper part has a highpass (LC) on the left and a notch filter inductor on the right. The board is now 240 x 240 mm. The C-C distance of the inductors as shown is 170 mm along the edges. The squares are other components and terminals. Can I get the board smaller by rotating the inductors better? I could follow the orientation rules for two or three, but what would be the preference when having 4 with these functions? I think I can leave the boards this large, maybe shift the smaller inductors further away a bit closer to the edges if there is nothing much to be gained from different orientation of the inductors. Any opinion is welcome.