Seems they ship to Europe only.
Maybe Gianluca can talk to them (same language) and maybe get some and ship to US?
The ebay store is this one:
https://www.ebay.de/str/movtech?_trksid=p4429486.m3561.l2563
Could not find a direct web to the company.......
Maybe Gianluca can talk to them (same language) and maybe get some and ship to US?
The ebay store is this one:
https://www.ebay.de/str/movtech?_trksid=p4429486.m3561.l2563
Could not find a direct web to the company.......
Hello All,
It was raining buckets so we went to the SoCal desert for a couple of weeks, see the attached photo. Still away from the bench.
I was recalling a TI.com tech note about a method for measuring low device noise that includes the DUT in the amplifier test circuit. See the attached link below.
https://www.ti.com/lit/ug/tidu016/t...rm%3Dnoise+measurement+amplifier%26nr%3D31348
For the P3 the bulk of the noise that is referred to the input originates in input stage that includes the 4 parallel JFETS. While we have OPA1612's or OPA1656's on the bench we can fabricate the second stage of the noise test circuit with 100 times gain / 40dB gain. The input stage has gain of it's own for about 1000 times / 60dB total gain. Typical of noise measurement circuits.
I can input a known signal voltage and the analyzer will calculate the gain.
The analyzer will output the measured circuit noise and divide by the known gain and calculate the noise referenced to the input.
Thanks DT
It was raining buckets so we went to the SoCal desert for a couple of weeks, see the attached photo. Still away from the bench.
I was recalling a TI.com tech note about a method for measuring low device noise that includes the DUT in the amplifier test circuit. See the attached link below.
https://www.ti.com/lit/ug/tidu016/t...rm%3Dnoise+measurement+amplifier%26nr%3D31348
For the P3 the bulk of the noise that is referred to the input originates in input stage that includes the 4 parallel JFETS. While we have OPA1612's or OPA1656's on the bench we can fabricate the second stage of the noise test circuit with 100 times gain / 40dB gain. The input stage has gain of it's own for about 1000 times / 60dB total gain. Typical of noise measurement circuits.
I can input a known signal voltage and the analyzer will calculate the gain.
The analyzer will output the measured circuit noise and divide by the known gain and calculate the noise referenced to the input.
Thanks DT
When measuring noise, the way the math works to get a meaningful answer the measurement amplifier's input noise should be 1/SQRT(2) the expected noise of the device under-test. You can get there with parallel (4x or 8x) ADA4898's and cross-correlation!
But it's about the music! As Dolby pointed out, the ear is most sensitive to noise in the vicinity of 5kHz to 8kHz. We are a lot more irritated about hum which Pearl 3 seems to deal with quite effectively. (I will order boards when they are available on the DIYAUDIOSTORE.)
But it's about the music! As Dolby pointed out, the ear is most sensitive to noise in the vicinity of 5kHz to 8kHz. We are a lot more irritated about hum which Pearl 3 seems to deal with quite effectively. (I will order boards when they are available on the DIYAUDIOSTORE.)
Dear Wahne,
what operationalamplifier do you suggest for the "Pearl 3". I prefer the old NE5532, maybe two single OP-Amp. - because the are in my stock.
Noise:
On regular base, i hear the noise of the entire circuit with an headphone, my ears tell me whats going on. My eyes looking on a good (normally analog) Scope, playing with the volume (turn it up to maximum), and with the setup /e.g. the settings of my oscilloscope. Normally that's my way to quicktest the development.
what operationalamplifier do you suggest for the "Pearl 3". I prefer the old NE5532, maybe two single OP-Amp. - because the are in my stock.
Noise:
On regular base, i hear the noise of the entire circuit with an headphone, my ears tell me whats going on. My eyes looking on a good (normally analog) Scope, playing with the volume (turn it up to maximum), and with the setup /e.g. the settings of my oscilloscope. Normally that's my way to quicktest the development.
Last edited:
In the video presentation Wayne mentioned that 5532 sounds very good in Pearl 3.
Video available here - https://burningampfestival.com/videos/
Video available here - https://burningampfestival.com/videos/
Take a closer look at the Pearl 3's theoretically timeconstants:
T3 = (R13||R28) * C3
= 73,822µs
fg3 is now 2155,93Hz
Can we ignore the load (R28 -jXC12) or the dynamic outputresistor, the source U1A - sourceresistor? No.
T1 = (R17+R14) * C1
= 3,182ms
fg1 is now 50,02Hz
Can we ignore the other component in the feedbacknet? Perhaps?
T2 = [(R17+R14) || (R15+R18)] * C1
= 318,56µs
fg2 is now 499,61Hz
Neglecting the influence of frequencydepand C8||C25 and the finite OL-Gain, the GBP, of the selected OP-Amp.
It's only an overview for clarification and for minor interest. The complete circuit fit the requirements and sound accordingly the twostage topologie.
T3 = (R13||R28) * C3
= 73,822µs
fg3 is now 2155,93Hz
Can we ignore the load (R28 -jXC12) or the dynamic outputresistor, the source U1A - sourceresistor? No.
T1 = (R17+R14) * C1
= 3,182ms
fg1 is now 50,02Hz
Can we ignore the other component in the feedbacknet? Perhaps?
T2 = [(R17+R14) || (R15+R18)] * C1
= 318,56µs
fg2 is now 499,61Hz
Neglecting the influence of frequencydepand C8||C25 and the finite OL-Gain, the GBP, of the selected OP-Amp.
It's only an overview for clarification and for minor interest. The complete circuit fit the requirements and sound accordingly the twostage topologie.
That amaze me not
.The very old NE5534 fits perfectly for allInOne-EQ-networks. I am sadly looking back in these times -> like the PMI-Device SSM 2134 is available on the market.
HBt.
(please excuse my very bad written english, i am better in hearing and reading - must be exercise myself better)
Last edited:
On the measured Pearl 3 I should mention R28 was 100K. It makes a tiny difference on the EQ. The good old 5534 and 5532 have made many great recordings.
HBT your English is total understandable. My German is about four words. My German speaking grandfather unfortunately died just before I was born.
HBT your English is total understandable. My German is about four words. My German speaking grandfather unfortunately died just before I was born.
Who is the enthusiastic guy, dance in front of the screen, seen in the Video?
By the way, it's a big mistake to belive that the currentsink Q9 can pull really the interna (OutputStage) of U1:B into operation known as Class A. It is only an additional load against negativ rail of the powersupply. Nothing less, nothing more.
The Video doesn't explain the reason for choosing split RIAA and especially not the network in feedbackloop. Otherwise (on the other hand) there are a million opportunitys for run alternative networks - all activ realisation.
I miss the explanation /more details. The adjustment of the overallgain /done with the devider on common node R4...7, R8 and R9, R22 is not special.
To play with the value of C1 and C12 is not an innovation ...
A little bit - lack of crucial difference between of other DIY-EQ.
Please don't misunderstood my thoughts and questions; i thing it is allowed to play with the offer, to force at its best. Pearl 3 must defend successfully against all others, Pearl 1 and Pearl 2 ... compete at his /her best.
By the way, it's a big mistake to belive that the currentsink Q9 can pull really the interna (OutputStage) of U1:B into operation known as Class A. It is only an additional load against negativ rail of the powersupply. Nothing less, nothing more.
The Video doesn't explain the reason for choosing split RIAA and especially not the network in feedbackloop. Otherwise (on the other hand) there are a million opportunitys for run alternative networks - all activ realisation.
I miss the explanation /more details. The adjustment of the overallgain /done with the devider on common node R4...7, R8 and R9, R22 is not special.
To play with the value of C1 and C12 is not an innovation ...
A little bit - lack of crucial difference between of other DIY-EQ.
Please don't misunderstood my thoughts and questions; i thing it is allowed to play with the offer, to force at its best. Pearl 3 must defend successfully against all others, Pearl 1 and Pearl 2 ... compete at his /her best.
Me.
I believe it’s much better than the previous Pearls. It will be interesting to hear other’s opinions.
You missed the part in the video when I pointed out that the output of U1:B sits at +.9V As forced by the closed loop. The load is on the positive rail. It doesn’t act like a bias potentiometer in a power amp on both output transistors just increased current in one.
Last edited by a moderator:
Has just been shipped (100 pcs):
Assume I get genuine ZTX851. Can see from his ebay shop that he sells a number of older vintage semiconductors. Germanium stuff also......
Ok.....price a bit higher than from "normal" providers. But in low numbers ZTX851 is not the cheapest......anyway.
i guess I get plenty now for one P3 build. I can match them very closely
Assume I get genuine ZTX851. Can see from his ebay shop that he sells a number of older vintage semiconductors. Germanium stuff also......
Ok.....price a bit higher than from "normal" providers. But in low numbers ZTX851 is not the cheapest......anyway.
i guess I get plenty now for one P3 build. I can match them very closely
Sorry Wahne,
i don't miss the point. It is obvious that Pin 7 of IC U1 must be on >0,7V. Nothing change the internal structure of the Operational Amplifier. The load, as seen from the output of U1:B is set (lying) between node Pin7 and node -15 (negativ supply rail) ... I know and understand the internal circuit exactly, for example the NE5534.
The complete resulting thing around Q7 and Q6 is , putting in the feedback, only change the capability to delivery more current at node PAD4. Typically for red led and a standing dc-current of 10mA set the biasvoltage to 1,6V ...
That's my point of view, a little bit simplifed but not wrong. I hope, that no one get disapoint now. The addition of "Gegentaktendstufe" /Q7 adnd Q6 is generally a good idea.
Don't forget (in mind), we don't drive heavy load at /with the output of an ordinary PhonoPreAmplifier.
friendly
HBt.
"I believe it’s much better than the previous Pearls. It will be interesting to hear other’s opinions." [quote 6L6]
No opinions - hard facts.
The use of integrated OP-Amps change the game decisive. Now is the underlying, main philosophy of the previous one (Pearl 1 and 2) gone /lost.
Where is the fascination of the twostage cascode, in between the passive equalization network. In the past the simplicity of the original was fascinating, and the typical singleended distortion footprint makes the sound - the behavior of deal with a little bit (temporally) to high signals (in German: "Übersteuerung").
The advantage of Pearl 2 was the Outputstage. To often the toshiba dual jFet on the scecond stage (of original Pearl) died, the only help was put in two independently resistors - from jFetSource to ground, change the value and add a crucial change to the devider at base of upper bjt ... take care of don't plug in or out when the powersupply is on. The discrete OP-Amp. of "Pearl 2" with singleended output mos-fet enhanced and save the handling, provide controlling and setup independly the gain, makes widely /largely indepently from load -> but is not a singleended cascode design. But, a big improvement from viewpoint. Pearl 2 change the typically soundstage. Perl 3 will change them one more time, but dramatic. Now it's no longer a design with the glory of discrete diy building without integrated circuit. It is much more ordinary, and must be compete against all other nicly and well build EQs based on OP-Amp. This is a fact.
This point of view disturb me, but not bother me to much - now is just a Competion. Thats why "Pearl 3" should be measure really perfectly.
RIAA-EQ is not a complicated thing, but every thing counts - thats make the development difficult. Powersupply is the next issue.
Hope you enjoy the fruit of your approach, it's totally understandable and justified. Who is the developer of Pearl 3, Wahne plus '6L6'?
friendly
HBt.
(short for Helge with my surname)
No opinions - hard facts
.The use of integrated OP-Amps change the game decisive. Now is the underlying, main philosophy of the previous one (Pearl 1 and 2) gone /lost.
Where is the fascination of the twostage cascode, in between the passive equalization network. In the past the simplicity of the original was fascinating, and the typical singleended distortion footprint makes the sound - the behavior of deal with a little bit (temporally) to high signals (in German: "Übersteuerung").
The advantage of Pearl 2 was the Outputstage. To often the toshiba dual jFet on the scecond stage (of original Pearl) died, the only help was put in two independently resistors - from jFetSource to ground, change the value and add a crucial change to the devider at base of upper bjt ... take care of don't plug in or out when the powersupply is on. The discrete OP-Amp. of "Pearl 2" with singleended output mos-fet enhanced and save the handling, provide controlling and setup independly the gain, makes widely /largely indepently from load -> but is not a singleended cascode design. But, a big improvement from viewpoint. Pearl 2 change the typically soundstage. Perl 3 will change them one more time, but dramatic. Now it's no longer a design with the glory of discrete diy building without integrated circuit. It is much more ordinary, and must be compete against all other nicly and well build EQs based on OP-Amp. This is a fact.
This point of view disturb me, but not bother me to much - now is just a Competion. Thats why "Pearl 3" should be measure really perfectly.
RIAA-EQ is not a complicated thing, but every thing counts - thats make the development difficult. Powersupply is the next issue.
Hope you enjoy the fruit of your approach, it's totally understandable and justified. Who is the developer of Pearl 3, Wahne plus '6L6'?
friendly
HBt.
(short for Helge with my surname)
Last edited:
Addition to my Post /Number 574 plus 575:
Compare it carefully with and without the additional PowerStage on the end. It's not a single and quick look, it takes time and measurment.
"(...) It doesn’t act like a bias potentiometer in a power amp on both output transistors just increased current in one." [quote Wahne]
I know it. But your reply does not cancel my argument.
The currentsink pull and bias (a tiny) the internal currentlimiting bjt, just a fraction of a trigger and depands on choice of what OP is installed - that's not a really good thing ... sometimes things get worse.
Compare it carefully with and without the additional PowerStage on the end. It's not a single and quick look, it takes time and measurment.
"(...) It doesn’t act like a bias potentiometer in a power amp on both output transistors just increased current in one." [quote Wahne]
I know it. But your reply does not cancel my argument.
The currentsink pull and bias (a tiny) the internal currentlimiting bjt, just a fraction of a trigger and depands on choice of what OP is installed - that's not a really good thing ... sometimes things get worse.
Last edited:
Matching Q5?
"(...) I can match them very closely" [quote MEPER]
Is not importend -for left and right channel balance, except your are looking for the very best Noisefigur in the bag. Honestly there is no need for doing so. Take a closer look at the jFets or increase R4,5,6,7 a little bit (with consequences!). Matching take place here, selecting the 2SK170 or your favorite jFet.
With no harm you can change Q5 to old Siemens BC550C or B-type, in the case of a Pearl 1,2,3 -> but there are so many fakes on the market, the doesn't fit and working well in lownoiseapplication with Vcc higher than 5Vdc (that's my observation and exasperating /sadly experience). Since that desaster working on a development, i measure all BC5xx-Stuff and selecting very carefully ...
"(...) I can match them very closely" [quote MEPER]
Is not importend -for left and right channel balance, except your are looking for the very best Noisefigur in the bag
. Honestly there is no need for doing so. Take a closer look at the jFets or increase R4,5,6,7 a little bit (with consequences!). Matching take place here, selecting the 2SK170 or your favorite jFet.With no harm you can change Q5 to old Siemens BC550C or B-type, in the case of a Pearl 1,2,3 -> but there are so many fakes on the market, the doesn't fit and working well in lownoiseapplication with Vcc higher than 5Vdc (that's my observation and exasperating /sadly experience). Since that desaster working on a development, i measure all BC5xx-Stuff and selecting very carefully ...
Now when I get 100 ZTX851 (+10 later in December) I will probably use the Mega328 tester to test if the 851 works and find 4 pcs that has relative close parameters. A Mega328 (and similar testers) writes out a few parameters (not a close matching at all but better than "nothing" and easy and fast to do ).
).