Greetings from Manchester...kind of
- By NdG1977
- Introductions
- 2 Replies
Hello everybody!
43 years old Mechanical Engineer decent to good DIYer and woodworker, decent electronics soldering skills almost inexistent competence of electronics circuitry (Recipe for disaster right?
) ...Italian living close to Manchester in the UK. Past experience in amateur car audio, not much experience in home HiFi.
I'm planning to modernize my system and I was quite sure of what I wanted (audio wise), then I spent weeks and weeks of lockdown reading this forum...Now I changed my mind a 100 times and I'm not sure anymore...
I'll be asking a lot of questions about setup and design!
Bye!
43 years old Mechanical Engineer decent to good DIYer and woodworker, decent electronics soldering skills almost inexistent competence of electronics circuitry (Recipe for disaster right?
) ...Italian living close to Manchester in the UK. Past experience in amateur car audio, not much experience in home HiFi.I'm planning to modernize my system and I was quite sure of what I wanted (audio wise), then I spent weeks and weeks of lockdown reading this forum...Now I changed my mind a 100 times and I'm not sure anymore...
I'll be asking a lot of questions about setup and design!
Bye!
What Causes Off-Axis Nulls?
- Multi-Way
- 32 Replies
On another thread, we were discussing the Behringer B2031P. It segued into a discussion about vertical nulls, and I thought it might be interesting to illustrate why they happen.
90% of you already know how this works, but I like making diagrams to help myself understand things. So here goes...
This is a Behringer B2031P. It has an 8" woofer, a 1" tweeter, and the vertical spacing between woofer and tweeter is about 8". (20cm.)
Using XDir, we can see that this geometry is going to lead to some deep nulls at 30 degrees off axis. Why is that?
Here's why.
When listening on-axis, we are equidistant from the woofer and tweeter. At 30 degrees off axis (vertically) there's a big null. The null occurs because there's a pathlength difference of 10cm between the woofer and the tweeter. 1700Hz is 20cm long; therefore a pathlength difference of 10cm creates a big fat null, because the two are 180 degrees out-of-phase.
Hope that makes sense!
If so, let's talk about beaming:
Here's the polar response from a typical 8" woofer. This is a Tymphany P22WO03.
At 30 degree of axis, at 1700Hz, the woofer is only "down" about 1dB from it's on axis response. This is because an 8" woofer has a diaphragm that measures about 6.5" in diameter. So the woofer isn't even beaming very much until it reaches about 2khz. (2khz is 6.5" long.)
Nulls occur when two sources are out-of-phase. If two sources are playing 90dB and they're perfectly out-of-phase, you get an output of 0dB. If one source is playing 90dB and the other is playing 89dB, the results won't be much different. There will still be a null, it just won't be as deep.
This is a situation where a waveguide can come in handy, because a carefully optimized waveguide can play about -6dB off axis. Basically it's a lot easier to reduce the off-axis energy of the *tweeter* not the woofer. The beaming of the woofer is largely determined by it's diameter.
In a situation like we see with the Behringer pictured here, there's a few options:
1) lower the crossover point. This is easier said than done; there aren't many 1" tweeters that can handle a xover point of 1khz.
2) The Behringer is largely a clone of a Genelec design from a couple decades back. The newer Genelec models simply moved the tweeter closer to the woofer. This works nicely.
In the measured vertical response of the Behringer, we can see that they simply lived with the off-axis null. It is very prominent at 1700Hz. Thanks to dantheman for the measurements : Great Waveguide List
90% of you already know how this works, but I like making diagrams to help myself understand things. So here goes...
This is a Behringer B2031P. It has an 8" woofer, a 1" tweeter, and the vertical spacing between woofer and tweeter is about 8". (20cm.)
Using XDir, we can see that this geometry is going to lead to some deep nulls at 30 degrees off axis. Why is that?
Here's why.
When listening on-axis, we are equidistant from the woofer and tweeter. At 30 degrees off axis (vertically) there's a big null. The null occurs because there's a pathlength difference of 10cm between the woofer and the tweeter. 1700Hz is 20cm long; therefore a pathlength difference of 10cm creates a big fat null, because the two are 180 degrees out-of-phase.
Hope that makes sense!
If so, let's talk about beaming:
Here's the polar response from a typical 8" woofer. This is a Tymphany P22WO03.
At 30 degree of axis, at 1700Hz, the woofer is only "down" about 1dB from it's on axis response. This is because an 8" woofer has a diaphragm that measures about 6.5" in diameter. So the woofer isn't even beaming very much until it reaches about 2khz. (2khz is 6.5" long.)
Nulls occur when two sources are out-of-phase. If two sources are playing 90dB and they're perfectly out-of-phase, you get an output of 0dB. If one source is playing 90dB and the other is playing 89dB, the results won't be much different. There will still be a null, it just won't be as deep.
This is a situation where a waveguide can come in handy, because a carefully optimized waveguide can play about -6dB off axis. Basically it's a lot easier to reduce the off-axis energy of the *tweeter* not the woofer. The beaming of the woofer is largely determined by it's diameter.
In a situation like we see with the Behringer pictured here, there's a few options:
1) lower the crossover point. This is easier said than done; there aren't many 1" tweeters that can handle a xover point of 1khz.
2) The Behringer is largely a clone of a Genelec design from a couple decades back. The newer Genelec models simply moved the tweeter closer to the woofer. This works nicely.
In the measured vertical response of the Behringer, we can see that they simply lived with the off-axis null. It is very prominent at 1700Hz. Thanks to dantheman for the measurements : Great Waveguide List
Change jbl synthesis sdp-40hd to get great sound.
- By Rinat
- Digital Source
- 0 Replies
Some time ago I decided to improve the sound quality of home theater - I really like opera. An analysis of what is on the market showed that the focus is on auto tuning, a huge number of channels, but not on sound. In general, I decided to make my jbl synthesis sdp-40hd (analogue of Lexicon MC-12HD) work as it should.
At the start, I was guided by the older brother of the Lexicon - Mark Levinson No.40, which is also on ad1853 in mono, a linear power supply and a price tag of 30k.
An external power supply unit for audio and video circuits has been designed and manufactured. For audio, two blocks on the m5230l (the Cref wima MKS 3.3μF capacitor, the rest in the Elna RFS strapping, in front of the microcircuits, decoupled from the main cans of Nichicon KZ), +/-5V and +/-16V. Separate +/-8V power supply for 7808/7908 video. In the main body, the pulsed universal power supply unit was replaced with a low-noise 5V from Traco, which supplies only digital circuits. From the external power supply, voltage is supplied to the processor through two cables. One by one, the voltage +/- 5V and +/- 16V are connected to the audio board. Otherwise, +/-8V is supplied to the video card and the inclusion of an external unit from the main device is controlled.
Removed power pull-up to I / U microcircuits, opa2134 replaced with ad823a, subwoofer with ad8512. The resistor in I / U is replaced with RQ73 3.01kΩ 0.1% 10ppm ("tantalum"). The tantalum capacitors in the ad1853 power supply are digitally replaced with x7r 10μF, the analogue of Elna RFS 100μF is decoupled from the 2.2ohm power buses, on the Nichicon KZ 47μF subwoofers. On “filtr” I installed wima MKS 4.7μF, supplemented with 1.5ohm. Ceramic y5v 0.1 μF capacitors in the ad1853 strapping is replaced by npo 0.047 μF (there is no more than 0805), a similar replacement in the power supply of the volume control. The filter m33078 is replaced by opa1656, by opa1662 subwoofers. Film capacitors MKP2 2,2nF are replaced by wima FKP2. Opa1656 inverted the left front and AUX channels at the input (I will explain later). I reconnected the channels - now the right channel goes to both front ones, the left channel goes to the both AUX channels. Removed analog signal switches as unnecessary.
Volume control microcircuits, except for the subwoofers, were replaced by PGA2311A, and the adjustment from the fronts to AUX was also thrown. In the power supply of Elna RFS 22μF - they are decoupled from the power rails by 1.2Ω. At the output op275 are replaced with LT1469. Removed all decoupling capacitors - one high-quality film at the input to the amplifier is enough.
I put Elna RFS and ROA on the power rails (I picked up the capacitance ratio by ear, I like the sound of old Silmic s, the latter don't sound like that, and adding ROA makes the desired sound) and removed tantalum capacitors scattered throughout the board - replaced with Elna RFS of small capacities Here is a general view of the board.
Thus, a true balanced signal goes to the XLR. At the same time, 2 ad1853, their own opamp and volume control with compensation for its noise work for each front channel. Also on the board of balanced outputs on the front channels lm833 was replaced with opa1656 and the balanced signal driver was removed as unnecessary. The output resistors have so far delivered what they were - the right ones on the road. The alteration was very long - delivery from the right places and every change was listened to and measured, it's good when there are 2 identical channels - auditory memory is not a reliable thing. First, AUX made it identical to the front one - I made changes only in it and compared it.
Here is the result of measurements of the modified E-MU1616 (without M) at the RCA outputs of the processor. One channel is measured better, this is the difference between the E-MU channels, the processor channels work the same. The output of the processor is identical. I do not quote the results of measurements on XLR, they are close, only there is no interference from the network at all, there is less noise, and distortion is 1 dB higher - the signal amplitude is high, you have to suppress the regulator - the dynamic range is reduced.
Volume control + 2dB, generator -2dB 2V rms.
IMD.
Volume control -10dB.
Here are the measurements of a number of processors / receivers conducted by ASR with much better hardware.
An additional task was to reduce the temperature in the processor case, which heated up over 45 degrees, which is due to the fact that the power supply for video and +5V audio analog was obtained by converting +/-16V pulsed power supply, which itself was very hot and, according to information from others owners were often out of order. Now the heating is about 37 degrees, which made it possible to install delicate audio capacitors.
As a result, I got a great-sounding processor that sounds exciting. I apologize for my bad English. If something was not clear, I am ready to answer your questions.
At the start, I was guided by the older brother of the Lexicon - Mark Levinson No.40, which is also on ad1853 in mono, a linear power supply and a price tag of 30k.
An external power supply unit for audio and video circuits has been designed and manufactured. For audio, two blocks on the m5230l (the Cref wima MKS 3.3μF capacitor, the rest in the Elna RFS strapping, in front of the microcircuits, decoupled from the main cans of Nichicon KZ), +/-5V and +/-16V. Separate +/-8V power supply for 7808/7908 video. In the main body, the pulsed universal power supply unit was replaced with a low-noise 5V from Traco, which supplies only digital circuits. From the external power supply, voltage is supplied to the processor through two cables. One by one, the voltage +/- 5V and +/- 16V are connected to the audio board. Otherwise, +/-8V is supplied to the video card and the inclusion of an external unit from the main device is controlled.
Removed power pull-up to I / U microcircuits, opa2134 replaced with ad823a, subwoofer with ad8512. The resistor in I / U is replaced with RQ73 3.01kΩ 0.1% 10ppm ("tantalum"). The tantalum capacitors in the ad1853 power supply are digitally replaced with x7r 10μF, the analogue of Elna RFS 100μF is decoupled from the 2.2ohm power buses, on the Nichicon KZ 47μF subwoofers. On “filtr” I installed wima MKS 4.7μF, supplemented with 1.5ohm. Ceramic y5v 0.1 μF capacitors in the ad1853 strapping is replaced by npo 0.047 μF (there is no more than 0805), a similar replacement in the power supply of the volume control. The filter m33078 is replaced by opa1656, by opa1662 subwoofers. Film capacitors MKP2 2,2nF are replaced by wima FKP2. Opa1656 inverted the left front and AUX channels at the input (I will explain later). I reconnected the channels - now the right channel goes to both front ones, the left channel goes to the both AUX channels. Removed analog signal switches as unnecessary.
Volume control microcircuits, except for the subwoofers, were replaced by PGA2311A, and the adjustment from the fronts to AUX was also thrown. In the power supply of Elna RFS 22μF - they are decoupled from the power rails by 1.2Ω. At the output op275 are replaced with LT1469. Removed all decoupling capacitors - one high-quality film at the input to the amplifier is enough.
I put Elna RFS and ROA on the power rails (I picked up the capacitance ratio by ear, I like the sound of old Silmic s, the latter don't sound like that, and adding ROA makes the desired sound) and removed tantalum capacitors scattered throughout the board - replaced with Elna RFS of small capacities Here is a general view of the board.
Thus, a true balanced signal goes to the XLR. At the same time, 2 ad1853, their own opamp and volume control with compensation for its noise work for each front channel. Also on the board of balanced outputs on the front channels lm833 was replaced with opa1656 and the balanced signal driver was removed as unnecessary. The output resistors have so far delivered what they were - the right ones on the road. The alteration was very long - delivery from the right places and every change was listened to and measured, it's good when there are 2 identical channels - auditory memory is not a reliable thing. First, AUX made it identical to the front one - I made changes only in it and compared it.
Here is the result of measurements of the modified E-MU1616 (without M) at the RCA outputs of the processor. One channel is measured better, this is the difference between the E-MU channels, the processor channels work the same. The output of the processor is identical. I do not quote the results of measurements on XLR, they are close, only there is no interference from the network at all, there is less noise, and distortion is 1 dB higher - the signal amplitude is high, you have to suppress the regulator - the dynamic range is reduced.
Volume control + 2dB, generator -2dB 2V rms.
IMD.
Volume control -10dB.
Here are the measurements of a number of processors / receivers conducted by ASR with much better hardware.
An additional task was to reduce the temperature in the processor case, which heated up over 45 degrees, which is due to the fact that the power supply for video and +5V audio analog was obtained by converting +/-16V pulsed power supply, which itself was very hot and, according to information from others owners were often out of order. Now the heating is about 37 degrees, which made it possible to install delicate audio capacitors.
As a result, I got a great-sounding processor that sounds exciting. I apologize for my bad English. If something was not clear, I am ready to answer your questions.
Anybody familiar with this transformer?
- By soldersmoker
- Solid State
- 6 Replies
I believe it's a power transformer for solid state audio, and I assume it's ISO-compliant, not from the Iso transformer company.
My Google-fu fails me on IDing this one.
Thanks!
Type 45 Parafeed Rebuild Complete
- By L0rdGwyn
- Tubes / Valves
- 20 Replies
Last year, I designed and built a 6J5-45 parafeed headphone amplifier, one of my first tube DIY projects. In a lot of ways it was a speaker amplifier adapted for headphones with a cold bias point. Over time I regretted not building it for speakers with a switchable headphone output.
Well yesterday I rectified that (no pun intended) and rebuilt the amplifier for speakers with a new mains transformer to get the full 1.5W out of the 45 😎
This will live in a small bedroom setup with a pair of Klipsch KG 1, so 1.5W is plenty for that space.
Anyway, I like to share photos of my completed projects, so here are some pics and a general schematic (final component values may be slightly different). Nothing fancy, a simple design, but the sound is quite nice!
Thanks for checking it out.
Well yesterday I rectified that (no pun intended) and rebuilt the amplifier for speakers with a new mains transformer to get the full 1.5W out of the 45 😎
This will live in a small bedroom setup with a pair of Klipsch KG 1, so 1.5W is plenty for that space.
Anyway, I like to share photos of my completed projects, so here are some pics and a general schematic (final component values may be slightly different). Nothing fancy, a simple design, but the sound is quite nice!
Thanks for checking it out.
Feedback: Two pole vs. output inclusive ?
- By mikebeth
- Solid State
- 31 Replies
Happy Holidays all! 
Does anyone have any experience with different advanced solid-state amplifier feedback methods - specifically two pole compensation vs. output inclusive compensation (as D.Self refers to it)?
I am currently evaluating the performance of having added output inclusive compensation (switchable via jumper: conventional or output inclusive) to my latest amplifier PCB build (ver3) and am quite impressed so far at how it reduces crossover distortion. I had considered designing the amp with 3 selectable feedback methods (conventional, output inclusive or two pole), but got lazy during the PCB layout and left out the two-pole option. I judged that the output inclusive compensation would more effectively combat crossover distortion, as other distortions in the previous PCB (ver 2.1) had been pushed below 0.001%. Now I wish I had included two-pole as an option for evaluation purposes.
I'd be interested to hear about/discuss/share anyone's findings in this design area. Currently I'm determining the optimal bias setting for both compensation methods and then will begin gathering performance graphs, etc.
Mike
Does anyone have any experience with different advanced solid-state amplifier feedback methods - specifically two pole compensation vs. output inclusive compensation (as D.Self refers to it)?
I am currently evaluating the performance of having added output inclusive compensation (switchable via jumper: conventional or output inclusive) to my latest amplifier PCB build (ver3) and am quite impressed so far at how it reduces crossover distortion. I had considered designing the amp with 3 selectable feedback methods (conventional, output inclusive or two pole), but got lazy during the PCB layout and left out the two-pole option. I judged that the output inclusive compensation would more effectively combat crossover distortion, as other distortions in the previous PCB (ver 2.1) had been pushed below 0.001%. Now I wish I had included two-pole as an option for evaluation purposes.
I'd be interested to hear about/discuss/share anyone's findings in this design area. Currently I'm determining the optimal bias setting for both compensation methods and then will begin gathering performance graphs, etc.
Mike
