It is important to understand that Sakis is in the repair business and that means by definition that he has a lot of components at hand. He knows that sooner or later he will use the rest of his stock. But when ranchu buys a bag of transistors to select closely matched pair he will have a bag of redundant transistors. We want to be responsible towards resources and do not want to leave a lot of electronic waste behind even if it means that our amp will be less than ideal. It is not only money problem...
True, and he also has the experience to know what will make a difference, and what is a waste of time... the rest of us are often just guessing. My P3A sounds pretty good with no exotic or expensive parts, minimal effort and no waste. I wish I could say the same about some of my other projects.It is important to understand that Sakis is in the repair business and that means by definition that he has a lot of components at hand. He knows that sooner or later he will use the rest of his stock. But when ranchu buys a bag of transistors to select closely matched pair he will have a bag of redundant transistors. We want to be responsible towards resources and do not want to leave a lot of electronic waste behind even if it means that our amp will be less than ideal. It is not only money problem...
I still haven't got around to taking final photos of inside my P3A - the reason being I've enjoyed listening to it in my living room too much!
I twisted every flow/return AC line and also the +/- DC power lines as suggested by Ian and Andrew. I also shortened and tidied up the other cables as practically possible. No noticable differences found but then none expected.
After some critical listening I did encounter something amiss with the treble, present only on some female vocalist tracks. Its difficult to describe: a "harshness" or distortion perhaps. After some research and investigation I realised that while all the ground lines were taken back to the filter bank capacitor bank, I had not connected the ground plane to chassis/earth ground! I have since remedied that and the distortion seems to have disappeared. Is it plausable that this could have been caused by improper earthing?
In the past couple weeks I've started working on the P88 preamp. I'm starting to wonder whether a preamp is really necessary. I seem to have a huge amount of gain in the computer source and this amplifier, so much so I have to turn the software volume control WAY down to achieve sensible listening levels. Perhaps a simple passive attenuator and source selector would suffice?
I twisted every flow/return AC line and also the +/- DC power lines as suggested by Ian and Andrew. I also shortened and tidied up the other cables as practically possible. No noticable differences found but then none expected.
After some critical listening I did encounter something amiss with the treble, present only on some female vocalist tracks. Its difficult to describe: a "harshness" or distortion perhaps. After some research and investigation I realised that while all the ground lines were taken back to the filter bank capacitor bank, I had not connected the ground plane to chassis/earth ground! I have since remedied that and the distortion seems to have disappeared. Is it plausable that this could have been caused by improper earthing?
In the past couple weeks I've started working on the P88 preamp. I'm starting to wonder whether a preamp is really necessary. I seem to have a huge amount of gain in the computer source and this amplifier, so much so I have to turn the software volume control WAY down to achieve sensible listening levels. Perhaps a simple passive attenuator and source selector would suffice?
Attachments
Hi
All these are valid points not to match the transistors.
However, i have high expectations for this amp and wish it to better the DOZ that i build point to point with matched parts.
BC546 is very cheap at my place (around US 20 cents ) and the shopowner allows me to bring my DMM and match it at the shop!
BTW DOZ is a fantastic amp and this is my first class AB build.All my previous amps are class A.
cheers
kp93300
All these are valid points not to match the transistors.
However, i have high expectations for this amp and wish it to better the DOZ that i build point to point with matched parts.
BC546 is very cheap at my place (around US 20 cents ) and the shopowner allows me to bring my DMM and match it at the shop!
BTW DOZ is a fantastic amp and this is my first class AB build.All my previous amps are class A.
cheers
kp93300
If you use PC as the source, you do not need P88 or any other preamp (including passive attenuator). But using software controls is not that elegant and you may decide to use more sources in the future, so having a preamp is not a bad idea. But, when I built P88 it was difficult to set the volume for low level listening even with only 6dB gain (input opamp was configured as unity gain buffer, output was set for 6dB gain). For modern sources 27dB gain in P3A is plenty!
I also have a CD player that I'd like to use - but can't until I have some means of adjusting volume control. The signal level from the CD player is very high, similar to the computer source. I'll set up the P88 with the minimum 6db gain as you say and if it proves difficult to fine-tune the volume levels I'll go down the passive attenuator route instead.
Kinda makes you wonder why people bother with active preamps if all the sources are modern with high gains.
Kinda makes you wonder why people bother with active preamps if all the sources are modern with high gains.
@ Ranchu: If you find yourself turning the volume on the preamp way down for normal listening levels, then almost NO preamp will give the best results.
Most people simply ignore this in the pursuit of ever more pure, and more exotic gear to put between the source and the power amp.
Preamps are a relic of days when phono inputs, tape inputs, receivers etc. all came from different equipment, and had different signal levels, sound signatures, and sometimes R-L balance. They are still very critical when this is the case, and in professional installations. And not to be neglected, fun to play with, so don't take this as a dismissal of a good topic...
Most of the noise added in a typical preamp is fixed amplitude, so a small signal comes out with more noise. On top of that, a small signal is more susceptible to noise of all kinds, magnetic, rf, rectifier switching noise, ground loops, etc. When you amplify that signal in the power amp, you amplify ALL that noise also.
So if all you have are modern digital sources, or high-output analog sources, a conventional preamp is counterproductive. This is not the fault of the pre at all... (with prior apologies for a bad paraphrase), The fault lies not in our gear, but in ourselves.
There is an article on this forum that explains it much better than I ever could:
Gain Structure
While I agree for the most with you, there is one relevant exception, that I personally tested several times with the original P3A.
That are MP3 players, smartphones and co. Their signal, when sourced from the "headphone" outputs (only a few devices have "line" outputs) is simply way too low and needs to be amplified of 2-4 times in voltage. That is NOT an impedance mismatch proble, as it is often mistaken for. The output impedance of those devices is very low (suitable to drive (8 tto 32 ohm headsets) so it does not pose problems with the P3A. It is just the signal level that is not enough to bring the output of the P3A on the brink of the clipping even at the maximum settings, considering its gain of 22.
I think you are correct, although I have not actually measured that on the P3A w. an MP3 player.
The problem with typical digital sources, like a typical PC output, is that adjusting the volume from the desktop reduces the digital signal in amplitude before the D/A conversion, so the number of useful bits used for the signal is also reduced. Given that most material is not actually mastered at 16-bits but something less, you don't get the best possible sound.
As long as most people insist on high-gain power amplifiers, and high-gain preamps, there will not be a good universal solution, but almost anyone can find a solution that will work well for their particular circumstances.
If P88 is built with balance regulator there will be additional 3dB loss, so 6dB in the second stage will be 3dB at the output. But I did not use balance regulator.
Another problem is volume pot value. With high gain amps and preamps it is usual to see 100-200k pots in Japanese products. With P88 value is just 10k so you have pretty loud signal at low settings.
Rod's P3A is really clever circuit. Everything is optimized for easy build and no matching of components. For instance resistance that inverting and noninverting inputs "see" are almost "balanced". Noninverting: 22k+1k+1k=24k. Inverting: 22k+1k=23k. Therefore there should be reasonably low offset even without matching of the the input transistors. This method is sometimes used with opamps to ensure that the DC offset is low without output coupling caps.
Another problem is volume pot value. With high gain amps and preamps it is usual to see 100-200k pots in Japanese products. With P88 value is just 10k so you have pretty loud signal at low settings.
Rod's P3A is really clever circuit. Everything is optimized for easy build and no matching of components. For instance resistance that inverting and noninverting inputs "see" are almost "balanced". Noninverting: 22k+1k+1k=24k. Inverting: 22k+1k=23k. Therefore there should be reasonably low offset even without matching of the the input transistors. This method is sometimes used with opamps to ensure that the DC offset is low without output coupling caps.
With reference to my earlier post, apparently Sakis DOES match channel-to-channel, as I see from a recent post in his P3A comparison thread:
http://www.diyaudio.com/forums/solid-state/164756-p3a-comparsion-table-long-21.html#post3572013
Personally, I do not see much point in this for listening with normal speakers, because most speakers, even very good ones, have a great deal of variation between them. (Sad truth is that the speakers you buy are usually not as perfectly matched and balanced as those sent to magazines and reviewers for evaluation.)
The things that seemed to make the biggest difference in my case was 1) a matched transistor pair in the LTP, 2) a better/cleaner power supply, in combination with bigger rail elcos on the boards (I split my p3A board in two halves, which proved to be unnecessary). I am sure that a better input coupling cap also helped, but I did not initially test w. good bass speakers, so I can't say I tested that.
Note that Sakis is a professional, and his P3A to ours is like a track car to a daily driver - its been modded and tuned "just a bit"...
http://www.diyaudio.com/forums/solid-state/164756-p3a-comparsion-table-long-21.html#post3572013
Personally, I do not see much point in this for listening with normal speakers, because most speakers, even very good ones, have a great deal of variation between them. (Sad truth is that the speakers you buy are usually not as perfectly matched and balanced as those sent to magazines and reviewers for evaluation.)
The things that seemed to make the biggest difference in my case was 1) a matched transistor pair in the LTP, 2) a better/cleaner power supply, in combination with bigger rail elcos on the boards (I split my p3A board in two halves, which proved to be unnecessary). I am sure that a better input coupling cap also helped, but I did not initially test w. good bass speakers, so I can't say I tested that.
Note that Sakis is a professional, and his P3A to ours is like a track car to a daily driver - its been modded and tuned "just a bit"...
There are speakers that are matched to within 0,5dBs for a pair but these are usually for studio use or (basically) studio speakers used for home listening (like Harbeth) but if one has so much money to buy them he usually has enough money to buy excellent factory made amp. I do not understand Lazy Cat's recommendation for loudspeakers to be used with VSSA amp. If you have enough money to buy ZuAudio speakers that he recommends than you should have enough money to buy decent factory made current feedback amp like, say, Krell S300i.
P3A is for diy-ers, it is not meant to be used with disproportionally expensive speakers, although it is probably better than many factory made amps. So I think that matching channel-to channel is really not for diy-ers.
P3A is for diy-ers, it is not meant to be used with disproportionally expensive speakers, although it is probably better than many factory made amps. So I think that matching channel-to channel is really not for diy-ers.
Finished P3A-- sound impressions
I have been listerning to this amp for about 1 week.
I use easily available parts Elna silmic 10 Uf input cap, panasonic FC 220Uf on board power supply caps , BC 546, 556, BD 139, 140 and 2sa5200 and 2sc1943. The transistors are matched.
I biased it at 75 ma.
My speakers are easy load 91db/w open baffle with powered subwoofer--V2 from GR Research. The source is Monica step ladder dac from diy paradise with Luckit usb to i2s converter.
My previous amp is a breadboard DOZ, dc coupled ---Shaan version.
I am pleasantly surprised at the sound . Overall impression is clean, clear, dynamic , tremendous control of the speakers.
The high frequency is a joy to listen. Bells has very nice ring and beautiful decay . Plucked string instrument such as the Chinese pipa has a nice bite with nice resonance from the box.
Mid is slightly lean but this is counterbalance with the MooseFet pre from Classic valve design . With this pre, mid range detail is amazing. opening of lips from close mic recordings are easy to hear . Singer is slightly in front of the musicians and well separated . Sarah Brightman's voice is seductive !
The bass is very tuneful and deep bass is well controlled. With big drums, the stick hitting the skin of the drums is well resolved. This is followed by the echoes in the recording room.
The set plays rock, classical and vocal equally well.
This amp changed my perception about class Ab amp. My previous amps are F5, alephJ, JLH69, and numerous Class T amp from 41 Hz..
cheers
kp93300
I have been listerning to this amp for about 1 week.
I use easily available parts Elna silmic 10 Uf input cap, panasonic FC 220Uf on board power supply caps , BC 546, 556, BD 139, 140 and 2sa5200 and 2sc1943. The transistors are matched.
I biased it at 75 ma.
My speakers are easy load 91db/w open baffle with powered subwoofer--V2 from GR Research. The source is Monica step ladder dac from diy paradise with Luckit usb to i2s converter.
My previous amp is a breadboard DOZ, dc coupled ---Shaan version.
I am pleasantly surprised at the sound . Overall impression is clean, clear, dynamic , tremendous control of the speakers.
The high frequency is a joy to listen. Bells has very nice ring and beautiful decay . Plucked string instrument such as the Chinese pipa has a nice bite with nice resonance from the box.
Mid is slightly lean but this is counterbalance with the MooseFet pre from Classic valve design . With this pre, mid range detail is amazing. opening of lips from close mic recordings are easy to hear . Singer is slightly in front of the musicians and well separated . Sarah Brightman's voice is seductive !
The bass is very tuneful and deep bass is well controlled. With big drums, the stick hitting the skin of the drums is well resolved. This is followed by the echoes in the recording room.
The set plays rock, classical and vocal equally well.
This amp changed my perception about class Ab amp. My previous amps are F5, alephJ, JLH69, and numerous Class T amp from 41 Hz..
cheers
kp93300
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.