Hello,
I have built the P3A several years ago, but for my taste, the sound is too clinical. I understand it is built for precision, and it does that very well, but it lacks the "warmth" that usually comes tweaked for branded amps. This warmth, as far as I understand, comes from an emphasis on bass and highs, with a decreased representation of the mid end, making it sound less-harsh and less-tiresome. ( like a V shaped equalization pattern ). So, I was thinking of making some adjustments, and with the help of Chat GPT, I identified some possible mods that I can make, but I would love it if anyone here had the time to validate if these can cause other issues with the amp and if these will actually affect the sound signature towards "warmth" :
low end :
high end :
It was also suggested to increase R13 and R14 to 0.47 or even 0.56 to tone down the vocals
Another thing suggested is to use a combination of ceramic and metal film resistors, rather than full metal film, even though it will slightly affect precision it will also slightly introduce a tiny bit of pleasant distorsion. Also to use electrolytics instead of Polyprop for the same reason.
What of these advices do you think are feasible ?
Thanks in advance !
I have built the P3A several years ago, but for my taste, the sound is too clinical. I understand it is built for precision, and it does that very well, but it lacks the "warmth" that usually comes tweaked for branded amps. This warmth, as far as I understand, comes from an emphasis on bass and highs, with a decreased representation of the mid end, making it sound less-harsh and less-tiresome. ( like a V shaped equalization pattern ). So, I was thinking of making some adjustments, and with the help of Chat GPT, I identified some possible mods that I can make, but I would love it if anyone here had the time to validate if these can cause other issues with the amp and if these will actually affect the sound signature towards "warmth" :
low end :
- increase C1 from 4.7 to 10 or even 22
- increase C2 from 100 to 220pf or even 470pf
- increase C3 from 100 to 220 or even 470
- increase C5 from 100 to 220 or even 470
- reduce R6 from 560 to 470 ( will this affect the recommended bias value ? )
- C+, C- increased from 100 to 470 or even 1000
high end :
- increase C4 from 100 to 150pf or even 220pf
- reduce R11 and R12 from 220 to 180 ( same question, will this affect recommended bias ? )
- add 2.2nF capacitor in parallel with R5
It was also suggested to increase R13 and R14 to 0.47 or even 0.56 to tone down the vocals
Another thing suggested is to use a combination of ceramic and metal film resistors, rather than full metal film, even though it will slightly affect precision it will also slightly introduce a tiny bit of pleasant distorsion. Also to use electrolytics instead of Polyprop for the same reason.
What of these advices do you think are feasible ?
Thanks in advance !
IMHO P3A is the less clinical amp among many I've built.
It is very much on the warm and mellow side.
What is your (or Chat GPT 🤣) reference branded amp?
It is very much on the warm and mellow side.
What is your (or Chat GPT 🤣) reference branded amp?
Some of those changes will have an audible effect if you're an elephant or a bat. Some make no sense. I have probed Chat GPT about electronics a few times, just for fun. Aside from general stuff you can find in Wikipedia, it has absolutely no idea. Please ignore.
Rather than messing with the P3A design, I think your best bet would be to put a 2 or 3-band tone control circuit in front of it and adjust to taste.
Rather than messing with the P3A design, I think your best bet would be to put a 2 or 3-band tone control circuit in front of it and adjust to taste.
Currently it's fed directly by a Cambridge Audio DAC but I also plan to hook it to a P97 preamp stage with this rebuild, which will give me some tone control, but I would love to also make it naturally have a more curvy response, hence the modifications. I am not debating if it's everyone's opinion, just asking if the modifications proposed would help in my goal without bringing in any risks to the circuit, and more importantly, which of them would have the highest impact.
Some modifications proposed by chat gpt will affect stability and may cause the amp to fail. Don't listen to chat gpt if you don't know what you are doing.
Maybe if ChatGPT knows it better let ChatGPT listen to it and do something else yourself instead?
Of all amplifier designs P3A is a thoroughly tested and proven one that is practically always fine. Does not need any change.
Of all amplifier designs P3A is a thoroughly tested and proven one that is practically always fine. Does not need any change.
ridiculous and pointless input filter.........change c2 to 2n2
try some ltp degen 100-200R and reduce c4 accordingly
try some ltp degen 100-200R and reduce c4 accordingly
increase C1 from 4.7 to 10 or even 22 - This will lower the low frequency F-3 from ~1.5 Hz to ~0.75 Hz or ~0.375 Hz (elephant frequencies)
increase C2 from 100 to 220pf or even 470pf - This will lower the high frequency F-3 from ~1.6 MHz to ~800 kHz or ~400 kHz (bat frequencies)
increase C3 from 100 to 220 or even 470 - Pretty much the same as C1
increase C5 from 100 to 220 or even 470 - No effect on frequency response
reduce R6 from 560 to 470 - No effect on frequency response, will throw off DC offset
C+, C- increased from 100 to 470 or even 1000 - No effect on frequency response, only ripple
increase C4 from 100 to 150pf or even 220pf - Reduce open loop gain by ~3.5 dB or ~6 dB, increased distortion across the band
reduce R11 and R12 from 220 to 180 - No effect on frequency response, will require quiescent current readjustment
add 2.2nF capacitor in parallel with R5 - I haven't checked the P3A specifically, but in a similar CFP output stage amp with the same 22k / 1k feedback, this reduces the phase margin from ~75º to ~3º, turning the amp into a very powerful oscillator
increase C2 from 100 to 220pf or even 470pf - This will lower the high frequency F-3 from ~1.6 MHz to ~800 kHz or ~400 kHz (bat frequencies)
increase C3 from 100 to 220 or even 470 - Pretty much the same as C1
increase C5 from 100 to 220 or even 470 - No effect on frequency response
reduce R6 from 560 to 470 - No effect on frequency response, will throw off DC offset
C+, C- increased from 100 to 470 or even 1000 - No effect on frequency response, only ripple
increase C4 from 100 to 150pf or even 220pf - Reduce open loop gain by ~3.5 dB or ~6 dB, increased distortion across the band
reduce R11 and R12 from 220 to 180 - No effect on frequency response, will require quiescent current readjustment
add 2.2nF capacitor in parallel with R5 - I haven't checked the P3A specifically, but in a similar CFP output stage amp with the same 22k / 1k feedback, this reduces the phase margin from ~75º to ~3º, turning the amp into a very powerful oscillator
Option 1 is the single largest thread generator of the tube section of DIYaudio.com. Absolutely worst advice possible. Sorry.
Apparently it is missed that no one needs a tube preamp with a gain of 100 when connecting a 2V rms source of the last 20 years to a semiconductor power amplifier that has normal 1V input sensitivity.
If I had the money I would pay for a stroboscopic flashing very annoying continuously popping up advertisement banner here to NOT connect randomly chosen tube preamps with random parameters to standardised semi amplifiers.
Apparently it is missed that no one needs a tube preamp with a gain of 100 when connecting a 2V rms source of the last 20 years to a semiconductor power amplifier that has normal 1V input sensitivity.
If I had the money I would pay for a stroboscopic flashing very annoying continuously popping up advertisement banner here to NOT connect randomly chosen tube preamps with random parameters to standardised semi amplifiers.
Last edited:
If anyone can provide a definition of that “warm” sound, please let me know. I once created a simple little DSP application that allowed me to add a variable amount of second harmonic to a signal. And whatever it is, the “warm” sound ain’t that. It sounded anything but warm.
Thank you very much sir, that is what I was looking for, thank you for your time. I will try building the P97 preamp and adding it to the signal path, and if I still don't like it I will try to search the internet for a different DIY amplifier schematic with more of a warm, vintage sound.
To njswede: to me ( and apparently others online ) , it looks like a certain level of V-shape curve rather than flat frequency response, ephasizing deep bass more, then gradually dropping to 0 db at around 350 hz, keeping a linear response up to around 3-4khz, when the curve goes up again in the highs.
Just an example of the general idea I found online fast, but with a wider 0 db line at the middle as explained before
Similar to what the "loudness" switch of many amps does actually, but the purpose there is to compensate for these frequencies at low volume levels, however the general idea is similar ( and no wonder on one of my older Sony amps that had a loudness switch I would just permanently keep it on regardless of volume level because the sound is much better imo rather than flat and "loud", tiresome )
I’ve always called that “the disco smile”. 😀
If that’s what you want, build yourself a set of Rod’s robe controls and crank them up to max. Done.
If that’s what you want, build yourself a set of Rod’s robe controls and crank them up to max. Done.
...sound signature towards "warmth"...
Consider getting different speakers or testing the ones now with different gear. The speakers probably color the sound more than the amplifier. Then consider the music source. Streamers such as Spotify and Tidal modify the sound in some way. Recordings from the previous millennium are remastered and upverted. Modern recordings are sometimes done by the artists themselves who are not recording engineers. I would prefer to get CDs.
Consider getting different speakers or testing the ones now with different gear. The speakers probably color the sound more than the amplifier. Then consider the music source. Streamers such as Spotify and Tidal modify the sound in some way. Recordings from the previous millennium are remastered and upverted. Modern recordings are sometimes done by the artists themselves who are not recording engineers. I would prefer to get CDs.
Numerous threads can be found regarding P3A modifications and such, just hit the search button. P3A is one amp that is very sensitive to changes, you may get it to work but it won't be for long. P3A when built right is totally stable with a very good SQ.
If you want to make some changes do not listen to ChatGPT, run a simulation instead.
Albert
If you want to make some changes do not listen to ChatGPT, run a simulation instead.
Albert