Ultralow distortion LM3886 chipamp
Hi everyone! It is my first message on this forum and I hope it will be interesting for you.
So, my power amp is based on LM3886. It has a very deep feedback - about 70dB at 20kHz, and much more on lower freq. It considerably reduces distortions compare to the standard schematic. I started this project almost one year ago, and now I have a very good working one. It has about -120dB THD at 11kHz, 4Ohms, 60W. IMD 19+20 is less, and I could not measure it.
Here is the schematic, parts list, PCB layout.
Some measurements.
That's it for today, I hope I'll post more soon.
Nick.
Hi everyone! It is my first message on this forum and I hope it will be interesting for you.
So, my power amp is based on LM3886. It has a very deep feedback - about 70dB at 20kHz, and much more on lower freq. It considerably reduces distortions compare to the standard schematic. I started this project almost one year ago, and now I have a very good working one. It has about -120dB THD at 11kHz, 4Ohms, 60W. IMD 19+20 is less, and I could not measure it.
Here is the schematic, parts list, PCB layout.
Some measurements.
That's it for today, I hope I'll post more soon.
Nick.
Last edited:
22kHz -120dB
33kHz -128dB
But -128dB is a noise floor, so visible harmonic is only 22kHz. I am going to improve my measurements abilities, so then I could say more precisely.
33kHz -128dB
But -128dB is a noise floor, so visible harmonic is only 22kHz. I am going to improve my measurements abilities, so then I could say more precisely.
Here are some pictures:
http://www.s-audio.com/?p=192
PCB is 4 layers. It was designed to reduce magnetic fields from output stage power lines. Power lines has a very low inductance and a lot of shunt capacitors shunt high frequencies currents as short as possible. This approach can considerably reduce the distortion at high frequencies in compare to conventional design.
http://www.s-audio.com/?p=192
PCB is 4 layers. It was designed to reduce magnetic fields from output stage power lines. Power lines has a very low inductance and a lot of shunt capacitors shunt high frequencies currents as short as possible. This approach can considerably reduce the distortion at high frequencies in compare to conventional design.
Last edited:
Contgatulations!! 
This looks like a very impressive design. Can you comment on its stability into capacitive loads?
Are PC Boards available?
This looks like a very impressive design. Can you comment on its stability into capacitive loads?
Are PC Boards available?
Thanks!al2002 said:Contgatulations!!
With an output filter (inductunce ~1uH + resistor 10Ohm) it has no problem to drive 1uF capacitor, with or without actual load.
Not yet. I have to test the speaker protection circuit, make some minor changes, and then....
Last edited:
Thanks for the BOM. I look forward to the final version and the PC Board layouts.
A couple of questions, if I may.
Is the 28 V LM 1084 regulated supply essential? Just wondering how much difference it makes to the final measured performance if U3 is just fed directly from an unregulated supply.
If a truly low-Z output signal from the preceding preamp stage is available to feed the U2+U3 composite amp, is the U1 buffer really needed?
Thanks in advance.
By the way your private e-mail is disabled.
A couple of questions, if I may.
Is the 28 V LM 1084 regulated supply essential? Just wondering how much difference it makes to the final measured performance if U3 is just fed directly from an unregulated supply.
If a truly low-Z output signal from the preceding preamp stage is available to feed the U2+U3 composite amp, is the U1 buffer really needed?
Thanks in advance.
By the way your private e-mail is disabled.
There is practically no difference in measurable perfomance. It was just done in order to make power lines clean as possible. Here is -28V oscillogram (LM3886 output is 20Vp-p on low inductive 4ohms). http://www.s-audio.com/forum/download/file.php?id=11&mode=viewal2002 said:
No, it will be optional. But if you have a noisy DAC it is better to install it, because it is a filter also. Actually U1(AD8065) have no measurable distortion (< -130dB).
I don't know why, may be because I am a "newbee" here. My email -
An externally hosted image should be here but it was not working when we last tested it.
Even if you enable your email button, your email address is hidden (spamproof). You can answer an email in two ways: 1 Via your email klient (and then reveal the address) or via diyaudio and keep the email hidden).Mykola said:
Nice work btw
You haven't thought of using Joomla for your homepage? Check mine for an example.
It also has -47 degrees of phase shift at 10kHz. Response is down 1.4dB at 10kHz and 4.2dB at 20kHz
No, you can see it's frequency response here. As you can see it mostly depends on the input filter parameters, and even with C3=220pF it has no more than -0.5dB at 20KHz. Probably C3=150pF could be a better value, it makes about -0.15dB at 20kHz. On the other hand, this filter is optional, this amp works fine without it.jackinnj said:
Always welcome 🙂.al2002 said:
I expect in the end of this month, but can't say for sure.
Actually a few Russians guys (from www.vegalab.ru) reproduced this amp, but they made their own PCBs. As result theó have a little more THD on high frequencies (the best result was -110dB at 10KHz).
I believe my e-mail button is not enabled because I am under moderation yet (as new guy).peranders said:
Joomla is quite easy to get going and there are lot's of nice looking templates. You have also a lot of people that use Joomla.
http://www.siteground.com/joomla-hosting/joomla15-templates.htm
http://www.siteground.com/tutorials/joomla15/index.htm
You should not be moderated by now.
http://www.siteground.com/joomla-hosting/joomla15-templates.htm
http://www.siteground.com/tutorials/joomla15/index.htm
You should not be moderated by now.
Interesting idea you have but personally I'm a bit skeptical because real life issues may surprise you. Feedback loops that you have may be a lot more complicated than they seem to be.Mykola said:
Re: Re: My ZD-50 ultralow distortion chipamp.
The first one was build to verify the idea. The second one was build to find out the potential perfomance. And it actually has about 0.0004% THD at 10KHz, 50W, and much less at lower frequencies. IMD 19+20 is less than 0.0002%. I still listen this prototype, for more than 6 months.
Also there were several listening competitions, so I am sure that's a really good amp, not just a Spice model 🙂
I believe you could find a lot of interesting information about this amp in my thread at Russian forum , but unfortunately it is on Russan language and the thread is really huge (more than 200 pages).
You are not the first 🙂 But it's not just a simulation. Each part was optimized in real life. As I wrote above, it is a 3-rd prototype.peranders said:
The first one was build to verify the idea. The second one was build to find out the potential perfomance. And it actually has about 0.0004% THD at 10KHz, 50W, and much less at lower frequencies. IMD 19+20 is less than 0.0002%. I still listen this prototype, for more than 6 months.
Also there were several listening competitions, so I am sure that's a really good amp, not just a Spice model 🙂
I believe you could find a lot of interesting information about this amp in my thread at Russian forum , but unfortunately it is on Russan language and the thread is really huge (more than 200 pages).