My ZD-50 ultralow distortion chipamp.

Nickolay Shvydky · 2008-08-05 9:27 am

Here is one of the ideas that was used in this amp.

peranders · 2008-08-05 10:08 am

Maybe you should test your ideas with LME49810/830 also :idea:

jackinnj · 2008-08-05 2:05 pm

Try examining this amplifier with a 10kHz square wave. What's the slew rate at full power?

The SPICE models for the AD8065 and LT1363 are available on the manufacturer's websites:

http://www.analog.com/en/amplifiers...mod-spice-models/resources.html?display=popup

http://www.linear.com/pc/downloadDocument.do?navId=H0,C1,C1154,C1009,C1022,P1108,D11759

Nickolay Shvydky · 2008-08-05 7:19 pm

jackinnj said:
Try examining this amplifier with a 10kHz square wave. What's the slew rate at full power?

Slew rate is limited by LM3886, and typically >15V/uS. Here is 20kHz square wave test.

The SPICE models for the AD8065 and LT1363 are available on the manufacturer's websites:

http://www.analog.com/en/amplifiers...mod-spice-models/resources.html?display=popup

http://www.linear.com/pc/downloadDocument.do?navId=H0,C1,C1154,C1009,C1022,P1108,D11759

Thank you. Actually LTSpice already contains the LT1363 model, and I used it.

Nickolay Shvydky · 2008-08-05 7:23 pm

peranders said:
Maybe you should test your ideas with LME49810/830 also

Yeah, its definetly possible.

Nickolay Shvydky · 2008-08-06 12:11 am

Here are some more oscillograms.

juma · 2008-08-06 1:36 am

Great work Nikolai ! :up:

Can you tell more about listening tests?

Nickolay Shvydky · 2008-08-07 5:58 am

juma said:
Great work Nikolai !

Thanks!

Can you tell more about listening tests?

I will hardly be able to do the justice to the amp in few sentences.
We have had extensive listening sessions where we compared the amp with Creek 4330MK2 and
Creek 5350SE hooked up to Revel M22 monitors and Audio Physic Spark loudspeakers.
The most noticable difference is in the bass, which is tighter and better controlled.
and in the soundstage, which is more solid and precise but, suprisingly, less expansive.
The high are grain free. Overall, the amp is very dynamic and musical.

KSTR · 2008-08-07 7:23 pm

juma said:
Great work Nikolai !

I'd like to second this.
And thanks for sharing the link to the Yewen composite amp article (which I haven't seen before), which really is essential understanding your amp (at first I couldn't quite disassemble all those multiple feeback paths, now it has become clear, a very clever design idea).

- Klaus

Max Volobuev · 2008-08-13 10:27 pm

I've made this amp on my own 2-layes PCB a couple weeks ago. Here are THD and IMD measurements. The right channel on the graph is the amp input, the left is the amp output.
So, what about the sound? Very transparent and clean, with good bass performance. Transfer of a vocal is very natural. I’ve compared this amp with my previous favorite amp http://astral-for.narod.ru/index/stones/amgen.gif. ZD-50 is just better at any type of music. My other components are 3-way homemade speakers (magnesium Seas Excel drivers) and AD1862 based DAC.

Actually, I think if 50W is enough for you, it will be very hard to find a better amplifier.

Here are some pictures
http://fotoplenka.ru/users/antecom/422197/

jackinnj · 2008-08-14 12:50 pm

Mbe I'm missing something here -- here's a phase-gain plot of the input stage --

An externally hosted image should be here but it was not working when we last tested it.

x-pro · 2008-08-14 1:08 pm

jackinnj said:
Mbe I'm missing something here -- here's a phase-gain plot of the input stage

That part is just an input LP filter and buffer. I guess that the simulated plots from the first post done without this LPF.

Cheers

Alex

jackinnj · 2008-08-14 1:20 pm

x-pro said:
That part is just an input LP filter and buffer. I guess that the simulated plots from the first post done without this LPF.

Cheers

Alex

Why use a 145MHz FastFET opamp in this application?

x-pro · 2008-08-14 1:27 pm

jackinnj said:
Why use a 145MHz FastFET opamp in this application?

You need a fast opamp in an LPF if you want it to work properly up to fairly high frequencies (i.e. CD-player output can contain some components in MHz range) . AD8065 used probably because it has very low THD.

Cheers

Alex

Nickolay Shvydky · 2008-08-14 9:08 pm

jackinnj said:
Mbe I'm missing something here -- here's a phase-gain plot of the input stage --

Personally I do not see anything wrong with this frequency response (-3dB at 75kHz). If you think you need a more flat response, just reduce the values of C2 and C3, or do not install the input filter (then the input is R7).
I already posted the amp frequency and phase response vs C3 graph http://www.diyaudio.com/forums/showthread.php?postid=1577734#post1577734

Nickolay Shvydky · 2008-08-14 9:22 pm

x-pro said:
You need a fast opamp in an LPF if you want it to work properly up to fairly high frequencies (i.e. CD-player output can contain some components in MHz range) . AD8065 used probably because it has very low THD.

Cheers

Alex

Hi Alex and welcome to my thread

!
Yes, because the AD8065 has a very low THD, even at very high frequencies. So I see no difference in the perfomance with or without it, except the SNR. Also the AD8065 has a very low input offset voltage.
Actually the U2 can be AD8065, LM318. The U1 I think can be ADA4898-1, but I did not test it yet...

Nickolay Shvydky · 2008-08-14 9:27 pm

x-pro said:
That part is just an input LP filter and buffer. I guess that the simulated plots from the first post done without this LPF.

Cheers

Alex

Alex, you mean the feedback depth plots

?

lumumba_p · 2008-08-15 4:03 am

Mykola,
Thank you very much for this excellent design!!! I have built a simplified version – one of 6 blocks that I need for my 3-way system. I could not believe my measurement results! Even with LM318 THD were lower than I can reliably measure with Juli@ (20W, 8 Ohm, 10kHz), that is below 0.0003%.
I also tested operation of the amplifier in Current Source mode. It works great!
I am not sure how to implement a power supply for 6 channels properly. Any ideas?

jackinnj · 2008-08-15 12:05 pm

lumumba_p said:
Mykola,
Thank you very much for this excellent design!!! I have built a simplified version – one of 6 blocks that I need for my 3-way system. I could not believe my measurement results! Even with LM318 THD were lower than I can reliably measure with Juli@ (20W, 8 Ohm, 10kHz), that is below 0.0003%.
I also tested operation of the amplifier in Current Source mode. It works great!
I am not sure how to implement a power supply for 6 channels properly. Any ideas?

Of course the THD% is lower, the amplifier is bandpass limited.

x-pro · 2008-08-15 1:05 pm

jackinnj said:
Of course the THD% is lower, the amplifier is bandpass limited.

There is no connection, unless the amplifier is band limited on the output, and that is not the case here.

Cheers

Alex

My ZD-50 ultralow distortion chipamp.

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member

Member