ES9038Q2M Board

Attachments

  • Front3.jpg
    Front3.jpg
    1,016.5 KB · Views: 385
You could probably get away with +-9v. The opamp or opamps need to be operated within their datasheet voltage limits, and the if using the dac onboard regulators to drop the + opamp rail down to +3.3v for the digital circuitry, then those regulators need to be operated within their datasheet dropout voltage specifications (they are likely to regulate better if not operated at the very lowest input voltage possible, it can help to have a few volts across them). Some similar dac boards dropped the +15v power down to +5v or +8v (see what it says on the first regulator), then from there down to 3.3v. If they are LDOs then the dropout voltages will be lower than if they are conventional regulators.

Regarding +- voltage rails for opamps, the voltage can affect sound quality. Usually higher tends to be better, but for AK4499 dacs using OPA1612 I/V opamps it turned out that +-11v sounded cleanest. Only why to find out is to try some different opamp rail voltage settings, then listen and or measure distortion for each voltage level.

As far as a recommendation goes, I probably wouldn't want to go below +-11v even though it would probably still work at, say, +-9v. Also, please use a good linear power supply for that, not SMPS or wall warts of any kind to start out with because its more involved to clean up the power coming from them. Better not to have dirty power in the first place, if possible.

For transformers, R-core are best for minimizing ground conducted noise from the AC line.
 
Last edited:
Daniel,
That sounds like a plan. Please let us know what you find, if anything.

Also, if it were me I would want to improve the sound of that board. One of the steps in that process involves separating the opamp power from the digital power. Your +-12v supply would probably be great for the opamp circuitry. On the other hand, a low noise +9v supply might be good for feeding the local 3.3v regulators that provide stable power for the clock, DVCC, and VCCA. That leaves AVCC, which is a tough one to get get right. We usually prefer to power it with opamp buffers and a low noise +3.3v reference. Its one of the first and most important things to talk about in more detail if you decide you want to improve dac sound quality.
 
Thanks,
Yes i was thinking about try to seperate the powerlines.
Is it that easy to cut the trace from the 7808 and feed something like 9v there?
I also have some two leftover 3,3v low noise regulator.
Maybe i can use them? But not sure where. I will use optical in so maybe some improvment there?
Also, in post #353 there was something about current mode?
Is it that easy that doing like in that post?

Many questions, sorry.

Edit: my Swedish spelling...

BR // Daniel
 
Thanks,
Yes i was thinking about try to seperate the powerlines.
Is it that easy to cut the trace from the 7808 and feed something like 9v there?

Yes, something like that. Best to keep wires from the power supplies to the dac board pretty short.

I also have some two leftover 3,3v low noise regulator.
Maybe i can use them?

I used three 3.3v low noise regulators mounted on the back side of the dac board (ground plane side). One regulator for DVCC, one for VCCA, and one for the clock. The MCU can stay on the original 3.3v regulator that's already on the board.

I will use optical in so maybe some improvment there?

Optical tends to be pretty jittery. Best to try to use good quality optical cable and keep it short, if possible.

Also, in post #353 there was something about current mode?
Is it that easy that doing like in that post?

The dac board comes with a one opamp voltage mode output stage design. It is *much* higher distortion than a proper 3-opamp current mode output stage design. The output stage and the AVCC power supply are two of the very biggest sound quality improvements that can be made. However, its some diy work to make it happen. Schematics for recommended output stage and AVCC supply are attached to post #3003.

A through hole output stage project with construction pictures is available at: Dropbox - Output Stage Instructions.zip - Simplify your life
 
There you had one person saying his opinion based on his measurements. I didn't see any phase noise plots for different cables, so I remain unconvinced of his claims. And, yes I am on the side that believes phase noise at low offset frequencies matters a lot. I can hear it and so can other people who bother to listen. Others differ, as we all know.
 
Yes, something like that. Best to keep wires from the power supplies to the dac board pretty short.



I used three 3.3v low noise regulators mounted on the back side of the dac board (ground plane side). One regulator for DVCC, one for VCCA, and one for the clock. The MCU can stay on the original 3.3v regulator that's already on the board.



Optical tends to be pretty jittery. Best to try to use good quality optical cable and keep it short, if possible.



The dac board comes with a one opamp voltage mode output stage design. It is *much* higher distortion than a proper 3-opamp current mode output stage design. The output stage and the AVCC power supply are two of the very biggest sound quality improvements that can be made. However, its some diy work to make it happen. Schematics for recommended output stage and AVCC supply are attached to post #3003.

A through hole output stage project with construction pictures is available at: Dropbox - Output Stage Instructions.zip - Simplify your life

Thanks for the helpful info. I will try to use my regulators to DVCC and clock.
Regarding the current output ##3003, it looks a little bit complicated for me :)

BR// Daniel
 
Regarding the output stage, I agree its its not the easiest thing to build. It probably would have been considerably easier if using SMD components, but some people didn't think they would be able to take on soldering things that small. Could be there is some 3rd party output stage that could be adapted, but can't think of one offhand.
 
Regarding the output stage, I agree its its not the easiest thing to build. It probably would have been considerably easier if using SMD components, but some people didn't think they would be able to take on soldering things that small. Could be there is some 3rd party output stage that could be adapted, but can't think of one offhand.

One of the Twisted Pair ones, Mercury or Legato perhaps, or even the SEN or its forthcoming BJT version??

John
 
IMHO, none of those output stages is ideal for this dac chip. At the very least, I think we should stick to opamp I/V, using an appropriate opamp such as OPA1611/1612. If we can't make that sound good, we are doing something wrong somewhere. Most likely doing something wrong in more than one place.
 
Sure, I'm using a PSU modified Soncoz SGD1. Same designer as for the Khadas toneboard. The designer solved the issue of the ESS linearity hump by correcting the opamp feedback capacitance/loading issue. It uses 2 X ESS 90382QM and fully balanced OPA1612 and RT 6863 Opamps for IV, supplied from separate regulators and transformer. Sounds and measures excellent out of the $470 box. Easy to do supply mods and versatile inputs. Saw off my heavily supply modded Katana with ease.

John
 
Interesting. The review at ASR says it has a very high ESS hump:SONCOZ SGD1 Audio DAC Review | Audio Science Review (ASR) Forum
...was that fixed in a later revision?

Anyway, how would you propose to compare output stages? Get two of SGD1, remove the output stage from one, and substitute another output stage?
 

Attachments

  • Soncoz.jpg
    Soncoz.jpg
    152.9 KB · Views: 285
Last edited:
Yep, it was fixed earlier this year. There is a thread there detailing how it was done and the recalculation of the filter cap and resistor values. I can't afford a second unit so I'll make the modifications a listen to the differences over weeks and months to decide which I prefer. Short term and even double blind testing doesn't work for me. Lots of listening does.

When I get a chance I'll post what mods I've made and why.

Note that the Soncoz makes use of ESS's own DAC chip regulator IC about which I know little.

J