ES9038Q2M Board

Okay, that is a start with data up to 352kHz. However, I have had my 9038Q2M up to 960kHz. Not sure if 9028 can go that high or not. If so, you probably want to allow for 300-400mA at least. Either that or measure the highest it can go first and then design the power supply, which would seem to be the more sensible approach.

Still probably wise to open a new thread for 9028PRO.
 
Last edited:
With all the recent activity in other threads I am afraid we may not be giving enough attention to ES9038Q2M modding progress. It sure would be great to hear from anyone who has some new build or mod progress to report. Any questions would be quite welcome too, of course. Anybody have anything? :)
 
With all the recent activity in other threads I am afraid we may not be giving enough attention to ES9038Q2M modding progress. It sure would be great to hear from anyone who has some new build or mod progress to report. Any questions would be quite welcome too, of course. Anybody have anything? :)

Why not to push it to the limits then ...

es9038q2m_A123.jpg - Google Drive

1xLifePo4 A123 3.3V for AVCC
1xLifePo4 A123 3.3V for CCHD-950-100Mhz

:)
 
@bravem01, the first couple of jumpers should select the input you want to use. For my 1.06 board with J1 on and J2 off, that selects optical (TOSLINK) input. J2 on and J1 off select SPDIF (coaxial).

The other higher number jumpers select reconstruction filters which will affect the sound, so you can chose the one you prefer. Wherever you purchased the board from should have a picture in the ad showing what the jumpers do. You should save a copy of that for reference.

Also, you can usually contact the seller by using the messaging feature at ebay or aliexpress, wherever you made the purchase, to ask questions about things like that. It may take them a day or two to respond but usually they try to be helpful.

Please let us know if that helps or if there is still a problem.

In addition, don't know how much you have read about the board in this thread but you shouldn't be surprised if the sound quality isn't very good until you start doing some modifications on it.
 
Last edited:
@TioFrancotirador, That looks like it should have very clean power. I do wonder with long battery leads if they make a big enough loop to pick up stray magnetic fields?

Also, wondering what you are using for an output stage? And what about after that for amplification?

Thanks in advance for any more information you may provide.
 
@TioFrancotirador, That looks like it should have very clean power. I do wonder with long battery leads if they make a big enough loop to pick up stray magnetic fields?

Also, wondering what you are using for an output stage? And what about after that for amplification?

Thanks in advance for any more information you may provide.

Yep shortening the leads should help. I will try it.
The output stage is discrete class A zero global feedback Legato 3.1 from twistedpearaudio (my favorite one). The amp is Modulus-86 from neurochrome (my best amp so far).
with Lifepo4 for AVCC it is sooooo easyyy listening. For the clock it did not matter that much. I already had LT3042 with Salas Ref-D as preregulator for clock. Seems AVCC is most sensitive.

Now

My 2xMaxwell supercaps 3000F (yes farads!) are one the way :) I will put two in series and charge them with cc/cv @ 3.4V. Their total output impedance will be 0.58mOhm! (LifePo4 has couple for mOhms). They can urge around 200A (Amps!) energy :). No electrical nor chemical noise. Although LifePo4 has the least chemical noise, yet there is still some.

So, more to come. I will let you know about my findings soon!
 
I did not try ESS recommended opmap circuit. What I did try though is latest version of trident dual shunt from twisteadpearaudio, paul hyness shunt, Salas Ref-D, lt3042. All of them were very good, yet for the image stability, clear background LifePo4 is the winner. It is like everything became, clear, calm and relaxing. It reminds me my pcm1704 dac that I used to build some time ago.
 
For myself, I used the ESS opamp circuit. But I didn't use any wires hanging in the air to connect it to the DAC board. I mounted on the back of the ground plane so the opamps outputs were right where they needed to be to connect into the AVCC inputs on the top of the DAC board. So, no wires to pick up magnetic fields or add resistance.

I did notice it helps to put everything in an enclosed steel case and to make the +-15v power very clean and very close to the DAC board. With upsampling to 960kHz I2S or DSD512 it sounds fine. Nothing wrong at all, although the two audio formats sound a little different perhaps due to different reconstruction filtering. Either way, the sound is smooth, clean, stable, and quite good overall. It was measured at right around what ESS says at -120dB distortion. Headphone amp packaged with it came in at around -105dB which was about what I expected for it. Not as good as the DAC line outputs, but good enough to hear the DAC sound quality quite well.

However, as far as I know nobody has built another one similar to what I did with tiny circuits built on the bottom side of the DAC board. It could be that it would be easier for most builders to get good results with LiFePO4 batteries, just so long as they are very careful not to mishandle them in any way.
 
Last edited:
...However, as far as I know nobody has built another one similar to what I did with tiny circuits built on the bottom side of the DAC board.


pls. check my post 1289 in this thread. Maybe not as tiny as yours but also no free wires hanging around and on bottom side of the board.
Built everything into a case now..

I am still waiting for my Schmitt trigger ic to reduce the clock rise/fall time. I want to place it instead of the 33Ohm resistor between clock and DAC and connect it to the PSU of my clock with a 0,1nF decoupling cap directly to the Vcc pin of the Schmitt trigger - any concerns/recommendations on that plan? After my NDK clock exchange the sound became less relaxing for long time listening and the soundstage is a little narrow..Not sure if it is caused by clock exchange or by the complete change to I/V circuit..I did not much listening in between these changes.

With opamp AVCC PSU only and DAC in voltage mode with orig. clock it was somehow more relaxing but dynamic was not as good as it is in current mode now..hope I can fix this with the schmitt trigger..
Really wonder if the hot running opamps are responsible for this, or better the root cause for the hot running opamps? Probably another MLCC in parallel to 10uF filter cap at AVCC opamp input could help? Or at offset (AVCC/2)?
 
Last edited:
Regarding warm opamps, I don't know. AVCC opamps seem to run cooler than IV opamps. As I said before I sort of presume there is some low level RF coming out of the DAC, too low level to see on my scope. Probably first thing to do would be to try ESS's 1st choice IV opamp, AD797 and see what happens with that. Also, again would point out that opamps seem to be running within their specified temp range. We humans sometimes don't feel comfortable when silicon does that, but it doesn't necessarily bother the silicon. Would be nice to understand it better though.

I guess another thing we could try when running lower sample rate audio would be to reduce the internal MCLK divider inside the DAC. That should reduce the highest frequency most of the DAC runs at and there could be some benefits from doing it when we know the sample rate will be low. One benefit is lower current draw by the dac, but could be less RF coming out too. Hard to say without trying it.
 
Last edited:
The hardest part of that is the pin lifting. I could provide some software. The little Arduino 'shield' board is optional, its just the way I did it. Only thing is you would need to be careful about not applying overvoltage to the Q2M and or the Arduino by improper power sequencing. Easy to prevent though if the Arduino is a 3.3v version and the I2C pulliups on the DAC board are used, then only protection needed would be a couple of schottky diodes to clamp the Arduino I2C pins from accidentally going too high. The relay and anything else would be optional. But its nice to have, makes it convenient to use. You could try harmonic distortion compensation if you wanted. You could try just about anything that's possible.