ES9038Q2M Board

Q9 ES9028 ES9028PRO HiFi Audio Decoder SUPER DAC assembled board | eBay

I am talking about the IV side. The AVCC side, it appears that they have provided separate regulators for L & R but only 4 channels of the 9028 are tied together and that apparently is strictly for superior SNR. This is not complete dual mono. If you want to get even better SNR ratio then I would suspect that the 9038pro would give that to you. However one thing I considered was that the currents out of the DAC are higher in the 9038pro and if you tie all 8 channels from a 9028pro, then it also would have higher currents.
When I read about this thing Russ White was warning about the significant increase in currents on the 9038PRO for the IV. I don't know how to assess whether or not this is handled by the chinesium DACs on Ebay as they appear to use the same IV circuits from the 9018 through to the 9038. Seeing this I stayed away from the 9038 despite being a only few dollars more. With the 9028pro I assessed the odds of it being ok and it tilted as lower risk.
This board does come with what appears to be shunt regulator for the IV circuit rather than a series type. Shunts are now the preferred route provided they are implemented properly like everything else. Also discrete op amps which are somewhat controversial for me.
The other thing to consider, again using the Benchmark product. They stayed away from the 9038. If it was a drop in replacement, they'd be on it. So something is up. Even Twisted Pear will recommend a 9028pro depending on your needs. At this time mine are purely FLAC and some treasured MP3s of tunes from my youth. So no DSD now for the time being.
 
Last edited:
Latest ES9028PRO ES9028 DAC HiFi Audio Decoder Q8 V2.3 | eBay

The AVCC decoupling on this board is suspect. The caps are too far away from the chip. It uses the lower capacity LT1763. Also it uses the LT317 to preregulate prior to an ADP150. It also uses a clock that is supposed to be not as good as the board I got. However NONE of the clocks supplied should be considered very high grade clocks. If you want that, you will need to take care of that yourself. Even the more expensive one comes with the same clock as the board I picked up for $140...
 
Last edited:
ES9028 ES9028PRO DAC Q8 Ver2.2 HiFi Audio Decoder PCM384K / DSD128 | eBay
Q8 Ver2.2 ES9018 ES9018S DAC HiFi Audio Decoder Finished in case | eBay

I was intrigued by the above unit considering that the when you consider the cost, the case, transformer and board. It looked like good value. Schematics are disclosed in the images.
Note it uses a TL431 based power supply for the IV circuit not a 317/337. Maybe a similar circuit to the one discussed here based on a TL431 Simple Voltage Regulators Part 1: Noise - [English]

The reason I did not go this route is because I have this crazy idea to build a tube based Balanced to Single converter in the hopes to have tube sound when I feel like it. So for most listening, the Tube converter is off. I use the single ended output from the DAC. When I want tubes, I will turn on the converter, and using the Balanced out of the DAC route that to converter and then to my preamp and just turn the selector switch. I now have like two DACs! One is SS and one has some tube qualities for those times I want it. I have all the boards etc for the converter. Just need to find some ECC88s.
 
Last edited:
Wow, those are a lot of PRO chip boards and options. Without some work, doubt any of them as-is could perform at a sound quality level as well as a well-implemented 9038 mobile chip board could. That is to say, it is more about the implementation than the particular chip when it comes to any of the 9028 or 9038 variants. The things the PRO chips offer is the possibility of lower lower noise, 2dB lower distortion numbers for 9038PRO, and more pins on the chip to simplify some interfacing options. The downsides are the PRO chips cost more ($75/ea in small quantities vs $15/ea for mobile in the 9038 line), and they cost more for higher current power supplies and higher current IV stages.

In other words, for the best sound quality at the lowest cost a good, basic ES9038Q2M would be really hard to beat, probably not possible to beat. I say 'would' because nobody makes a good, basic board with the cost put into sound quality instead of other stuff.

I sure wish there was some light at the end of the tunnel, but there probably isn't. The DAC markets don't seem to work that way or there would probably already be such a thing.
 
Last edited:
We know that all of them can be improved by a DIYer. All a question of the resources you are equipped with including knowledge and equipment. I've been down that road many times. In the old days of the Audio Amateur. POOGE
POOGE - Progressive Optimization of Generic Equipment (aka tweaking) | AcronymFinder

I've begun dry fitting of my DAC on the other thread. It is starting to look decent and to my ears it sounds decent as well. At the end, the equipment will be able to sit with the rest of DIY stuff with no excuses that it looks homemade. I'm done with those days. Moving from 9038Q2M to a 9028Pro board
 
Last edited:
Since we seem to be showing off packaging ideas, I have my test circuitry in part of a steel file server case. I can test different power supplies and SRC options, etc. using this.

Eventually I will put some selected components in a smaller box, but i don't want people to start getting the idea I am opposed to cases. I'm not. But, for a DIY project I would like to see performance come first and then a case found that fits rather than finding a case and adjusting sound quality to fit.
 

Attachments

  • TestCase.jpg
    TestCase.jpg
    504.9 KB · Views: 555
Q9 ES9028 ES9028PRO HiFi Audio Decoder SUPER DAC assembled board | eBay

I am talking about the IV side. The AVCC side, it appears that they have provided separate regulators for L & R but only 4 channels of the 9028 are tied together and that apparently is strictly for superior SNR. This is not complete dual mono. If you want to get even better SNR ratio then I would suspect that the 9038pro would give that to you. However one thing I considered was that the currents out of the DAC are higher in the 9038pro and if you tie all 8 channels from a 9028pro, then it also would have higher currents.
When I read about this thing Russ White was warning about the significant increase in currents on the 9038PRO for the IV. I don't know how to assess whether or not this is handled by the chinesium DACs on Ebay as they appear to use the same IV circuits from the 9018 through to the 9038. Seeing this I stayed away from the 9038 despite being a only few dollars more. With the 9028pro I assessed the odds of it being ok and it tilted as lower risk.
This board does come with what appears to be shunt regulator for the IV circuit rather than a series type. Shunts are now the preferred route provided they are implemented properly like everything else. Also discrete op amps which are somewhat controversial for me.
The other thing to consider, again using the Benchmark product. They stayed away from the 9038. If it was a drop in replacement, they'd be on it. So something is up. Even Twisted Pear will recommend a 9028pro depending on your needs. At this time mine are purely FLAC and some treasured MP3s of tunes from my youth. So no DSD now for the time being.
The super DAC ES9028PRO board seems to be a much better deal than the basic ($138) board. It includes USB, upgrade TXCO, upgrade opamp and better regulators. But the seller does not sell it in fully assembled box!
 
upgrade TXCO, upgrade opamp and better regulators.

TXCO is worthless, it needs ultra-low jitter clock not temperature controlled clock. Two different things, and important to keep apart in one's memory. (Note: a clock could have temperature control and be ulta-low jitter too, but the temperature control part doesn't help with DAC performance.)

Also, discrete opamps are only for looking cool and adding distortion. There is no better opamp for one of these DACs than AD797 or LME49720 family (including LME49860).

Understandable in that there are lots of things to remember about what is important for making a good DAC, but we may need to keep reminding ourselves of what is eye candy and marketing foo, vs what actually helps for making the best DAC. Remember, the marketing guys know what makes you tick better than you do. You may not think so, but they do.
 
Last edited:
TXCO is worthless, it needs ultra-low jitter clock not temperature controlled clock. Two different things, and important to keep apart in one's memory. (Note: a clock could have temperature control and be ulta-low jitter too, but the temperature control part doesn't help with DAC performance.)

Also, discrete opamps are only for looking cool and adding distortion. There is no better opamp for one of these DACs than AD797 or LME49720 family (including LME49860).

Understandable in that there are lots of things to remember about what is important for making a good DAC, but we may need to keep reminding ourselves of what is eye candy and marketing foo, vs what actually helps for making the best DAC. Remember, the marketing guys know what makes you tick better than you do. You may not think so, but they do.
Unfortunately, I have neither the technical knowhow and the golden ear to hear the best sound and always refers back to your comments. I am seriously consider this instead of the SU5 Q2M.
HiFi ES9018 ES9028PRO ES9038PRO+XMOS U208 USB DAC 32Bit/384K DSD 64/128/256 XLR Output U208 USB Card-in Amplifier from Consumer Electronics on Aliexpress.com | Alibaba Group
The seller mentions this:
ES9028 with USB :OP Amps are 2PCS LME49720NA+2PCS LME49710NA
Transformer: 1PCS OFC 50WR Transformer

I hope to hit the jackpot like Mikett did and get something good AS IS and can be upgraded as I go along. This looks like a more open case for mod and hopefully, the design did not make the same mistake as the SU5 board.

For now, I am still hanging onto the MShow Q2M box running off a LT1083CP linear PSU.
 
Last edited:
@keilau, I looked at the pictures. Maybe better but still a ways to go from being as good as my modded board. Can't tell exactly, but might need new AVCC supply and new clock. Output stage resistors look like probably 1% metal film, whereas I would use .1% or better thin film for better channel matching, and of better quality to minimize nonlinear distortion. There are film caps where there should be C0G in the signal path, and probably not enough good power decoupling for opamps with 55MHz GBP where there should be 10uf tantalum in parallel with maybe .05uf to .1uf X7R, instead of a single .1uf film cap. SMD tends to be better than leaded to keep inductance low.

We don't know what the bottom of the board looks like and if there is proper ground plane. That could raise some issues.

Unfortunately, we never really know the story on these boards until someone gets one at starts looking at it close up.

We do know how to fix most everything of course, or you could just leave it and enjoy it for whatever it is.

None of all of these boards seem to allow for upsampling, which for whatever reasons can help. Unfortunately, some of these boards don't give the user choice of reconstruction filters so it may not be possibly to fully benefit from upsampling by make use of a broad transition band to minimize group delay in the pass band.

The only good news is there does seem to be some general trend towards better implementation as price goes up, but the effect seems small for each price increment.
 
The super DAC ES9028PRO board seems to be a much better deal than the basic ($138) board. It includes USB, upgrade TXCO, upgrade opamp and better regulators. But the seller does not sell it in fully assembled box!
That is the same board including regulators as the one I got for $137. Couple things. It appears to not come with heat sinks on the LT1963. Other than the op amps already upgraded...it is the same board for $138. If you look around you will find the same board over and over on Ebay Aliexpress and Tao Bao. It definitely is a 4 layer board. Was changing some bypass capacitors today and the ground side needs more heat than the other side indicating possible a ground plane sandwiched in there. That is important.
One thing to note is that the bypassing on the op amps indeed is somewhat inadequate for the wide bandwidth LME497xx family and AD797. It is easy to fix that just as Markw4 had suggested. The op amps are not expensive but you have to shop at a reputable seller or else you could end up with fakes.
 
Last edited:
@keilau, I looked at the pictures. Maybe better but still a ways to go from being as good as my modded board. Can't tell exactly, but might need new AVCC supply and new clock. Output stage resistors look like probably 1% metal film, whereas I would use .1% or better thin film for better channel matching, and of better quality to minimize nonlinear distortion. There are film caps where there should be C0G in the signal path, and probably not enough good power decoupling for opamps with 55MHz GBP where there should be 10uf tantalum in parallel with maybe .05uf to .1uf X7R, instead of a single .1uf film cap. SMD tends to be better than leaded to keep inductance low.

We don't know what the bottom of the board looks like and if there is proper ground plane. That could raise some issues.

Unfortunately, we never really know the story on these boards until someone gets one at starts looking at it close up.

We do know how to fix most everything of course, or you could just leave it and enjoy it for whatever it is.

None of all of these boards seem to allow for upsampling, which for whatever reasons can help. Unfortunately, some of these boards don't give the user choice of reconstruction filters so it may not be possibly to fully benefit from upsampling by make use of a broad transition band to minimize group delay in the pass band.

The only good news is there does seem to be some general trend towards better implementation as price goes up, but the effect seems small for each price increment.

Yes, the incremental upgrades are like going to a car dealer. I looked at the upgrades and they are pricey for what you get. Again choose your compromise.
 
Last edited:
TXCO is worthless, it needs ultra-low jitter clock not temperature controlled clock. Two different things, and important to keep apart in one's memory. (Note: a clock could have temperature control and be ulta-low jitter too, but the temperature control part doesn't help with DAC performance.)

Also, discrete opamps are only for looking cool and adding distortion. There is no better opamp for one of these DACs than AD797 or LME49720 family (including LME49860).

Understandable in that there are lots of things to remember about what is important for making a good DAC, but we may need to keep reminding ourselves of what is eye candy and marketing foo, vs what actually helps for making the best DAC. Remember, the marketing guys know what makes you tick better than you do. You may not think so, but they do.

Hi Mark

What are the discret opamps you have tried?
Serge
 
What are the discret opamps you have tried?

I am expecting to try some soon, but so far I haven't tried any for DAC IV use.

However, they generally do add some distortion compared the opamps recommended by ESS for Sabre DAC IV outputs.

Sometimes some people like a little tube distortion or Class A distortion. No question, some people like it. However, and this is the important thing to remember, you can't make a DAC with remaining implementation problems sound good by adding some so-called 'good' distortion. You need to have a very clean, well-implemented DAC first. If you have that and then you want to see what a little Class A distortion or Tube distortion can do to the sound, that would the correct point at which to try it.

Therefore, my advice in this thread is definitely fix everything you can fix and get a good clean DAC going first. If you get to that point then you can stop there and have something very nice that you can enjoy and that can serve you well for maybe 10 years or more. And if you want to play around with Class A or Tube options to your taste then you can do that too.
 
Why this board has no succes ?

Maybe because it is SPDIF input only. And maybe because the clock appears to go through a SPDIF receiver instead of directly into the DAC. Maybe because no single-ended output. Maybe all the above. It's not what most people want and I wouldn't want it either. Somebody would have to give me one to try if they want to prove what they have done is a good idea. So far, nobody seems convinced the architecture makes sense to get the best sound.
 
Last edited:
That is the same board including regulators as the one I got for $137. Couple things. It appears to not come with heat sinks on the LT1963. Other than the op amps already upgraded...it is the same board for $138. If you look around you will find the same board over and over on Ebay Aliexpress and Tao Bao. It definitely is a 4 layer board. Was changing some bypass capacitors today and the ground side needs more heat than the other side indicating possible a ground plane sandwiched in there. That is important.
One thing to note is that the bypassing on the op amps indeed is somewhat inadequate for the wide bandwidth LME497xx family and AD797. It is easy to fix that just as Markw4 had suggested. The op amps are not expensive but you have to shop at a reputable seller or else you could end up with fakes.
I got my opamp from DigiKey. No only get ease of mind of getting authentic part. They are often cheaper than most ebay sellers.

The super board has the USB subboard that the basic board does not have. That's why I think it may be worth it.

I am looking at some assembled ES9028PRO boxes for $300 on AliExpress. The opamp is already upgraded to LME89720/89710. I plan to do the dual toroidal transformer upgrade myself. Then I worry about the other mod you and Markw4 suggested.
ES9038 ES9028PRO ES9018 DAC audio decoder Amanero USB interface Support DSD coaxial optical with Remote Control Shield USB cable-in Digital-to-Analog Converter from Consumer Electronics on Aliexpress.com | Alibaba Group