China ES9018PRO ES9028PRO 9038PRO mods&upgrades

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
pictures of the boa

20170404_131739854_iOS.jpg . DSC02430.JPG . DSC02433.JPG . DSC02435.JPG
 
That looks better than many of the other es boards I’ve seen, looks much more diy friendly anyways.
I have a similar ak4495 board that shares some of the regulators.
First thing I would do would be to straighten out the loads on the onboard regulators before getting wrapped up in any external regulation schemes. The differences between what is provided and some of the super low noise regulators will be washed out in the application.

The LT1963 will work with an output capacitor that has an esr as low as .050 ohm, and the smaller ams1117 will work with capacitors that have an esr as low as .300 ohm.
What I have found to work well is to use something that has values like this with each part.
The Sanyo Oscon 47uf is going to be more well suited to being placed after the LT regulator than after the smaller ams part.
The ams1117 I have found will work well with a 100uf, 16v, starndard electrolytic such as a bc 037 for example, or the Nichicon VX.

That along with putting a .01uf COG on the clock for decoupling will make a nice difference, and there is still likely much to be gained from the output stage as well.
 
The regulators are made such that they can only be stable, and provide the advertised specifications for noise and transient response by having a certain range of output capacitor specs.
This is, in this case, not necessarily determined by the first number, or UF value of the part, but by the esr, or resistive component that these parts all have. Often they are described as impedance curves, and are really what makes one type of part, or dielectric more appropriate for the different uses that we deal with in audio, or anything else really.

Kind of like tires on vehicles; a tractor tire might fit onto a Porsche rim, and same with the car tire on the tractor rim but isn’t the ideal way.

The common commercial regulators are usually meant for mass produced consumer goods that will be using lower cost, and ready for automated assembly types of parts, like ceramic, and tantalum, as are mentioned in many of the data sheets. Most parts have been made by several manufacturers, and by digging, you can usually find more information on the various components.
The same regulators on one of my projects has responded very well to the changes mentioned above, after I had zealously applied a bit too much in the way of low esr parts to the circuits. This included the oscon types as well, which I am finding only have limited uses in digital audio. They seem to be great as long as they don’t exceed the minimum esr of the regulators, or resonate with the smaller ceramic parts that are required as local decoupling.

After the on board power supply is improved, a snubber circuit can be added to the diode bridge for a great bang for the buck, global improvement. I always make sure and put a .1uf polypropylene X type of capacitor across the incoming power as well on any line level gear.
 
Tbh I have no idea about the other layers. I got ahold of a faulty board with a short circuit between layers, so could use it for educational purpose.


So far I have upgraded the following:


- Clock changed to Crystek CCHD 950
- burson v6 opamps
- Changed Wima MKS capacitors to MKP.
- replaced counterfeit mika CDE capacitors with original CDE 1% 100 pF.
- fake Elna Silmic II are still there waiting to be replaced with originals.


IMG_20190102_165347.jpg
 
I have had great results with using a combination of Siemens KS, or EMZ polystyrene along with the Wima FKP2. Looking at the circuit, I will place the polypropylene caps in series with the signal path, and the polystyrene in the shunt positions to ground.
This has given a nice balance of punchy bass, with smooth airy treble, at least in my experience/opinion.

Those 100pf can be swapped out for better parts too...

I didn’t see any local bypass on the op amps when I looked at the boards, if you can, try .01uf film caps under the sockets, directly on the pins.
 
In that case you would have more work to do, as I recall those need additional high/medium esr parts on the pins and also tied to a ground plane that needs to be very close. I try to do that with bipolar op amps if I ever use those.
I don’t have any experience with the details of any es output filters, I avoided them since they weren’t forthcoming with data sheets.
 
As far as the output stages for any of these Sabre dacs goes, the output stage we recommend using in the 'ES9038Q2M board' thread should work quite nicely so long as the I/V resistors are scaled appropriately. OPA1612 can handle the current from a single ES9038PRO channel. Don't forget enough stacked film caps on the +-15v rails. 100uf per rail has served quite well for ES9038Q2M dacs, have to see if still sufficient when using ES90038PRO with higher currents. Likely okay.
 
Last edited:
As far as the output stages for any of these Sabre dacs goes, the output stage we recommend using in the 'ES9038Q2M board' thread should work quite nicely so long as the I/V resistors are scaled appropriately. OPA1612 can handle the current from a single ES9038PRO channel.
Is there someting wrong with the original output stage with busron v6 opamps?
The SQ is quite ok for me. Still didn't clean the power supply though.
Don't forget enough stacked film caps on the +-15v rails. 100uf per rail has served quite well for ES9038Q2M dacs, have to see if still sufficient when using ES90038PRO with higher currents. Likely okay.
seems a huge capacity for film caps, There might be a problem with space, not sure they would fit. would electolitics work instead?
 
Is there someting wrong with the original output stage with busron v6 opamps?
The SQ is quite ok for me. Still didn't clean the power supply though.

seems a huge capacity for film caps, There might be a problem with space, not sure they would fit. would electolitics work instead?

I have not heard Burson opamp, and the output stage might be okay. Don't know. I find the dacs tend to sound better with more output stage filtering than many designs use. The output stage we use for ES9038Q2M has good filtering and all good components, is all, so you kind of know it will will not be worse and might be better than what the dacs sometimes come with stock.

Also, elctrolytics did not work instead of film caps when I tried it. I used 10,000uf, .053 ohm ESR electrolytics and they sounded bad compared to 110uf of film caps. I don't know if it really takes a whole 100uf or so of film caps though, I only tested the bunch of those that were already soldered together. You could try will less and see how many you need before there is no benefit to adding more.
 
Last edited:
Moderator
Joined 2002
Paid Member
680 ohms seems very high I/V resistor. 220-240 seems more appropriate for 2Vrms out, and obviously the capacitor value should scale (upward) accordingly. With 10mA output current available from 4 DACs per channel, 680R would be really the wrong choice.

This is an output stage meant for some other, lower current DAC that has been transplanted into this circuit by the 'designer'. Not that you should expect anything more out of these kind of products - and I wonder if anything at all is optimised correctly for the ESS chips.

The best mod for this stuff is the dustbin - the fact it has a 9038Pro means very little. To maximise performance you will have to spend a lot on modifications (and some mistakes cannot be modded out) that you might better spend on a better DAC to begin with. I do know this is DIYAudio, but at least your starting point has to have some merit.
 
A 680 ohm I/V resistor is probably about right for ES9038Q2M, but not for ES9038PRO. About 3v peak to peak voltage swing at the output if the I/V amp keeps it operating in Class A bias, since the input and output are offset by AVCC/2. Maybe not a bad idea. Any additional needed gain could be picked up in the differential summing and filtering stage, if desired.

There are ways to get exceptionally good sound quality performance out of either ESS or AKM chips, at least some people think so. Benchmark DAC-3 is based on ES9028PRO and Stereophile rates it A+ recommended in terms of sound quality, and SOA in terms of measurements.
 
Also, elctrolytics did not work instead of film caps when I tried it. I used 10,000uf, .053 ohm ESR electrolytics and they sounded bad compared to 110uf of film caps. I don't know if it really takes a whole 100uf or so of film caps though, I only tested the bunch of those that were already soldered together. You could try will less and see how many you need before there is no benefit to adding more.

Very Low ESR Electrolytics can sound exactly as you found - not so good.

My recommendation, before spending $ on film caps is to first try Elna Silmic II caps. They are fairly cheap at Digikey, measure very well and sound natural.

As I've previously stated, start with just the electro before paralleling any MLCC or film caps. Once you start paralleling, especially with hi Q MLCC's, tank circuits form and it can affect the sound.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.