ES9038Q2M Board

Although a proper vacuum solder pump with a rework station is the ultimate, the second best and most cost efficient is this: Big Blue Solder Sucker.
It beats all of the smaller one's hands down (and I've used many), for it's vacuum capability, and IMNSHO any of the other options for desoldering components is just a pain in the A$$.

Thanks for that tip. I've never got along with any of the vacuum suckers I've tried, but they were all much smaller than that one.
 
whatever would allow you to understand how the circuit works.
so, what about "+" for the cap #9 then? does it measures +12v or something, or is it still short to ground?

Pardon me for being a bit slow today, migraines seem to reduce my IQ by a good 50%...

Do you mean measure the voltage from the + pin of cap 9 to the point where the PSU (at the moment, a wall wart) connects to the board? Or from the + pin on cap 9 to the pad I marked as 1 on the last picture I posted?
 
I liked Mark's suggestion to use a stainless steel needle to poke through solder locking holes in the board, so I made a little tool from a needle and some epoxy clay to form a handle:

attachment.php


The only progress I've made on my board today is to fit a set of screw terminals in preparation for powering it from an external PSU.

attachment.php


Because this blue version doesn't have holes drilled in it for mounting such terminals I just expoxied the terminals onto the board.

attachment.php


Then I used three little pieces of copper wire to connect the pins of the terminal block to the solder pads on the underside of the board:

attachment.php
 

Attachments

  • Hole poking tool.jpg
    Hole poking tool.jpg
    70.8 KB · Views: 372
  • P7150002.jpg
    P7150002.jpg
    116.6 KB · Views: 368
  • P7150003.jpg
    P7150003.jpg
    128.4 KB · Views: 396
  • P7150004.jpg
    P7150004.jpg
    143.6 KB · Views: 378
Okay, I better come back in.

Ian, You have the white wire shorted to ground, it sure looks like. The points on the board you have labeled 1 and 4 are negative power and ground, respectively. The point you have labeled 3 is the positive power. The point 2 would be where the dual supply would connect to ground (assuming you cut the trace I put an X on in the picture I posted. You should be able to check by measuring from point 3 to ground, it should charge up some capacitors so the resistance should appear to climb and eventually stop. It should not be low resistance to ground.

Also, on the bottom of the board where you have soldered on some bare wire to the terminal strip, that crosshatched looking part of the board under the bare wires is the ground plane. There is only a very think layer of green solder mask insulating that area and solder mask is not really intended to be a reliable insulator. So, it looks like a big short circuit waiting to happen.

I will post a picture of where you need to measure. You should measure to the power pins on the opamp, pins 4 and 8. 4 is negative and 8 is positive. Obviously they should not be shorted together or measure a low resistance between them.

Also, if you go back to the first picture I posted for you of this part of the board I labeled the polarity of the power pins for you. You remember that, right?
 

Attachments

  • Ian.jpg
    Ian.jpg
    105.2 KB · Views: 93
Hi Mark

That's very helpful, I'm understanding now. I connected my repair wire to point 4, so it is connected to the negative power (1) but also to ground because I haven't cut the trace connecting 4 to ground yet. I do indeed remember the picture you posted, I saved it to refer to.

Should I now go ahead and cut that trace between 4 and ground?

I didn't realise that about the ground plane on the underside, I'll have to fix that, I'll replace the bare wires with insulated ones, thanks for catching that error!

So, first thing I will do is replace those bare wires with insulated ones, then I shall make the measurements to pins 4 and 8 on the opamp and report back. Cheers!
 
Aah, that's why I haven't cut it - because I'm restricted to using a wall wart or battery at the moment as I'm still waiting for the slow boat from China to arrive with my voltage regulator before I can build a dual rail PSU.

I've put together a little battery to power the board for testing from four 18650 li-on cells in series, I found a barrel connector in my box of bits so I can hook it up to the board using the connector it already has.

So, don't cut that trace until I've got a working dual rail supply ready to connect to the board, right?
 
I took some measurements with my meter in continuity mode:

pin 4 on the op amp to the negative solder pad I labelled 1 shows good continuity

pin 8 on the op amp to the positive solder pad I labelled 3 doesn't, the meter shows a reading of 1.601 but doesn't beep to indicate good continuity

Hmm, not sure what this means.
 
Ian, you posted a route, which looks ...

@eziitis, Thanks for trying to help out Ian with some stuff that is obviously new to him.

For my own part, I am trying to start backing away a little from daily support in this thread. There are still issues I continue to think about for improving these DACs, and that I may decide to do something about. If that happens, I will post about it here.

Also, a review of the Allo Katana DAC should be forthcoming after I get it up and working, hopefully very soon. (Still waiting for an email reply to my query.)

However, hopefully those who have been helped by information posted here in this thread by others would be willing to help those who happen to come along later. That is to say, if you have been helped by what has been posted here, then it might be gratifying to consider contributing back by returning the favor and helping someone else who happens to come along later (even if it can seem a little frustrating at times ;) ).
 
Last edited:
pin 8 on the op amp to the positive solder pad I labelled 3 doesn't, the meter shows a reading of 1.601 but doesn't beep to indicate good continuity

Might mean the trace under the burned cap lead is also a little burned on top of the board underneath the cap body. Maybe a good idea to get the cap out of there and take a look. I would recommend heating up the lead that is not burned while adding some fresh solder or at least flux and trying to wiggle one side at a time up as much as it will go without excessive force. Once a little gap starts to open between the body and the top of the PCB you can stick a little screwdriver in there a pry while heating and adding solder/flux on the bottom to continue coaxing it out. A small vice to hold the PCB can make it go a lot easier. No doubt the cap can be removed and any trace/board damage repaired. Please ask if unsure about any problems you may run into.

EDIT: Might also help to cut the cap leads off flush before starting, that way there will be less to pull up and out.
 
Last edited:
So, I removed the cap and everything looked fine underneath so I threw the cap away, soldered a new one in place and also changed the jumper wire so it goes direct to the negative terminal.

I get a good circuit between 1 and 2, but there is no continuity between 1 and 8. What puzzles me is that there is also no continuity between 2 and 3, which are the two pins of the capacitor, and I do have the cap the right way round, I triple checked that.

attachment.php


So now I'm confused....
 

Attachments

  • Repair-6.jpg
    Repair-6.jpg
    329 KB · Views: 413
New SQ Improvement Finding

Was doing some testing with my modded DAC today in preparation for a three way comparison between it, DAC-3, and Allo Katana.

Found the setting of the AK4137 that gave the closest match to DAC-3. It was 11.2MHz DSD with the Short-Sharp AK4137 filter. It made the cymbals slightly more EQed up as compared to DAC-3. Also tried Short-Slow filter which made the cymbal EQ sound a little low compared to DAC-3. Overall, Short-Sharp seemed to be the closest fit.

Also noticed that AK4137 DSD mode seemed less distorted than PCM mode. In addition with PCM mode there are many more filter combinations to try since there are the ES9038Q2M filters and the AK4137 filters.

In DSD mode there is only one default Q2M filter, so only the AK4137 filters to try.

Also noticed that even using DSD my modded DAC sounded a bit more distorted than DAC-3. The distortion sounded like jitter to me, not analog output stage or AVCC type distortion.

After thinking about it a bit more I remembered there are some registers in ES9038Q2M to control the DPLL bandwidth, and that best sound quality is obtained using the lowest bandwidth setting that can stay locked in with good stability.

Using my Arduino and shield setup I tried reducing the DSD mode DPLL bandwidth as much as possible and sure enough the distortion was made less. Turns out the microcontroller on the Chinese DAC leaves the DPLL settings at their defaults. (There is one setting for PCM/I2S and one for DSD, two in total.)

Anyway, it definitely and audibly reduced the distortion and now in DSD mode my modded DAC sounds closer than ever to DAC-3. (I am also dropping the maximum digital level in Windows to prevent clipping in any ASRC processing in the modded DAC and AK4137, which is also important to do). While I can still tell a difference between DAC-3 and my modded DAC now is it is getting small enough I might almost be satisfied if I only had my modded DAC. I can see the game from this point on would be to continue to work on getting jitter down before it goes into the ES9038Q2M so I could reduce the DSD DPLL bandwidth even more. If I could get it down low enough it might give DAC-3 some decent competition. Every notch that DPLL setting goes down the better it sounds and there are maybe 3-4 more notches for DSD. Makes me realize I should have tried it before.

So, bottom line, my recommendation would at this point is to use AK4137 (and a there is new cheap one out on ebay for $33, which I expect to have one here to try in the next few days). Might be worth putting a better clock in that and seeing if I can get jitter down more, or maybe try to see if I can hook up SRC4392 in I2S mode and see how much I can get DPLL bandwidth down in PCM mode.

Anyway, looks like I am back in the game for awhile longer between distortion reduction experiments and Allo Katana reviewing. :)
 
Last edited: