ES9038Q2M Board

kimschips,
Maybe it would help if you would tell me what you tried so far: Do you have a Windows PC? If so did you download and extract the .zip file to a folder on your local computer? If so, what programs did you use to try to open the RTF file? If it didn't open did you get any error messages, and if so what did they say?

It works fine on my windows PC
Thanks
 
WRT 11V vs 15V PS rails, that is surprising. Normally the opposite is the case and they nearly always measure slightly better with higher rails.

Yeah, I was surprised too.

WRT the I-V schematic, have you checked what LPF function it is or if it is 'linear phase'? There have been plenty of published I-V LPF's that didn't have
linear phase response which is a no-no.

The one I suggested is Gaussian. Of course, while optimal for phase it does that at the cost of very gradual frequency response roll-off. I didn't check to see if the trade off there was optimal as used with ES9038Q2M.

WRT NPO ceramic caps, they do measure very low distortion but I've nearly
always used Wima FKP (Film+Foil polyprop) caps for LPF circuits.

One could try that. NPO/C0G work well for some pretty good dacs costing in the thousands of dollars.

Tell you what, why don't you help us out for once and post your suggested design schematics, BOMs, and layouts for I/V stages, reclockers, whatever else you got? Then we can question you about if you thought of this or that.
 
Mark, I have suggestion for something you might try with ak4499.

AK4499 NOS mode sounds pretty poor, but of course the internal filters arent really much better, however using NOS mode in conjunction with HQ Player providing external filtering gave some absolutely amazing results.

Since its PCM you run Closed form M at max 705.6 easily, which is probably a big part of the amazing sound. No idea what a ''closed form'' filter is other than its based on another ''closed form'' filter from Schitt and is computatively demanding, but it sounds a hell of a lot better than the other filters, even the standard tap version.

I havent tried DSD with Ak4499 but this clearly sounds better than PCM with no software oversampling which is a first, would be interested to hear if you think this sounds better than DSD.
 
Since its PCM you run Closed form M at max 705.6 easily, which is probably a big part of the amazing sound. No idea what a ''closed form'' filter is other than its based on another ''closed form'' filter from Schitt and is computatively demanding, but it sounds a hell of a lot better than the other filters, even the standard tap version.

When I have heard Mike Moffat of Schiit talk about his 'closed form' filter it seems to me he conflates two things - the way the filter operates and the way its designed. In my understanding 'closed form' is a mathematical term that says there's an equation to come up with the coefficients which gives precise values. Pretty much how you'd design say a 2nd order Sallen-Key filter - there's an equation you plug your corner frequency and impedance into and out come the precise component values. Digital filters of FIR form aren't normally designed like this, they tend to use a numerical algorithm developed by Parks and McClellan - Moffat appears to look down on this method as 'an approximation' as if successive approximation/optimisation to converge to a suitable result is something bad.
 
Laserscrape, May I ask if you personally have been listening to find out how AK4499 sounds or if you have been reading someone else's experiences?

I'm trying to find out where and how the 'sounds poor' words/opinion came about or originated.

Also, I agree with Abraxalito about the meaning of the term 'closed form.'

My own experience with AK4499 is that how it sounds with PCM verses DSD and how sample rate dependent that is all depends a lot on how AK4499 has been implemented. It can sound quite good in the best case, and if mistakes are made it can sound more like as you seem to be describing. Ultimately though, I probably like DSD256 best. But making that choice is hardly the end of it. If the dac doesn't sound good with native 16/44 then something is still wrong with the implementation IMO.
 
Hej guys
What do you say to AKM4493 versus ES9038?

AK4493 is probably a lot easier to implement well, would be my opinion. That said, if ES9038Q2M is implemented at its best I think it would probably win a comparison over AK4493. Since getting the best out of ES9038Q2M requires some specialized expertise, I would probably suggest AK4493 as an attractive option.

However, the AK4493 dac you linked to is another of the cheap Chinese dacs not too different in some ways from the ones we mod in this thread. They are all made with the cheapest possible parts except for the dac chip is real. Other parts with brand names or similar looks to brand name parts are often fakes.

A quality implementation of AK4493, more like you would want to end up with, might be the one offered by JL sounds at: Products - I2S over USB Audio
As you may see, it still needs a USB board and a controller/display, an output stage, power supplies, etc. By the time all is said and done a fair amount of money will be invested.

Same thing about investing money if you go with a cheap Chinese dac only you will have to upgrade the components to get the best sound, if that's what you are seeking. You would still need a USB board, maybe a SPDIF/TOSLINK board, maybe an ASRC board (AK4137), output stage, power supplies, etc.

So, thing is you can learn and have fun doing diy dacs or you can spend money to have others do the work for you. Either way costs will add up if you want near-SOA performance.

If you would be happy with a nice little 16-bit dac for cheap, I would suggest you go check out Abraxalito's thread and see how that appeals to you. Much lower cost for some diy fun and good value for money.

Another option besides JL Sounds could be a more complete dac board with USB already built-in but still needs a power supply. In looking for something like that I would consider the new: Revolution DAC - Digital input USB (uUSB) - Analog output RCA 2.25Vrms ...those guys also have a thread here in the vendor part of the forum.
 
Laserscrape, what super slow roll off PCM filter, and what is 'equivalence to NOS with stair-stepped output'? You are feeding it high sample rate PCM with a zero-order-hold similar to a lower sample rate with no interpolation filter? Is that it?

It would sure help us understand if you would be willing to reveal a bit more about what you are up to, maybe write a paragraph or two describing the hardware and any custom software (even post a pic or a maybe a few). I mean, if you are planning to sell something and just want to keep it a trade secret then we could understand that. Otherwise if you are reticent to avoid possible criticism of a diy project or something like that, we could always take it to PM if that would be better for you. As it is now I am having trouble understanding where you are implementation-wise. What sounds good/best depends a lot on that.
 
Last edited:
There is nothing really special about my suggestion, it should be naturally possible to do it with nearly any AK449X dac, commerical or DIY.

I assumed you would be aware of the built-in ''super slow'' filter mode these AKM DACs have, it is zero order hold ''filter'' like you describe.
In this mode, when oversampling was provided externally from HQPlayer it sounded very good.
Better, I think, than when HQP was still providing oversampling but the AK4499 was set to any of the other real filter modes it has.

I was oversampling to 352.8kHz in HQP, the max rate for my Amanero interface.
Maybe at 705.6kHz the internal filter would be completely overridden, so whatever internal filter the ak4499 was set to would make no difference to sound.

This is all regarding PCM, if that wasnt clear.
 
This is all regarding PCM, if that wasnt clear.

Okay. For the external filter to dominate over the AK filter its corner frequency and transition band have to be lower/sharper than the built in filter. In that case the external filter largely dominates, but not completely.

So then next thing to understand please, are you comparing 384kHz in both cases (just changing the filter in AK to different settings), or are you comparing 16/44 in one case to 24/384 in the other case?

Also, may I ask what the peak digital level going into the dac is in both cases, with and without external filtering? Are you keeping it to -3dFS peak or something else?
 
Last edited: