ES9038Q2M Board

[freezebox]

Thank you eziitis. Maybe I will try later, but for now happy with 12,75V.



Markw4, I still have some questions regarding DAC powering:
1) Do you power both VCCA rails with same external powered LDO keeping onboard ferrite filters and Caps for each VCCA rail?
2)Do you power MCU by external PSU (same PSU as 1)?) or from onboard +15V line->7805Vreg(guess you changed it to better one)->AMS1117-3,3?


I am straight thinking how to handle my LP5907-3,3V LDOs probably substituating the ferrite filter resistors for both VCCAs, Clock, DVCC and feeding the currrent 3,3V lines with 5V externally as source for the LDOs. Guess both VCCAs should be fed from same LDO, no?

[Markw4]

Originally Posted by freezebox

Markw4, I still have some questions regarding DAC powering:
1) Do you power both VCCA rails with same external powered LDO keeping onboard ferrite filters and Caps for each VCCA rail?

I am currently tying the two VCCA pins together using a jumper wire between the filter caps on each side of the dac chip, bypassing the inductors. However, the LT1763 is connected in series with one of the inductors. Don't know if it makes any difference. In any case, both VCCA pins should both be fed from the same power supply one way or another.

2)Do you power MCU by external PSU...

The MCU is still powered as it was originally. No changes. The external 5v power supply I use feeds the original 3.3v regulator on the board as well as all the new LT1763 regulators.

[freezebox]

Thank you Markw4


I understand VCCA,



So do I undertsand right, MCU is also powered by external 5V supply via ams1117, right (which for me is not originally)? So orig 5V pre-reg is not used, correct?


@Mikett: my new PSU 5V rail is adjustable from 4,5V to 5,4V only from stock...

[Itsmee]

Originally Posted by Markw4

Probably using LT304x as a buffer for the LTC6655 3.3v reference on the AVCC...

A simple tweak could improve the performance - if it works, reconfigure the LT304x op amp as a sallen key low pass filter, I need to buy one to try.

Originally Posted by eziitis

...Vref apparently is TL431...

TL431 are noisy, I use them in my shunt version of the Super Reg with dedicated 12dB sallen key, ~ 1.5Hz (47K & 2u2F).

[Markw4]

Originally Posted by freezebox

So orig 5V pre-reg is not used, correct?

Yes, correct.

[eslei]

hi, there,

there is a es9038q2m i2c linux driver on github,

GitHub - VinnyLorrin/ES9038Q2M-Linux_Driver

i'll compiler and try it in next a couple of days.


cheers

[freezebox]

Hi,
for DAC and clock powering by LP5907 LDO I wonder how to handle the info about max output cap size for this reg. It is written in datasheet as maximum output cap size 0,7 to 10 uF. For both VCCA I have 2 tantalium filter caps (probably 10uF each)+ 2 ceramics already on the board. Is it necessary to place a resistor (e.g.47Ohm) or the ferrite bread between the LDO output (with 1uF cap to ground before the bread) and the VCCA / Clock filter caps to not exceed max cap size though?


Would it make sense to place a ferrite bread at AMS1117 (which feeds MCU and DVCC) input to not pollute the 5V rail which also powers the other LDOs for VCCAs and Clock?



another thing about opamps. Markw4 mentioned that AD797 needs resistor before a low ESR filter elko (if bigger than 4,7uF) - I remember a professional German tuner for Linn streamers and oppo and sonos players (Dr. Gert Volk) uses LME49720 with very low ESR FP organic polymer caps (7mOhm, 330uF) in both power lines and feeds them with a bank of caps through a 100 Ohm resistor. No additional ceramics or tantal. He also recommended to place a 100nF film cap over the input power pins if other solution is not feasable. Maybe this would also help for our device to get ridd of the slightly brightness with those opamps?

[Markw4]

@freezebox, The AD797 data sheet warns about stability issues with decoupling cap size/cap ESR. Other opamps are available that don't have that same issue. I may try an OPA1612 for AVCC, but I think LME49720 can work okay too. No reason why I have to stick with AD797, but if I felt I really needed it then I would add a little series resistance with the AVCC power rails to stabilize it. Probably not worth the bother at this point. Besides, I want to keep the AVCC power rail impedance low.

Regarding remaining brightness, I need to redo harmonic distortion compensation since I added more filter caps and changed opamps to OPA1612. Also, I probably should look at adding a little RF filter between the dac outputs and I/V inputs, just to see what happens with that. That would probably require redoing harmonic distortion compensation one more time though. Another place to look for distortion might be in the interpolation filter, and is why I started looking into using Spartan 6. To the extent that may be a problem, a fix may be a ways off. No time to work on it recently and probably won't get back to it until Katana listening comparisons are over.

Another area affecting distortion is DPLL bandwidth. I was able to get it down to 3 or 4 out of 10 yesterday. It is rather tricky, and dealing with it empirically all the more so. I want to see if I can rig up something to measure with. Have to see on that. I will write some more about all that pretty soon, but the important thing for people to be aware of is that register tweaking is a must for best sound quality and lowest distortion.

Will try to keep picking away at remaining imperfections. However, please remember I am comparing the modded dac to the SOA DAC-3 which uses additional technology that we are not using in order to get its sound. That and well equipped electronics lab to do a lot of measurements. The multi-layer PCB (6 layers?) it uses is well past its 2nd revision. It is also their third dac whereas this is my first, and at rev 2, or maybe only 1.2.

And as I said before, we probably don't need to worry so much about the stuff we are already doing pretty well, we need to worry about the stuff we aren't doing at all.

I also don't know how far we can get using off-the-shelf 100MHz clocks rather than buying reels at a time of custom spec'ed lower frequency clocks. There are lots of issues here if one wants to keep moving towards SOA.

I could probably go on and on, but no use in doing so. It would help if we had a few people caught up to me with their builds so we could compare experiments, or break up experimental work between us in a thoughtful, organized way. How about that?

[Markw4]

Regarding immediate plans, now that Katana 1.2 is here, it is a good time to finish up where I am right now, and do some listening comparisons. Who knows, maybe I will learn something. For that, I need to redo harmonic distortion compensation for the modded dac, and then just leave it at its present iteration.

I will take some time to compare different power configurations, and board choice configurations for Katana. Want to try it with power that Allo recommends, and with power that I use for my modded dac. I figure that will have to be good enough power to go with a $250 dac board, besides its the best I have .

When I get Katana sounding as good as I can, then I will do a careful 3-way listening comparison and let you all know where things stand.

Also, while I have Katana here, I will take some pictures of the boards and if there is interest, I will explain about some of the design details that can be gleaned by inspection.

[eziitis]

Originally Posted by freezebox

... for DAC and clock powering by LP5907 LDO ....

perhaps you may try out this one