Soekris Dam Dac

[Bmxmen]

Okey it works, i can download only on putty. I have give dam ale 99 filters and fpga,uc. When i have umanager 0.9 i must give all from 0.99?

[gstew]

First, some background...

Early this summer, I took the mods on my DAM1021 Rev1's almost as far as one could go without going to LiFePO4 or Ultracap power, as documented here and in a few following posts:

https://www.diyaudio.com/forums/digital ... 83813.html

Also around the same time (give or take a couple of months), I got some recent ESS-DAC-chip based setups running... a couple of Twisted Pear Audio Buffalo-IIIPro setups (one the ES9028 version with a Legato I/V stage, the other the ES9038 version with their Mercury I/V stage), Ian Canada's prototype ES9028Q2M RPi DAC with Onetics output tranformers, and the Allo.com Katana. As I summarized in this post:

https://www.diyaudio.com/forums/vendor- ... 60908.html

"At their best, they beat my modified Soekris in detailing and start to get near it in naturalness, thought the Soekris still has that 'R2R' sound that is closer to good vinyl than most digital reproduction."

So I'm looking at ways to increase the resolution of my Soerkis.

Some thoughts...

First there is the LiFePO4 / Ultracap supply direction. I DO intend to go there after Ian has his offerings available. So that's a direction I will be going, just not right now. As I suggested in this thread of Ian's:

https://www.diyaudio.com/forums/pc-base ... 26395.html

... his simple but well-designed RPi DAC with Onetics transformers on the outputs on an RPi->Allo Isolator->Allo Kali reclocker stack (with each of the other boards in the stack powered by Uptone Audio LPS-1.2s... 2 in parallel for the RPi) is in many ways the best digital setup I have running. AND this is without running it on a linear-modified RPi. In this case, I'm running it with an attached charging source similar to the LiFePO4 float charging using an LPS-1 (sounded better than a LPS-1.2). I did hear a tiny bit if SQ drop with the charger connected, but not enough to lose the always-on no-hassle convenience. Ian did hear a sufficiently significant SQ drop in his experiments with this to not pursue it. I SUSPECT my use of an off-the-grid charging supply reduces that SQ drop, I still need to confirm that and see if I can get Ian to try it too.

I'm planning to wait for his boards so I can develop a 'systems' approach I can use with the other various DACs I have running here, just as I currently use 2 sets of multi-rail backplane supply to power those DACs.

I know this is a necessary step to the best available sound. Soon.

Then on the current mod-set, there are some additional steps worth trying. AnalogSA mentioned in this post it being easy to use a separate regulator to power the DAM's clock separately from the FPGA and other housekeeping chips by removing a 1R resistor:

https://www.diyaudio.com/forums/vendor- ... 54152.html

That should be an easy mod to implement in my current setups and will be one of the next ones I do. Also Peter used the technique also shared on the DIYAudio main DAM thread to use an output transistor to convert the final stage VREF 3.3V regulators into mini-Superregs. Even though I know this will be replaced by the direct LiFePO4/Ultracap power at some point, it is worth trying now to see what other sonic gains are available in a standard-powered setup. I'll report how I make out on these.

Then there are the other mods that AFAIK pretty much only people here have tried that I haven't implemented yet...

- Replacing the FPGA to shift-register ceramic coupling caps with SMD film caps.

- Bypassing and removing the I2S isolator chips.

I'm assuming both of these will provide some level of increased resolution. My question to those of you that have done these (looking at you Nigel and Randy), what level in resolution increase did you hear doing these mods?

Then of course there is using later and improved filters. I'm using the last filters posted on Paul's dead and lamented DAM DAC website, which I felt were definite improvements over the stock filters at the time. Since that time additional ones have been been developed and made available on the DIYAudio thread (here: https://www.diyaudio.com/forums/digital ... 2r-29.html ) AND Soren has release new firmware increasing the available size of the filter and new filters to go with them, though comments on the main DAM thread indicate the firmware still contains some bugs. So I may play with some of the crowd-sourced filters until Soren has debugged the firmware. BUT my (now distant) impression from going to Paul's filters over Soren's stock ones is that they provided some minor yet worthwhile improvements, but the level of improvement in overall resolution was minor at best.

Finally, I don't know if any of you caught this... In this post:

https://www.diyaudio.com/forums/vendor- ... 41476.html

... Soren mentions that the firmware for the DAM1941 and his other recently-release boards/products with an on-board USB interface go into a FIXED CLOCK MODE when using that interface (and if you did and mentioned it here, I've missed it). THAT could mean (if we lobby Soren a LOT) that could be made a hidden option for the DAM1021/1121 boards, which would open the door to use of better fixed clocks AND source sync mode with local flip-flop-based reclockers. My gut feel is that there's a lot of resolution left on the table with Soren's clock choice(s), clocking scheme, and the FPGA-based reclocking on the DAM1021 (which some comments by Soren suggest was supplemented by off-FPGA reclocking on the DAM1121, but I've never seen definite confirmation). Shall we start a groundswell on this? I'd be willing to pay a reasonable fee for Soren to implement this, perhaps as a 'limited release' or 'undocumented' mod. Anyone else?

Greg in Mississippi

P.S. I forgot to ask, anything else I should consider on improving the DAM1021's? I HAVE thought about going to the DAM1121, especially if he makes more with the 570 clock, but that's a pretty expensive option.

[nige2000]

ive been taking a break from diy electronic hobby these last few months
so im really far from up to speed

things that may help resolution

improve Power supply/ remove all electrolytic and ceramics over 1uf (or swap using low value film <1uf)
replace fpga caps with film
low pass filter, i just removed mine

some sort of signal buffer on output?

think fixed clock option will unlock a door

[gstew]

Thanks Nige!

I didn't mention in my earlier note, but I've gone to LCR polystyrene for the output low-pass filter caps. Nicely detailed, natural, and extended highs with no noticeable emphasis or tizziness. Much better than the Panasonic PPS SMD film caps, MUCH MUCH better than the stock ceramic caps. Plus they come in tight tolerances and are not wildly expensive. I get them from Newark, but they are typically drop-shipped from the UK.

I found no-filter cap too hot and 1200pF still a little too hot. I'm trying some small trim caps... currently have 150pF added to the 1200pF, have 180pF, 200pF, and 220pF here to try if needed.

& I guess I just have to bite the bullet and do the FPGA output caps. I have the 100pF PPS caps here, just have been avoiding doing that.

THANKS!

Greg

[randytsuch]

I found both the FPGA caps and removing the isolators to be an upgrade.

For the isolator mod, I also put a potato FF right next to my dac, as close to the connector as possible.
If you have a nice, well clocked signal coming from next to the dac, the isolators probably add skew.

For the fpga caps, it was an improvement but a pain to install the caps. I got better as I went along, but the film caps are very heat sensitive. Apply heat once, quickly to mount the caps. I would mount a few, then test to make sure the dac was still working. A scope helped to troubleshoot if I had problems, but if you only do a few at a time then you can replace them again, but it will take some time. Buy a bunch so you have extras for the ones you mess up.

Randy

[nige2000]

fpga caps replacement and isolator removal would be a good bet,
maybe some fpga caps that dont melt can be as good
i just accepted i might burn or drop more than a few of the pps

[randytsuch]

fpga caps replacement and isolator removal would be a good bet,
maybe some fpga caps that dont melt can be as good
i just accepted i might burn or drop more than a few of the pps

They are cheap enough so it wasn't a cost thing with me.
It was just finding the bad one if I overheated the things. Normally I touch up solder joints to make them look better, but with these caps I stopped doing that and that helped a lot.

[gstew]

Randy / Nige,

Thanks. I did figure removing the isolators and using a reclocker at the DAC would also make a useful improvement.

One of my Soekris setups is fed from an SDTrans384. I will be using that one to compare the SDTrans to the the EBay cheapie SD card as a source. That'll be the one I'll initially setup with a reclocker.

I have 2 options for reclocking I'll try. First, the bread-board version as well-documented in Randy's blog. Second, I have several Acko S03 boards which are designed as synchronous reclocking boards for USB-I2S interfaces and the BeagleBoneBlack single-board-computer. They do have isolators, but only on the I2S input side and have a direct connection on the output side. As I remember, they use the same or a similar Potato Semi chip for the reclocking, so except for any degradation from the isolators on the input side (which should be much less than having them between the reclocking and the Soekris input), they should be functionally very similar to the bread-board version.

& I'd already planned to mount whichever board I'm trying right at the Soekris input connections, to minimize I2S transmission issues.

My other Soekris setup is fed from an RPi -> Isolator board -> Kali reclocker stack, which is a lot harder to run with a synchronous reclocker. OTOH, Ian Canada will be releasing an RPi-sized all-in-one isolator-reclocker board with replacable clocks in the near future, so that should be able to fairly easily accept a separate clock from a synchronous reclocker.

On replacing the FPGA caps, I DID buy a bunch, maybe 1/3 more than needed. I am getting a lot of experience in SMD builds, so I should be able to do this fairly well... but on the other hand, I also have enough experience with soldering the small PPS film caps to know that I'll ruin 1 out of every 5-10.

I'll be doing this later in the fall, will let you know how it goes.

Greg in Mississippi

P.S. I'm getting some evidence, though I haven't done head-to-head comparisons yet, that the Rubycon PML caps are better sounding than the Panasonic PPS ones. BTW, PML is just Rubycon's proprietary acrylic cap film. I tried to get some Rubycon 100pF's for the FPGA caps, but they were out at Digikey and not carried by Mouser or Newark, as far as I can see.

[Bmxmen]

I have bough New Xmos witch fifo reclocer and BIG OLED display , have connected it to soekris dam1021 dac384khz. Im useing foobar with resampler. Before i nad an amanero 384khz usb i2s source and resampler was seted on 352800khz sound was very good. When i chaned amanero to xmos ani with fifo reclocer trying to set on 352khz it was no plaing, soekris dac dont have i2s connectionn so trying 384khz with same results.only plaing max on 192khz i have seen on display that xmos probybly plays even 384khz but Why soekris dont seen it? Amanero was seen it on 352khz. I have download free newest driver. It is pasible to get from you newier drivers for xmos 768khz? very thanks for helping me. For now my soekris is powered by 8lifepo4bateries.

[Bmxmen]

Hello Again i have modded soekris to Lifepo4 powering but have an problem,dac plays but very silent and dirty. I have disolderet 5v Regs, 3.3v reg and 1.2 switcher, reference 4v opa some of The electolitic caps, one resistor near clock. And buffer. I have Done one mistake,but think it is not big mistake i have soldered in wrong place gnd from shift,clock and fpga. On yellow i have Markt my wrong gnd. But i have changed on upper pad where capacitor have minus. But stil dac plays very silent and dirty. Did i have killed him? I have 4 bateries for each of shift rail, one batery for clock and fpga(fpga have 1.2v reg) , one batery for isolators, two bateries for xmos, and two bateries for oled display. Please help me.
<a href=https://zapodaj.net/a5cb7df0fefbd.jpg.h ... 930.jpg</a>
<a href=https://zapodaj.net/793452d0a2de2.jpg.h ... 819.jpg</a>