Did you supply the transformers?
I supplied them. Here is the Edcor link I saved from Ted Smith since I can never find them!
1 Like
Ted, in the interest of trying to reduce the rework burden for some component testing and such Iām thinking of temporarily dropping down to (1) or 1 pr of OA per channel. Thinking could still do comparative testing that way with alot less rework required.
Thinking we could basically just lift the input and output resistors (and the output differential C) to disable the OAās. BUT am thinking about possible transmission line issues for the bitstream lines; so thinking would probably use the outer OA pairs per channel (to end-terminate the lines), and disable the inner OA?
Thanks, T
If you really want to only modify one opamp per channel you should go back to Snowmass or earlier. Starting in Windom I drive the opamps a little differently and need at least two opamps per channel at times.
With Windom and beyond you can skip two opamps per channel (LA2, LA4, RA2 and RA4), but not three:
You probably want to leave the opamp inputs connected so the input drivers have the correct terminations and so the opampās inputs arenāt floating which could lead to oscillations or extra current draw. (Those terminations rely on the opamp driving its inputs together.)
Lifting R9, R10 and C10 will keep the unmodified opamps from messing with the output.
1 Like
Do you need one or two of the transformers for the mod?
Two transformers.
2 Likes
Yes, two transformers.
1 Like
Ok, thank you! T
Take a look at post 2610. It costs little and can be applied quickly. It gives a much greater improvement in sound than replacing the output transformers. There is a lot of theoretical considerations about modifications, and a tremendous fear of touching analog PCB.
Your 2610 post has great detailed explanation of the MOD. I wish I have the gut to try it but the last time I tried the soldering my parent had to stop me from burning the house down. I may have been in middle school?
More likely I will be continuing enjoying my ābasicā DS (enjoy it greatly) until the new one comes out eventually. Thanks for all the pointers 
Donāt feel bad, me neither hahaā¦ 
1 Like
I didnāt quite follow the OA numbers you specified for skipping. Seems to me I could use the inner (nearest to the re-CLKer) pairs RA3, RA4 and LA3, LA4. No?
And they do seem to be anti-pairs, such that the +/- outs are on opposite sides of the output cap.
Per channel: The two inside opamps are on one reclocker output and the other two outside opamps are on another reclocker output (see the traces in the picture above.) So you can either do (1 and 3) or (2 and 4).
1 Like
So I cant use just one re-clocker on a pair of OA?
With Snowmass or earlier you can do whatever you want. With later releases you want to keep at least one opamp per relocking pair (per channel.)
1 Like
Ok thanks, Can you help me understand why? Thanks, T
I vary the bits between channels slightly at times to get a better sounding signal. Either side by itself will be a little noisier than both together. I didnāt have this feature before Windom.
1 Like
Apologize in advance, this will sound like heresy to some. But Iāve changed the OA circuit (radically) to get ~ perfect step response. Iāve also brought in some tantalum resistors, at the input on the low side of the feedback leg which I believe is likely the most critical R in the circuit (based on past experience with power amplifiers). Iām doing a functional listening test after modding 1 OA circuit per channel. The other (3) per channel have their outputs disconnected from the output nets.
And even though only 1 OA per channel currently it sounds very encouraging so farā¦ Iāll scope it tomorrow and make sure all is well, then do the 2nd OA per channel per Tedās advice.
Whatās driving this super deep dive is a somewhat oddball DAC that dropped into my life 2 weeks agoā¦ It has changed everything, rocked the foundations down to the philosophical level even, so to speak. Weāre talking the sound of partsā¦ the sound of wireā¦ Germanium transistorsā¦ Valvesā¦ Etc.
Can we get the DS to do similar greatness? Weāre trying to seeā¦ Iāll try and link a clip of the test so far. PS- none of this is endorsed at all by PS Audio, of course.
Link:
Lots of fly specks that used to be in the OA circuits 
Yes, weāre working hard here.
Man straight up, Iām hearing some stunning music from this thing this am, just streaming Pandora 
Music, music, incredible music de naturalā¦ Clarity, transparent as h-ck with NO distortion(!) It has also stepped out of the back row and no longer somewhat too āreticentā; seems as forceful and up front as the music asks. Airy, spacious as h-ck. Its been freed to really sing if you will, the sonic boot is lifted off the neck. Wow!
It has never played even close to this level before. Itās obvious it really liked what weāve done so far thank G-d 
We have our answer it seems.
With that tremendous encouragement weāll try to complete the 2nd OAās and get the PSU stuff re-installed. Iām not sure how much weāll get completed this week though as we are headed to Axpona (woot!) in the pre-dawn am Friday. Wayy too early for this kid lol but direct flights at least.
We also need to move to the pure Cu amorphous transformers that arrived last week.
Iām also thinkingā¦ maybe wire each of the (2) transformer primary coils to a separate OA output. The 2 OA in a pair will no longer connect to each other then, each OA will drive a primary coil separately. Hmmā¦ 
And think weāre also going to disconnect the OA from the board output nets and wire straight from the OA to the transformer, and from there to the XLR out using a special treated wire. Itās probably the best I know of; noticeably more clear/natural than others. Similar manner as Audience AU24SX interconnect for example. We may strip the insulation and route through teflon tubing; weāll see.
Iām excited, I think itās going to be truly spectacular T