They are on internal planes. I just forgot to run an anti-trace around the PLL section of the power plane. The FPGA is a BGA and all of the power is straight up thru vias right next to the appropriate pad.
3 Likes
Ted, may I ask what these two resistors that you recommend upgrading are doing in that circuit? I donāt know if a schematic has been posted or not, but Iām guessing that one might be the input resistor and the other the feedback resistor for the op-amp?
Also, Iām assuming that 16 resistors total should upgraded (two for each op-amp), and that 8 of these should be bypassed by a ~50uF cap. Is that correct?
The Vocm mod is modifying the resistor divider that sets the output common mode of the video opamps. This is, in essence, the ground reference for the outputs. Thereās resistor divider for each opamp. There are two caps on the bottom resistor, but they have a SRF near 20MHz. Adding some caps to stabilize the Vocm reference in the audio frequencies is very useful. The input impedance of the video opamps in the DS is much lower than I thought when I designed the circuit, so a stable input is important. The video opamps in my prototype and the TSS and Mk II have a much higher input impedance (like youād expect with an opamp.)
The schematic has been posted multiple times, but the search here isnāt always helpful.
4 Likes
Ted, what is the total voltage across the divider which provides the Vcom to the op-amps? And is there some reason that people were suggesting using a non-polar cap for this application? Since this is a DC divider, using a polarized cap seems appropriate to me.
7V. Youād have to ask them about non-polar. Thereās no reason not to use a polar cap and thatās what I do in other DACs there.
3 Likes
Thanks. Iām assuming that the lowest ESR is the most critical specification for this application?
Thatās my choice.
3 Likes
Do you happen to know the power rating of the stock resistors? I need to make sure that the new resistors are properly sized for the board.
I usually use ERA-6AEBā¦V for anything near the audio.
Ā±0.1% 0.125W, Ā±25ppm/Ā°C, 1/8W Chip Resistor 0805 (2012 Metric) Automotive AEC-Q200 Thin Film
Obviously since essentially no current flows, this is overkill.
3 Likes
I assume all of you Hot Rodders are aware of this parts source, but just in case:
2 Likes
You can add this one to the list as well (in the US):
2 Likes
Could you snap a picture and let us know about the challenges when you are finished so we can see the changes? Also let us know how it sounds 
I might try some of these if they are simple to do. Iāll have to get my Modder to do them. It looks like some of the components are pretty big to put on a SMD board.
Thanks for the synopsis of your changes. It lays them out pretty clearly. It puts together some of the things weāve learned from Ted and the Modderās over time.
-Alan
My opinion, and trying to help based on actual listening and experience.
C10 SMD PPS is likely already sonically superior in this case to the big Cu foil cap. But as well in this application C10 needs to have excellent high freq characteristics (way above the audio band) which means lead length, construction, etc matters.
And youāll have āsome difficultyā fitting and attaching large through-hole parts for SMD C10 and R9/R10 
If you locate them remotely somehow and use small jumper leads to the pads, that also alters the sound.
Recommend- use instead a nude z-foil SMD 0805; theyāre also non-inductive constructed. Same high freq issues applies as above.
Vocm resistors- plan for 499R (common value) rather than 500R (hard to get). The stock value is 499R. If you want the quietest, most well behaved you can find there use a nude z foil.
C704 is not installed on the stock DS. Probably for a good reason Adding one likely will not improve things sonically IMO. However if you insist, use an SMD PPS and you likely wonāt harm the sound much either, which is a good thing
Hope that helps someone. T
Interesting. I have 2 DACs here with no C704 installed. IMO, just leave the PPS. T
I had it put in. It gave a better high-end and more air.
-Alan
Jeff, the owner, is a great guy. Iāve done business with him before. He upgraded a pair of speakers I had by redesigning the crossover and using higher quality parts.
Yes he is, and super super smartā¦ T
I would stick with Panasonic if I could as that is a known (for me), but would hope the Wima are ~ equivalent. They do appear to be promoted quite a bit lately. T
Great that you start to document this!
It would be very helpful if we finally had a description or picturing of commonly agreed to doās and documentation of parts used for the Ted approved VOCM mod that is clear enough for non EEās to be handed over to an experienced modder who didnāt read this thread.
I assume weāre not at this stage yet.
Yeah thatās much better I thinkā¦
Re: 4), Iām not sure where you plan to fit these huge films. And I donāt think they are necessary to achieve the goal on Vocm. It can be accomplished with tantalum or even electrolytic in this case. Regarding using the electrolytic, Vocm already has 2 x 1uF local high speed bypass as well (C9 and C?).
The resistors are excellent IME but not cheap. Most people have a limited budget for these things. In light of that, thinking about which ones should give the best bang for the buck. Just some personal thoughts here. Weāll use the same old image again for illustration.
Based on experience with high open loop gain/closed loop amp circuit, and letās say if I could only change 2 pairs of R, think I would consider putting the money into R7/8 and R3/4. These are feedback network resistors; they also handle our signal directly. Feedback R in the traditional sense that is; in this case things are complicated a bit by the additional R1/2 at the very front (MFB topology). Which also handle our signal directly as well. In light of that Iād probably pony up for all 3 locations.
In a traditional 3 stage power amp for example, R7/8 would be about the most audible in the system for example. So much so that it is critical to optimize this R. And most commercial designs I have seen make no extra effort to do this IME, ie they basically donāt know or regard it as a donāt care. However the connection Iām thinking of is a bit different so maybe not a direct comparison really.
There is another location that is even moreso IME, in fact itās a make or break location. Even more than say, the R in a passive I/V converter at the head of a signal chain. Ie, its a resistor noise amplifier basically and if not optimized with the very best it stops the whole show basically. Iāve seen designs with basic $0.10 junk R there, and thatās how they sound. The designers simply donāt know. However, that location does not apply here thankfully.
The R5/6 here are a simple voltage divider circuit off to the side that creates the reference voltage Vocm, and they do not handle our signal per se. And the more bypassed and quieter we make the Vocm node (via the C across R6), the less wiggle there should be on that node and even less so critical then I would say. I would not expect these R to have as much impact as above; however Ted did say they were able to hear R changes at this location as well so they are audible.
Regarding resistor quality, I know from experience that the Y series Z foil generally slaughter most through-hole. But not so sure where things rank vs a good SMD thin film. I have not personally conducted an extensive comparo of SMD R. But I think in general a good thin film SMD is quite good in comparison to most through hole.
Anyway just wanted to share these thoughts about trying to improve the sound. None of the above discussion is guaranteed or set in stone and YMMV. T
1 Like