Valab NOS DAC - subjective listening and modifications - Page 106

I think he thought you were talking about bypassing the output caps?

I for one was talking about bypassing the DAC caps not the output caps, bypassing the Dac bias caps with a styroflex or replacing them with somethig decent made quite a big difference in my previous dac with the 1541 gold crown.

The best results with bypassing imho are to be had when you connect the bypass between the most central spot of the common ground and and as close to the DAC pin as possible.

BTW ; Styro's need a good burn in as well, and the sort of styro makes a lot of difference!

No one got a schematic of the Valab?

Well just to add to the confusion, I was talking about bypassing the reg that supplies the vref to pin 7 of the dacs, I was hoping to get a cleaner supply there.

And, at the same time, I was asking if the 0.1uf smd were necessary if you already have good larger coupling caps on the dac chips.

if you have a better cap on the vref you would not need the smd, yet the larger the cap gets the worse the noise filtering gets at higer frequencies and for that you need bypass caps...
0.1 smd is quite large already if we're talking about bypasses.
Given that you have something like 200 uf on the Dac vref you'd need something much smaller to decouple (or filter) high freq noise, typically one fifth to one tenth of the largest cap (in several steps) is what seems to work best.
So 200uf (or thereabouts, not hard science) combined with 20 uf, with 2 uf with, 0.2 uf with several 100 nf in ever increasing quality caps would be the recipe (electrolytic, MKT, MKP, MKV, Styro, Silver/Mica) that would be ideal as each dielectric has its own specific frequency range.

Now granted that space is limited you might want to cut down a bit on the number of bypass caps and only use the smaller sizes.

Just play around with caps this way, you'll gain a lot of insight.
I personally do not like this in the signal path as the signal somehow gets smeared but for DC stabilizing/decoupling this works OK imho.

The stock 10uf and .1uf smd are across the (Gnd, Vdd) pins 4,5. The vref pin 7 doesnt have a cap, (it looks like it does, because the .1 smd is next to pin 7) it was the voltage supply from the LM334, thats the one I was trying to cap.

There isnt a LM334 in that position on dacs earlier then v2.5, they used the ground differential to provide the vref. But the Chameleon uses the LM334, so might be gains to be had there if the supply is cleaner.

However thanks for the info, I did also want to know why we need the .1 smd if we have better qual caps replacing the stock 10uf, and you have answered that, with a lot of nice freeby info as well

And if you have any ideas on how we could provide a cleaner voltage to the vref please share, I think this might help lower the noise floor.

Cheers

PRP vs Texas

Just started burning in the PRP on the IV, will give it a couple of days then swap back the Texas.

Initialy, they walk all over the Takman carbon comps, which arent bad at all.

Very transparent, full extention of frequency range, clean and warm, without being warm, and fast but relaxed?!"&*^

This might be closer then I expected, battle of Titans.

hmmm, wait for the edge, I found them a bit fatiguing personally

I've had the Takman carbon film in for a couple of weeks, initialy bright, but still lean in that direction even now, not unpleasant though. I expected the PRP to be more balanced, and perhaps more neutral like carbon comp and Texas, perhaps somewhere between the two.

Well so much for expectations...

They sound more like a punchy, lively version of the Texas, but not coloured as indeed the Texas are def not. But when they have to be gentle they are, this contrast/nuance shows up on even on a single note, more then I can remember hearing with the Texas.

Pat what would you recommend for min burn in for the PRP? I will then do I direct comp to my Texas. I will look out for subtle differences like fatigue and edge as qusp has suggested.

I've got more texas on the way 175r, and some 180r prp for the Chameleon here already so I can listen with my ears and not my wallet.

Looking at the schematic it appears that benefits could be had from applying a small bypass (several 100 pF or nF styroflex for example) between the source of the Vref at the LM334 and pin 7 of each DAC and likewise (probably more important) between the 7.6 V and pin 5 of each DAC.

Not sure but it appears that the LM334 has no cap whatsoever in the current design, looking at the spec sheet it may not be absolutely required but I am wondering whether anyone can shed some light on that topic.

Reading up on the current scene of IC regulators I read in thespec sheet for an ultra low noise version that tantalum works better for IC voltage regs due to the usual ESR of those.

Hey there, can see the bypass between Vref and LM334, but I'm not clear (read, too dumb) about why between 7.6v and pin 5?

Yes it certainly doesnt look like the LM334 has a cap for some reason, I'm sure that voltage can be smoothed more if it did. Just need to locate right spot on the board.

Pin 5 of the TDA 1543 receives the 7.6 Volt supply, the crap picked up on the the way from the PSU to the DAC can be attanuated by attaching the small Capacitor directly to pin 5.
Not sure whether the LM 334 will benefit from any cap, it is just too new for me ;-) it is a current source not a voltage regulator so ripple may not be an issue(I was more thinking along the line of filtering any high freq digital noise possibly caused by the thing itself).

BTW anyone know a source for those PRP resistors in Europe? I only found a US supplier with hefty shipping prices, and I do not want to use my US collegues as freight station too often ;-)

partsconnexion

Oh well thats canada actually, havent seen any europe ones