Hi All,

I'm starting this thread because I have been able find little useful information on the net about tweaking the Buffalo32S (B32S). There are some threads on diyaudio.com but they all seem to be centered around the development and/or promotion of entirely different DAC's and not how to get the better out of what is already very good. Importantly, I have been unable to find much advise from people who actually possess a Buffalo32S. I was extraordinarily satisfied with the B32S in its "out-of-the-box" configuration and improving it is therefore a difficult and interesting task. I think that the usually kind and collaborative help of Russ and Brian in tweaking and (hopefully) improving B32S is essential as this is their baby and as they know much more about this circuit than anybody else.
My idea for this thread is simply to have a place to show and discuss modifications and implementations of B32S. Not to compare it to other DACs. Not to question its basic qualities. In essence a constructive thread for those of us who already enjoy it and have the passion (or obsession....) to improve it. All of this hopefully with the help and support of Russ and Brian. From experience, I believe that the B32S could be made much more tweaker-friendly without compromising anything in the overall topology of the circuit and I hope that maybe some of the things we do could be rendered more easy to implement in future incarnations of B32S.

As an appetizer (hopefully) I will try to attach pictures of my current, but unfinished, implementation of B32S.

Thanks for attention,

Nic

Output: In stead of mounting the optional gain resistors on the PCB I soldered these onto an Elna 4pole 3way switch and connected this to the PCB (white wires). This lets me vary the gain on back of the DAC (between 4, 2 and 1 Vrms in my configuration). Furthermore, I have two sets of XLR outputs: one for headphones and of for my pre/power amp. I have the traces cut on the PCB (according to the manual) but the wiring between the headphone XLR's and the gain setting resistors (mounted on the switch) by-passes this cut and restores the low-impedance output. These modifications were only done to increase the versatility of the final implementation of the DAC and I have not noticed any negative effect of these mods on the sound.

Power supplies: The B32S is very good with the TPA power supplies and changing these was mostly done out of curiosity as I have experienced very significant improvements of other DACs improving the power supplies. I first changed the PS for the analog sections with two sets of Jung super regulators (JSR03/05 from Sjostrom audio) set to +/- 15V. This, I think, improved the sonics significantly in a positive way, meaning that I have tried to switch back and fourth between the two PSU configurations and that I definitely like better the JSRs. The B32S with the TPA LCBPS is already very good so the improvement is, however, little in absolute terms. This is not a good bang-for-the-buck upgrade if you already have (dual) LCBPS but should be tried if you want your B32S at its best. I seem to recall that others have had similar experiences and with Russ and Brian coming up with the Placid supply, I think that there is a fairly wide consensus that there is something to gain by upgrading these supplies.
I also changed the supply for the digital section to a 5V JSR, but even if I changed back and fourth I could not hear any significant difference between the two. I think that that either there is little or nothing to gain in improving the supply for the digital section of B32S or that the 3 on-board Vregs fed by this external supply is in some way limiting the potential beneficial effect of improving the off-board supply. The logical next step was therefore to try to replace the the three on-board regs (two 3.3V and a 1.22V reg). At the moment, I have replaced the two 3.3V regs with shunt types from Paul Hynes, I still need to do the 1.2V reg. I think the change has a quite a significant beneficial effect but as the tweaking, the way I did it, is kind of irreversible it is difficult for me to do A/B comparisons (well unless I buy a second Buffalo32S.........). I would like to underscore that the risk/benefit ratio of replacing these regulators is quite high - at least for people like me with medium-to-low experience in smd soldering. It would take very little modifications to the B32S pcb layout to greatly facilitate the testing of alternative power supplies for the digital section of B32S......

If other Buffalo owners should be interested I will be happy to provide further details of what I did.
Thanks,
Nic

Edited by user Saturday, September 5, 2009 1:49:16 PM(UTC)  | Reason: Not specified