Salas BiB regulator
These are absolutely excellent regulators with very low noise (measurements have been published from multiple sources), and output impedance, when implemented properly. But they do need to be tuned appropriately for the circuit which they are powering. In my experience they do not like to see a lot of capacitance on their output. Designer Salas is highly responsive to folks using them, and will help in tuning the regulator to work with various circuits via the appropriate DIYaudio threads.
When trying to make a measurement of output noise using an ordinary DMM, you are not going to be able to measure the actual noise level as if the regulator is working correctly the noise will be at the µV level. Short your DMM probes tother and see what it reads set to AC... my Fluke, for example never reads lower than 3.5 mV in the presence of nothing. If Sidiropa's DMM reads 10 mV or just slightly under (9.8-9.9 mV) it is likely the regulator is working properly, not oscillating, and has very low noise (µV levels), but if his DMM reads much lower when shorting the probes (like mine at 3.5 mV) then the regulator is still oscillating reading 10 mV. My Salas BiB regulators all read around 3.6-3.5 mV AC using my DMM (same as shorting the probes) when powering the Legato (with stock 100 µF caps replaced by .1 µF films).
Sidiropa, what does your DMM read for AC voltage when shorting the probes? If it reads right around 10 mV then you are good with the caps removed (except you may want to get confirmation from Russ on the advisability of this in his circuit, taking into account the very low output impedance of the Salas regs and the use of sense wiring), but if your DMM reads much lower, then you still have an unstable regulator and the next step would be to consult with Salas on the BiB threads.
Of course if you know someone with a decent 'scope who knows how to use it, it would be nice to use that to check for oscillations directly, as this would leave out the imprecise measurement ability of the DMM and guesswork.
These are absolutely excellent regulators with very low noise (measurements have been published from multiple sources), and output impedance, when implemented properly. But they do need to be tuned appropriately for the circuit which they are powering. In my experience they do not like to see a lot of capacitance on their output. Designer Salas is highly responsive to folks using them, and will help in tuning the regulator to work with various circuits via the appropriate DIYaudio threads.
When trying to make a measurement of output noise using an ordinary DMM, you are not going to be able to measure the actual noise level as if the regulator is working correctly the noise will be at the µV level. Short your DMM probes tother and see what it reads set to AC... my Fluke, for example never reads lower than 3.5 mV in the presence of nothing. If Sidiropa's DMM reads 10 mV or just slightly under (9.8-9.9 mV) it is likely the regulator is working properly, not oscillating, and has very low noise (µV levels), but if his DMM reads much lower when shorting the probes (like mine at 3.5 mV) then the regulator is still oscillating reading 10 mV. My Salas BiB regulators all read around 3.6-3.5 mV AC using my DMM (same as shorting the probes) when powering the Legato (with stock 100 µF caps replaced by .1 µF films).
Sidiropa, what does your DMM read for AC voltage when shorting the probes? If it reads right around 10 mV then you are good with the caps removed (except you may want to get confirmation from Russ on the advisability of this in his circuit, taking into account the very low output impedance of the Salas regs and the use of sense wiring), but if your DMM reads much lower, then you still have an unstable regulator and the next step would be to consult with Salas on the BiB threads.
Of course if you know someone with a decent 'scope who knows how to use it, it would be nice to use that to check for oscillations directly, as this would leave out the imprecise measurement ability of the DMM and guesswork.
Last edited:
Just saying - if that measurement is accurate (I can only go by what I read) - then the regulation is awful... not necessarily because it is a bad design - but well because it appears to be working awfully in the situation it finds itself. 
The Mercury is a capacitive load - you are likely best off with a regulator comfortable driving it - where it comes from - I don't really care - but I make a couple great choices!
I meant no offense to the circuit's designer.
Cheers!
Russ
The Mercury is a capacitive load - you are likely best off with a regulator comfortable driving it - where it comes from - I don't really care - but I make a couple great choices!
I meant no offense to the circuit's designer.
Cheers!
Russ
No worries Russ! I am sure Salas would not be offended, as his regulator designs are actually excellent (and have been independently measured to be so), they just have to be implemented properly, and as I mentioned, he is very responsive to inquiries about how to tune them for different circuit applications.
My best regards to Salas, but in any case - let's please not be debugging third party regulators in this thread - it is not the appropriate place.
Not because I don't wish the user to get help - I just want the right audience to hear the questions. At the same time - I don't want those interested in the Mercury to somehow get he impression it is affiliated with someone else's regulator.
Cheers!
Russ
Not because I don't wish the user to get help - I just want the right audience to hear the questions.
At the same time - I don't want those interested in the Mercury to somehow get he impression it is affiliated with someone else's regulator.Cheers!
Russ
Sounds like you have received yours? If so - I would really love to hear what you think of it. I know what I think - but I am not very objective.
Yes! Got it assembled and ready to go and had two days to listen to it. Basic build here. BIII 9028 Pro, LCDP HO for DAC/USB input, Placid HD BP 2.1 for I/V board, Amanero/Cronus. Still waiting for 4.3" tft to see if my recently assembled DimDim shield works so running on onboard controller/dipswitch. Went with Apodizing Fast Linear Phase filter.
I started out with my old faithful Legato. I run balanced out with ground lifted on Legato so no common mode DC between ground/pos and ground neg. No coupling caps needed. Sounded better than the 9018 so I slipped Mercury in. Due to higher output of Mercury, I used Radio Shack to keep listening level the same. Wow, it's like a different DAC. More detail and separation IMO. I'll be retiring the 9018 with Legato and investing in a new HiFi 2000 chassis for my new DAC! High marks Russ!
Like Russ said, extremely low DC between balanced pos/neg and between ground pos/ground neg. I was able to connect the balanced ground again. It helps eliminate spikes when switching to/from digital source.
This DAC won't be as extreme as my first BIII...
Attachments
Last edited:
Man that's going to be a fantastic DAC! Nice work! Thanks for the report.
Cheers!
Russ
Believe me - this is by design - I have been working on this output stage for more than a year - and it started out discrete... Mercury is essentially Counterpoint but using OPA1632 where I have been using discrete components. It sounds as good as it does - because I was meticulous about designing and testing it... over and over and over....
10-4
And hence my advice to post the questions in the appropriate other forum threads.
My best regards to Salas, but in any case - let's please not be debugging third party regulators in this thread - it is not the appropriate place.
Not because I don't wish the user to get help - I just want the right audience to hear the questions.
Cheers!
Russ
And hence my advice to post the questions in the appropriate other forum threads.
Hi Russ,
Would you be willing to talk a bit about the details of the analog filter circuit? Is it accomplished by just the two 1nF film caps in association with the present resistance of the circuit, or is there more to it? In the first post you mentioned filter corners of 20 K or 30 K, is that correct (seems very low?), and what is the corner frequency with the included parts (1nF)
Thanks for any information which you care to share.
Would you be willing to talk a bit about the details of the analog filter circuit? Is it accomplished by just the two 1nF film caps in association with the present resistance of the circuit, or is there more to it? In the first post you mentioned filter corners of 20 K or 30 K, is that correct (seems very low?), and what is the corner frequency with the included parts (1nF)
Thanks for any information which you care to share.