Ciao Gaetano, the Soekris manual stated this:
Positive Rail: 0.18A @ 10V
Negative Rail: 0.06A @ 10V
Total: 2.4W
The positive supply draw about 3 times as much current as the negative; the current is
almost independent of input voltage.
for R1 you should use this formula R1=600/CCmA and add the spare headroom.
there are varius build around UltraBiB and dam1021 on the web Dimdim blog as a lot info about.
Positive Rail: 0.18A @ 10V
Negative Rail: 0.06A @ 10V
Total: 2.4W
The positive supply draw about 3 times as much current as the negative; the current is
almost independent of input voltage.
for R1 you should use this formula R1=600/CCmA and add the spare headroom.
there are varius build around UltraBiB and dam1021 on the web Dimdim blog as a lot info about.
Thank you very much for your kind answer! I will also ask some advises to Dimdim. For now, I will try building the power supplies, then I will go ahead!
Ciao, Gaetano.
Measuring at C1 i read exactly 26 Volts. Do you advise me to turn down the output voltage? Is my margin too small?
Its still a viable margin at this current level but try 20V rail output too. It will relief a little M2 heat and will make M1 parasitics less capacitive. Listen to it and if its alright for the phono keep it. Better policy for little lower grid voltage peak demand hours of day too.
Thanks for your advice, i will give it a try. Means also re-adjusting the phono preamp circuit for enough symetrical headroom. The emitter of the telescopic cascode sits at about 11.3 Volts, thats why i was aiming for a rail voltage of about 22.5 Volts.
I dont want to use a 2x12Vac Toroid because i dont know where to burn the excess voltage of about 7V in my phono circuit.
I dont want to use a 2x12Vac Toroid because i dont know where to burn the excess voltage of about 7V in my phono circuit.
oooops, i missed this fact. Back to calculation again. Looks like i will have to order another toroid..
With a 2X9 Tx and 22.5Vout M1 burns 0.7W at 0.2A CCS. M2 burns 3.375W with 0.15A spare at 22.5Vout. With a 2X12 Tx M1 will burn 2W for same CCS & Vout. M2 stays at its previous dissipation. Utilizing 2X12 Tx and a 3.6 Ohm R1 without readjusting the phono seems like the most easy plan. Better balanced for dissipation and voltage headroom. Could possibly sound bit better in the highs too because M1 will drop further its parasitic capacitance. More voltage over it keeps making it faster up to a point.
Salas, to confirm, you've mentioned that with the Ultrabibs it's best to use at least 5cm of 12-14 gauge cable - is that correct? If I can go shorter, is it best to leave the length at 5cm?
If there is a film or ceramic rail decoupling cap located on the client board very near its input power connection better keep the 5cm. If you don't have a scope to confirm the rails are oscillation free with shorter length that is. Oscillation problems are not common at all in the Ubib build reports up to now but you never know with powering various gear.
Thanks. For the befit of others, having compared sets of Ultrabibs with the MUR120s and MSRF860Gs powering the TPA Mercury I/V stage and Buffpro38 I much prefer the MSRF860G. The sound is less bright and more full bass. This has made a much larger difference to the sound than playing with current settings.
Seems a fitting time to ask a question that has been in my head recently.
Regarding power rail capacitors on the input of whatever we may power with Ubib...
I recently read on the Ian Canada thread a member there had removed all the input psu rail caps from his dac and output stage, citing SQ SQ improvements. I assumed he was on battery/ultracap PSU but no he was on 'conventional ' regs....I dont know what.
For instance I was about to change the input rail caps on my dac IV stage as they are 16v and I shall be powering with 15v Ubib eventually. I figured 25v caps would be better. But now I consider removing and not replacing.
What does Ubib prefer to see on its outputs?
Regarding power rail capacitors on the input of whatever we may power with Ubib...
I recently read on the Ian Canada thread a member there had removed all the input psu rail caps from his dac and output stage, citing SQ SQ improvements. I assumed he was on battery/ultracap PSU but no he was on 'conventional ' regs....I dont know what.
For instance I was about to change the input rail caps on my dac IV stage as they are 16v and I shall be powering with 15v Ubib eventually. I figured 25v caps would be better. But now I consider removing and not replacing.
What does Ubib prefer to see on its outputs?
I built 4 pairs of Ultrabibs to compare dual mono - 2 sets with MUR120 and 2 sets with MSRF860G. I built one set with MSRF860G first, then one set with MUR120 and tried dual mono. I felt like I lost a bit of the sound I liked but gained in some other aspects and also used dual 50VA transformers. I then built a second set using MUR120 and tried the universal values recommended by Mark Johnson. Using dual mono with the MUR120 versions, I felt like I lost the sound I really liked in the first iteration, so I can't say whether the snubber did anything for the sound. I recently finished the second set using MSRF860G and the sound is back to what I enjoyed at the outset, but neither of these sets is using a snubber. At some point, I will try the universal values on both sets using the MSRF860G.
I am still playing with current levels, but have one set running at 275mA and one at 325 mA (leftover from using it to power both channels). I believe in dual mono each rail of the Mercury will consume about 90mA and have found R1 values of about 2.2 ohm, 2.4 and 2.5. This gives about 275mA - 240mA. I tried the first MUR120 set at 200mA, but after not liking the sound switched both the MUR120 sets to 300mA. The sound didn't improve, which led me to go back to the MSRF860Gs. I think the current on these can be lowered a bit to get the temps on the tabs to the high 50s Celcius, although these are only mounted to aluminum l bars and will be connected to a chassis at some point to cool further.
I am still playing with current levels, but have one set running at 275mA and one at 325 mA (leftover from using it to power both channels). I believe in dual mono each rail of the Mercury will consume about 90mA and have found R1 values of about 2.2 ohm, 2.4 and 2.5. This gives about 275mA - 240mA. I tried the first MUR120 set at 200mA, but after not liking the sound switched both the MUR120 sets to 300mA. The sound didn't improve, which led me to go back to the MSRF860Gs. I think the current on these can be lowered a bit to get the temps on the tabs to the high 50s Celcius, although these are only mounted to aluminum l bars and will be connected to a chassis at some point to cool further.
Last edited:
Ubib is alright for working with any downstream capacitors given the whole system is still stable, just allow some wire or copper track distance in between. Still, when not really needed capacitors of substantial value exist down the rail, they can partially dominate over some aspects of performance. Slowing down the power interface.
Sometimes extra capacitors are there just as local filters and maybe dispensable but some other times they are there as necessary local decoupling elements. In this case the client board becomes unstable or dirty without them. Nobody except the designer knows exactly why they are there until measuring with them and without them. Watching the rails and signal output for disturbances. Takes mainly a scope and possibly some kind of FFT analyzer.
Only comparing sound before and after to decide is always blind work done. Like finding your way out of a park in the dark. People walk dogs there and you may very well find the gate in the end but you may had also stepped on something they had left behind. 😀
btw. I'm the guy who removed all big caps from the rail with a unexpected good result for the LT3045 by keep only the 4.7µF MKS Capacitor in the rail as shown in the datasheet of the LT3045. In my system this was not only a taste, it was more as day and night 🙂
But the UltraBIB on its way. I'm curious about comparing this LT3045 setup with the UltraBIB. I will let you know the result 🙂
But the UltraBIB on its way. I'm curious about comparing this LT3045 setup with the UltraBIB. I will let you know the result 🙂
Last edited: