Correct.
Correct
Also correct.
Every situation/listener is different, but I have not. But I have older ears. For te record, I use a single Placid HD BP supply.
Typically, when people want to use a full dual-mono configuration, this goes all the way back to individual transformers. The idea is that one channel would have no impact on the other, so full separation. We allow for that, but I personally don't do it. To each his own.
Depending on the DAC output level, and gain used, it is safe to plan for 500-600mA per rail (+/-) to allow for ample headroom.
There is also the "pseudo dual mono" approach to powering the Mercury. I did it this way:
Single tranny, two secondaries (100 VA in my case, but 50 would probably be enough), then I rectified the two secondaries, and sent one to two, biggish, 4 pole caps. Each cap output feeds a single regulator for two - rails and two + rails. The four pole caps provide a fair degree of isolation, not the same as having two trannies, but close, and easier to implement.
Honestly i cannot say I have ever heard a benefit of dual mono supplies vs. either "pseudo dual mono" or true dual mono as long as the power supply has low output impedance. Theoretically sound staging and perhaps dynamics would be better with dual mono, but I am not so sure it matters much even in the context of a very high performance system.
Single tranny, two secondaries (100 VA in my case, but 50 would probably be enough), then I rectified the two secondaries, and sent one to two, biggish, 4 pole caps. Each cap output feeds a single regulator for two - rails and two + rails. The four pole caps provide a fair degree of isolation, not the same as having two trannies, but close, and easier to implement.
Honestly i cannot say I have ever heard a benefit of dual mono supplies vs. either "pseudo dual mono" or true dual mono as long as the power supply has low output impedance. Theoretically sound staging and perhaps dynamics would be better with dual mono, but I am not so sure it matters much even in the context of a very high performance system.
Brian, per the below, if you're using a Placid BP you're saying the supply should be set to deliver at 600ma per rail? How much should it be shunting?
If going dual mono, you would then set each rail of the power supply to deliver 300ma (in which case you would be delivering 300ma to L+, 300ma to L- and same for R+ and R-)?
Separately, if you wanted to try changing the 330uf caps on the Mercury, what are the required specs for replacement caps? Many of the oil/film caps have much lower uf so I'm not sure how far you can stray from the specs of the provided caps.
Thanks.
If going dual mono, you would then set each rail of the power supply to deliver 300ma (in which case you would be delivering 300ma to L+, 300ma to L- and same for R+ and R-)?
Separately, if you wanted to try changing the 330uf caps on the Mercury, what are the required specs for replacement caps? Many of the oil/film caps have much lower uf so I'm not sure how far you can stray from the specs of the provided caps.
Thanks.
Last edited:
Although very late to this party, I am in the process or making identical, Mercury-containing DACs with BIIIse boards having all three ESS chips! Thus, a quick read through this thread for hints, etc.
Among early observations, I was surprised by the saturation of transients with 9038/Mercury! Sometimes I wanted the drummers to just chill a bit.
I dialed the Mercury supply voltage down to 14.5v and was a little more comfortable. Bass resolution is unprecedented! Early in the process, the BIIIse boards are running with default on-board firmware. I can hardly wait to dig into the control registers of the 9038, and let that Mercury fully 'burn-in'.A bigger surprise has been how delightfully musical the older 9018s sound! ...especially with >48kHz PCM. I'm experiencing absolutely zero listening fatigue with the 9018/Mercury. Those 9018s have already made a ton of enjoyable music, and now with the Mercury I/V they will be back in service for family members.
I particularly value the fact that Russ and Brian's components are upgradeable in this way. Plus, the high quality of TPA designs is providing great service longevity in addition to excellent sonic performance.
Frank
I wanted to confirm how everyone is thinking about the current consumption of the Mercury. Brian has mentioned to plan for 500-600mA per rail to allow for ample headroom but I've also read that actual current consumption is far less (and the legato only drew about 350ma TOTAL).
I'm going to try different shunt power supplies and need to set the current to allow for approx. 100mA above peak consumption.
If I'm going to bridge the L/R power inputs, should I be planning for say 550mA + 100mA of headroom, so 650mA total (shunting 100mA) for + and also 650mA for - ? If I am running dual mono, then I would halve those numbers and provide approx. 325mA to each of the L-, L+, R- and R+ rails. Am I thinking about it correctly or do I not need that much current cushion?
Similarly, if I'm going to run a BuffPro28 with Legato, it seems I should be planning for approx. 200ma per rail, but this seems too low given Brian's recommendation for the Mercury, so I am not sure.
https://www.diyaudio.com/forums/twi...achieving-escape-velocity-21.html#post6053184
I'm going to try different shunt power supplies and need to set the current to allow for approx. 100mA above peak consumption.
If I'm going to bridge the L/R power inputs, should I be planning for say 550mA + 100mA of headroom, so 650mA total (shunting 100mA) for + and also 650mA for - ? If I am running dual mono, then I would halve those numbers and provide approx. 325mA to each of the L-, L+, R- and R+ rails. Am I thinking about it correctly or do I not need that much current cushion?
Similarly, if I'm going to run a BuffPro28 with Legato, it seems I should be planning for approx. 200ma per rail, but this seems too low given Brian's recommendation for the Mercury, so I am not sure.
https://www.diyaudio.com/forums/twi...achieving-escape-velocity-21.html#post6053184
I have a 9038/Mercury running on the bench so I just measured. When idle, Mercury is drawing 171 and 176mA from the two rails. At full output into SE 300ohm headphones, Mercury is drawing an additional 15-16 mA. My shunt current then dips from 90 to ~75mA (low resolution time scale on multimeter). Signal is some very bass-heavy electronica that would be very damaging if I ever put those phones on my ears!
The B3se9038 Centaur linear power supply is also driving another board so I won't measure, but with extreme signal the Centaur heatsink is touchable for 10 sec.
Frank
The B3se9038 Centaur linear power supply is also driving another board so I won't measure, but with extreme signal the Centaur heatsink is touchable for 10 sec.
Frank
Thanks and that makes sense. In this case, since I will be setting a fixed current limit on a shunt regulator, I wanted to make sure I won't be shunting excessive current. It seems the Mercury generally draws about 170mA per rail, doubling would point to 310mA. If I want to shunt 100mA, I'd set it up to deliver approx. 410mA but if I'm using 5-600mA as guidance, I would set it up to deliver 6-700mA per rail and could be shunting 500mA. That being said if aiming to deliver 5-600mA per rail makes sense I'll set the power supply up that way.
I have the Placid HD bipolar supply set up to deliver about 360mA per rail and I read it shunting about 170mA. When setting up a different shunt supply and using a fixed resistor to set the current limit, it seems setting the limit at about 360mA, or even slightly lower, would work. Am I thinking about this the right way?
Just remember that however you set the shunted current, that goes to warming the heat sinks, which are less efficient in a chassis than in open air. I’d be curious where Russ or Bryan set their Mercury supplies. With Legato I couldn’t hear any difference between 75mA vs. higher shunt amounts so 80 was my cold setting.