My test setup has a massive ground loop, being based on one soundcard. As a result all amps have 50Hz + harmonics, and I am unable to get the noise floor below -95dB even if any of the tested amps are actually better than that.
Even so, the Mullard returns a stellar result, and bears double checking. Will post later when and if I have the ground loop sorted out. The competitor was Juma's BJT + Hexfet design based on the F5, but this first cut execution was running bias of only 500mA.
Deferring to the needs of the measurement suite, I have not posted images of the results. They simply look ugly.
Even so, the Mullard returns a stellar result, and bears double checking. Will post later when and if I have the ground loop sorted out. The competitor was Juma's BJT + Hexfet design based on the F5, but this first cut execution was running bias of only 500mA.
Deferring to the needs of the measurement suite, I have not posted images of the results. They simply look ugly.
Attachments
Thanks for the update - THD looks great, as predicted by simulation. IMD is not bad either - in fact, very good. Both are superb for 0.5A output-stage bias. The Darlington or MOSFET versions with IRFP150 or similar looks very workable and should have a bit more durability.
I need to get back to this in due course, but I was distracted by lots of other stuff on my plate, including the Tymphany drivers, Tabaq cabinets, various discrete opamps and so on.
Incidentally, this should be great for driving full-ranges like the Tabaq or near-field bookshelf setups.
The 500mA bias figure was for the F5/BJT design. I suspect at full bias I can get it down by at least 6-7dB. I am facing some difficulties in getting to bias correctly, and we already have ambient of 37 degrees, that is not helping. Also since it has extremely low gain I need to drive the soundcard at -1dB, which is causing it to lose performance.
The Mullard was tested at -12dB input level, which massively reduces the distortion from the soundcard. The problem is that RMAA expects a minimum level at the input, and I was not able to reach this level using my jig (10dB-50dB attenuation).
The Mullard was at full 1A bias (1.1A, actually) and 20V rails, and it's a monster task to keep it cool. Two channels puts 80W into a single sink. I still manage to keep the device spreaders below 65 degrees in this weather, but it's probably going to end up as a winter amp.
It also works remarkably well with regular speakers, I'm using it on the Usher S-520 and the match is stellar.
The Mullard was tested at -12dB input level, which massively reduces the distortion from the soundcard. The problem is that RMAA expects a minimum level at the input, and I was not able to reach this level using my jig (10dB-50dB attenuation).
The Mullard was at full 1A bias (1.1A, actually) and 20V rails, and it's a monster task to keep it cool. Two channels puts 80W into a single sink. I still manage to keep the device spreaders below 65 degrees in this weather, but it's probably going to end up as a winter amp.
It also works remarkably well with regular speakers, I'm using it on the Usher S-520 and the match is stellar.
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.