I guess this could look fine if made accurately. Especially when using husks for the drilled wire ends. This is at least my assembly plan for it...
I've just fitted the pieces of the "SiC" puppy to a piece of pine board - with 3 transformers, this is one amp that won't be walking across the table when the bass gets heavy. It'll sport a Corian top plate (faux brown granite) to keep inquisitive fingers out of the high voltage. Tomorrow night I'll finish the power supply, and sand and paint the base board satin brown to complement the top. Having leaned my lesson with "Le Mutant", I've positioned the amp modules so that the hot air exhaust gets ejected immediately out into the open and doesn't linger around. This one also looks like it'll make it to BA. I have somewhat higher hopes for the sound of the Puppy as compared to the other two Passish amps I'm bringing, but comparison with a good source and speakers will tell...
So sad, I just finished getting the SiC Puppy up and running on its pine board. The sound is crisp and incisive (my favorite of the 3 Passish amps I've built thus far), but overlain with a patina of hum. Whether this is spray from the power transformer, power supply hum (should not be an issue with capacitance multipliers, but I need to check), or some snarky oscillation that just raised its head is unknown. Running the amp on an electrophoresis supply appears to kill the hum, so it's probably something to do with the supply. Anyway, as I really don't feel like hauling the electrophoresis supply to BA along with all the rest of the crap I'm bringing, this boy stays home until he's well. Too bad, as he may be my favorite child of the projects I've been working on lately. Le Mutant and Half Nelson are working, so they'll make the trip.
I'm really interested in this thread and where it's going. I'm a tube guy and I can relate to this, though my ss knowledge is poor. I'm wondering about putting chokes in the power supply if there's a hum problem - tube technology! Also wondering about a tube front end with interstage.
I've been looking at Edcor transformers - they do 150, 300 and 600 ohm primaries. The Lundahl LL1627 could be used as well - that's a lot more expensive. Actually the Edcors are substantially larger than the UBT1 and Lundahl, which both weigh around 2.5 kilos. The Edcors are more like 4.5 kilos and rated for 400ma. I imagine there's a delay to get them of over a month, but still interesting and cheap too.
Andy
I've been looking at Edcor transformers - they do 150, 300 and 600 ohm primaries. The Lundahl LL1627 could be used as well - that's a lot more expensive. Actually the Edcors are substantially larger than the UBT1 and Lundahl, which both weigh around 2.5 kilos. The Edcors are more like 4.5 kilos and rated for 400ma. I imagine there's a delay to get them of over a month, but still interesting and cheap too.
Andy
Attachments
Last edited:
I used the UBT-1s because I had them - no doubt better stuff could be had, as I consider the One-Electron stuff to be pretty much MOR. I might expect better things from the Triode Electronics monsters I have on hand, though their bias current capability is not that much higher than the UBT-1 (225 vs 160 ma). The reason I also tried this was that people jumped on to this suggestion when I made it in another thread and claimed that the bias current capability of the UBT-1s would be nowhere near the 085's sweet spot. I want to find out how much that matters. So far I've been pleased with the results. I only wish I'd had time to exorcise the hum issue so I could have brought the amp to BA this year. Experiments with higher bias current might prove interesting. I didn't want to deal with the extra 200 dollars and the wait time.
Any experiments with low or high bias current levels are limited by the power dissipation capability of the 085. It's in a TO-247 package (a good thing), but because of the small SiC die, the junction-to-case thermal resistance is pretty high, much more so than something more proletarian like an IRFP240. The SiC die itself can tolerate high temperatures, but the packaging material, the mount down, and adjacent components are another matter. This would have been a more usable part in a TO-3 package or one of those funky MIL-style metal deals. A company in New Jersey (United Silicon Carbide or something like that) has a data sheet out on a part looking a lot like the 085 (down to the funky gate leakage), but has it in a MIL-style package with glass to metal seals. I'm pushing ~45 watts through my 085s, using aluminum oxide washers (for both voltage standoff and thermal conductivity), and crossing my fingers. As soon as I can get off my butt long enough to lay my hands on some Arctic Silver, I'll be using that, too.
As for transformer limitations on low and high end, folks have been dealing with that in tube amps for decades with audibly satisfactory results.The brief here is semis in a tube-type circuit, how's it sound? Other games without transformers can be played at a later date. The circuit I'm trying also uses just one 085 per channel and doesn't require any byzantine matching games, a distinct plus when you're trying to suss out an unfamiliar and hideously expensive part.
I have a layout going that uses the IXYS depletion mode mosfets in a setup similar to what Nelson used in his First Watt 085 amp. An extra minus supply is needed so that the direct-coupled input circuitry works right with the small bias voltage required by the Ixys devices, but I'm cool with that. When I play with the actual circuitry and get some sensible results I'll post yet another thread.
Any experiments with low or high bias current levels are limited by the power dissipation capability of the 085. It's in a TO-247 package (a good thing), but because of the small SiC die, the junction-to-case thermal resistance is pretty high, much more so than something more proletarian like an IRFP240. The SiC die itself can tolerate high temperatures, but the packaging material, the mount down, and adjacent components are another matter. This would have been a more usable part in a TO-3 package or one of those funky MIL-style metal deals. A company in New Jersey (United Silicon Carbide or something like that) has a data sheet out on a part looking a lot like the 085 (down to the funky gate leakage), but has it in a MIL-style package with glass to metal seals. I'm pushing ~45 watts through my 085s, using aluminum oxide washers (for both voltage standoff and thermal conductivity), and crossing my fingers. As soon as I can get off my butt long enough to lay my hands on some Arctic Silver, I'll be using that, too.
As for transformer limitations on low and high end, folks have been dealing with that in tube amps for decades with audibly satisfactory results.The brief here is semis in a tube-type circuit, how's it sound? Other games without transformers can be played at a later date. The circuit I'm trying also uses just one 085 per channel and doesn't require any byzantine matching games, a distinct plus when you're trying to suss out an unfamiliar and hideously expensive part.
I have a layout going that uses the IXYS depletion mode mosfets in a setup similar to what Nelson used in his First Watt 085 amp. An extra minus supply is needed so that the direct-coupled input circuitry works right with the small bias voltage required by the Ixys devices, but I'm cool with that. When I play with the actual circuitry and get some sensible results I'll post yet another thread.
Attached is a picture of the beast in question - not too big, but with 3 transformers, almost as ponderous as the Dutchy whale. I'll be looking for that hum tomorrow (tonight I'm relaxing by drilling and threading a bunch of holes on the Mutant's new heat sink). I have a capacitance multiplier on each amp module, but if those aren't working right, ripple would sneak through. Hanging a scope probe on the B+ will give me an answer for that in an instant. The transformer I'm using is split bobbin and kinda undersized for its 160VA rating - it may spray flux like a tom cat.
Attachments
Attached is a picture of the two options I have for power transformers for the Puppy. The one I'm using currently is on the right. It's supposedly rated for 160VA, but feels kinda undersized. It also has a split bobbin, which is nice for primary-secondary isolation, but also increases the leakage/stray-spray flux. The transformer on the left is ~250VA, but weighs around twice as much - I trust it a little more, but it weighs a ton. It has a shield, and the end bells should help to contain stray flux. Grounding the core would also help. I had a bid in for an Antek 400VA toroidal isolation XFMR (no longer available from Antek), but I was inattentive and the end of the auction slipped by me.... I may gin up an SMPS for this guy if I get sufficiently frustrated with conventional iron.
Attachments
Last edited:
Just to be sure (a belt and braces approach), I modelled the capacitance multiplier I'm using in the SiC Puppy circuit, and it behaves as I expected it would, drastically slowing down the B+ rise time at turn-on (no thump), and killing the line frequency ripple. This points to zorched components or a real bonehead implementation mistake somewhere. Once I get to the bottom of the matter, i should be able to enjoy some music from this amp without mains hum in the foreground.
SiC Puppy is now working on the bench using an enclosed Ampex transformer rated for 170V output - no hmmmmm, so the split bobbin isolation transformer I was using before was probably polluting its precious bodily fluids with a lot of spray flux. This one will probably make to BA once I get things tidied up and tied down.
The SiC Puppy is definitely go for BA 2012 - it just passed an extended workout in the basement, and will wander up to the living room for some more listening tests. The enclosed Ampex transformer runs things without heating up at all, and no, I repeat, no hummmmmmm.....One of the Antec or the new Diyaudio cases w/heat sink may allow the amp to run without fan, which would get rid of the last of the background noise. That won't happen until after BA, though.
Now that Semisouth has imploded, this will be another one of the orphan amps with no devices to fill the sockets... At least I got mine while the getting was good. It makes it all the more imperative for me to bring up my "L'Fake HV" some time soon. Unfortunately, not in time for Burning Amp this year.
Last edited:
