Suzy,
Did you have a schematic for the final version of this? I've been meaning to get something along these lines together for a long time but just haven't had the time.
Yup, it’s in post 47: https://www.diyaudio.com/forums/equ...supply-using-lateral-mosfets.html#post6551955
ha!! I saw the thread title and thought 'awesome, I have everything for this in my parts draw' then they talked you out of laterals ... lol. I have a bunch of the dual die semelab/allfet and exicons and can easily spare 4-6 (I reckon I can perhaps run them hotter, due to them being the big daddy parts?) then the forum talked you out of it ... hehe. I reckon i'll push ahead and play with the spice and do up a pcb of my own for the laterals. been meaning to knock something like this up for years and never got around to it, but I figure I can either mount them remotely on the sink and use them for something else if I really need them, but this way I can maybe add a second set of higher voltage rails and use it as a psu for the J2 i'm building when i'm not using it for a lab supply. only thing i'm missing is a suitable transformer.
Good stuff Suzy!!
Good stuff Suzy!!
I'm also adapting this design for my purposes... won't be as nice as Suzy's implementation, but I also don't need ultra-low-drift for my application. I've got other power supplies for that.
What my challenge is going to be is to get a PAIR of 70V 8A power supplies into a single 2U enclosure. I've got a pair of heatsinks and a chassis from a Crest CA9, so we'll see how successful I really am. I think I also have a couple of 2U HP server fans for it too.
In any case, I strongly recommend using a transformer tap-switching system for something like this, otherwise the amount of heat being dissipated gets insane. A buck converter as a pre-regulator is better, but also more complicated to implement.
No reason you couldn't use lateral FETs for this, and I suspect you could get the current sharing worked out. The bigger issue is that they're (IMO) way too expensive for a power supply.
What my challenge is going to be is to get a PAIR of 70V 8A power supplies into a single 2U enclosure. I've got a pair of heatsinks and a chassis from a Crest CA9, so we'll see how successful I really am. I think I also have a couple of 2U HP server fans for it too.
In any case, I strongly recommend using a transformer tap-switching system for something like this, otherwise the amount of heat being dissipated gets insane. A buck converter as a pre-regulator is better, but also more complicated to implement.
No reason you couldn't use lateral FETs for this, and I suspect you could get the current sharing worked out. The bigger issue is that they're (IMO) way too expensive for a power supply.
Yeah, the expense doesn't worry me.
A. They were bought and paid for long ago, so they are effectively free.
B. As a result, it would be more expensive (for me) to buy something else for the job.
C. I have built the amps I bought them for, plus spares for those, as well as some set aside to build Juma's SuSy BF862->OPA1632->Laterals circlotron, as I find that schematic intriguing and then some left over.
D. Moving forward I can't see myself building too many high powered Class A/B amps, as i'm using digital XO and thus it calls for multichannel/ lower individual power-amps.
Got plenty of verticals too, but nothing as beefy as the ALFETs and I suspect they would be more of a PITA to get right, so I figure the time saved its worthwhile by itself.
I dont have a lot of experience with high powered Bipolars, always favoring FETs in my amps, so dont have any stock to draw on.
Yes, I was looking at the various ways of lowering losses for lower VOUT. Since I will likely be running an active bridge, perhaps I can switch the secondaries between series/parallel that way, vs normal relays. Perhaps i'll treat it as a proof of concept for such a design for my amps, allowing lower rails for lower power when the volume is low. The Buck convertor angle is something for another day.
A. They were bought and paid for long ago, so they are effectively free.
B. As a result, it would be more expensive (for me) to buy something else for the job.
C. I have built the amps I bought them for, plus spares for those, as well as some set aside to build Juma's SuSy BF862->OPA1632->Laterals circlotron, as I find that schematic intriguing and then some left over.
D. Moving forward I can't see myself building too many high powered Class A/B amps, as i'm using digital XO and thus it calls for multichannel/ lower individual power-amps.
Got plenty of verticals too, but nothing as beefy as the ALFETs and I suspect they would be more of a PITA to get right, so I figure the time saved its worthwhile by itself.
I dont have a lot of experience with high powered Bipolars, always favoring FETs in my amps, so dont have any stock to draw on.
Yes, I was looking at the various ways of lowering losses for lower VOUT. Since I will likely be running an active bridge, perhaps I can switch the secondaries between series/parallel that way, vs normal relays. Perhaps i'll treat it as a proof of concept for such a design for my amps, allowing lower rails for lower power when the volume is low. The Buck convertor angle is something for another day.
I read all suzyj´s threads with great interest and so with this thread. Thank you for so generously providing models etc to the forum.
Considering I have more than an handfull of older (90ies) standard vertical mosfets, with SOA-DC specified I decided to do something with that. So here is my model variety with a local feedback around each mosfet. Yes, a lot of OPs are required. They are kind of obsolete when it comes to switching applications in some cases too...
Still requres some tinkering, as the models are not the same as for my mosfets (STW38NA50 in SOT-227 or STW50NB20 in TO-247) but as a general proof I guess it works. Might be a bit unstable due to the loop-in-loop stuff unless properly compensated...
The original setup works fine with multiple BJTs replacing the mosfets if using a predriver in a BJT darlington config, NPN/NPN or sizklai style. . I still might have a fistful of MJE21194/5s self-desoldered from an PA-amp where the main transformer accidently was feed 230VAC on 115VAC setting, which not only led to magic smoke release but real smoke too. And some cleaning up afterwards.
Most of the silicone, caps and Xformer was lost but not the outputs...
Considering I have more than an handfull of older (90ies) standard vertical mosfets, with SOA-DC specified I decided to do something with that. So here is my model variety with a local feedback around each mosfet. Yes, a lot of OPs are required. They are kind of obsolete when it comes to switching applications in some cases too...
Still requres some tinkering, as the models are not the same as for my mosfets (STW38NA50 in SOT-227 or STW50NB20 in TO-247) but as a general proof I guess it works. Might be a bit unstable due to the loop-in-loop stuff unless properly compensated...
The original setup works fine with multiple BJTs replacing the mosfets if using a predriver in a BJT darlington config, NPN/NPN or sizklai style. . I still might have a fistful of MJE21194/5s self-desoldered from an PA-amp where the main transformer accidently was feed 230VAC on 115VAC setting, which not only led to magic smoke release but real smoke too. And some cleaning up afterwards.
Most of the silicone, caps and Xformer was lost but not the outputs...