That's a gorgeous chassis. Bob Carver has really had some incredible designs over the decades. I once bought a Carver unit online before online shopping was a thing. It was so inexpensive I couldn't resist. But what I got was a really bad receiver made in China before the good stuff was being produced there. Just a blip in his long career of turning out great amps.
I finally threw together the test rig into a scrap chassis...
These tubes (6DQ6GTB, triode connected) make 25WRMS.
These tubes (6DQ6GTB, triode connected) make 25WRMS.
On the bench early in the wee hours of the morning is the new XSA Labs Series One A40 amp PCB getting installed onto the heatsink. Some drill and tap operations now needed for the new locations of the the underhung MOSFETs and bridges.
The interior will be cleaned up with having two trafos and a new PSU control board that contains a remote enabled triac and SSR soft start plus some safety shutdown thermal switches and all connections for IEC and trafos. That board is currently under development. For initial evaluation testing, I’ll be using a prototype based on an advanced version of the SFP with remote on off.
Renders:
The interior will be cleaned up with having two trafos and a new PSU control board that contains a remote enabled triac and SSR soft start plus some safety shutdown thermal switches and all connections for IEC and trafos. That board is currently under development. For initial evaluation testing, I’ll be using a prototype based on an advanced version of the SFP with remote on off.
Renders:
Looking good, and reminds me that my next order if any should be at Mouser and not Digi-Key...
I love that no-nonsense back panel design and layout! Super clean.
Heatsink drilled and tapped.
Installed boards with MOSFETs, thermostat (80C) and rectifier bridges:
Ready for testing:
Installed boards with MOSFETs, thermostat (80C) and rectifier bridges:
Ready for testing:
X,
Very interesting. What is the basic design topology of this design? Looks like an All Cees power supply integrated onto the main board and a single N channel and P channel set of Mosfets. Input Wima cap. Perhaps a differential input pair BJT’s along with current source.
Beautiful work as always.
Best,
Anand.
Very interesting. What is the basic design topology of this design? Looks like an All Cees power supply integrated onto the main board and a single N channel and P channel set of Mosfets. Input Wima cap. Perhaps a differential input pair BJT’s along with current source.
Beautiful work as always.
Best,
Anand.
@poseidonsvoice: It’s based on Hugh’s Alpha Nirvana with an on board CRC supply. The design is licensed from Hugh Dean and Aspen Amplifier for commercial use. My 23v trafo is running a little hot +/-29.5v since I borrowed it from another project.
@Vunce: The heatsink is 300mm x 150mm. Best part is all the tapped holes are 7mm deep to allow standard 6mm deep brass standoffs to go all the way in without having to trim the thread shorter (my pet peeve with many commercial chassis from Italy). The 6 new holes I tapped are also deep and requires use of a bottoming tap to make sure it can fit a standard 6mm deep thread.
@Vunce: The heatsink is 300mm x 150mm. Best part is all the tapped holes are 7mm deep to allow standard 6mm deep brass standoffs to go all the way in without having to trim the thread shorter (my pet peeve with many commercial chassis from Italy). The 6 new holes I tapped are also deep and requires use of a bottoming tap to make sure it can fit a standard 6mm deep thread.
The heatsinks were running 48C before the bridges were put on them. Probably 52C with the bridges is my guess. But final product with have lower voltage trafo. Right now it’s 29.5v and should be closer to 27v. That should cool it a bit.
Back panel was 42C with old config.
I powered up one side and no smoke but there’s a problem. Bias current is balanced on both P and N but there is +24vdc offset on output. N channel is 0v at Source but P channel is 24v at Source and -29v at Drain. So something is off in the active CCS on the bottom half.
Back panel was 42C with old config.
I powered up one side and no smoke but there’s a problem. Bias current is balanced on both P and N but there is +24vdc offset on output. N channel is 0v at Source but P channel is 24v at Source and -29v at Drain. So something is off in the active CCS on the bottom half.
Last edited:
The other channel works fine. Rail is +/-29.9v and bias current is 1.69A for 101W total dissipation. Bridges will be another 4W for 105W total. DC offset is 7mV.
Playing music now on one channel:
At least now I can do a side comparison for what’s wrong.
Heatsink is measuring 52C at hottest spot with IR thermometer. About 47C at coolest spot.
Playing music now on one channel:
At least now I can do a side comparison for what’s wrong.
Heatsink is measuring 52C at hottest spot with IR thermometer. About 47C at coolest spot.
Last edited:
I found the problem. A small solder bridge on the bottom side of the board. It unbalanced the LTP input stage. Cleaned it up and it’s all fine now. Playing in stereo with dual monoblocks in same chassis.
Hehe, there are indeed quad independent rails. But since each amp requires dual rail, the “blocks” refer to the amplifier music channels.
One is lying flat because I was trying to make room for the terminal block in the old config where the PSU boards were on the floor of the amp. I can now rotate them both flat in the new version.
All the wires and miscellaneous SCR and always on AC/DC 5v PSU will be made neat and moved onto the new SSR soft start and controller board based on Jhofland’s SFP.
I may have to redo the rear panel layout to have the RCA’s and binding posts closer to each channel to reduce wire run length and put the IEC in the middle. Visually, I don’t like the symmetric IEC “nose” in the middle look as much.