I've just finished hooking up the innards with the exception of the global feedback leads. Tomorrow I'll be doing a cautious power-up. I decided to forego the Garter feedback for now in the interest of simplifying the biasing/debugging. I have a quad of random used Sovtek 6BQ5s in place now that will be replaced with a JJ matched set once I trust the circuit.
The power transformer was more lively than I thought it would be, given that it had its origin in a fairly modest console with small OPTs. B+ is close to 400V, and I padded the screen supply such that it's around 320V. I'm far more concerned about frying the screens than the plate on the JJ EL84s I'm using. Those who have more experience with the JJ EL84 are welcome t chime in with suggestions for sane values of B= and screen voltage.
My big concern is that the bias balance drifts, which mat be due to the 1-turn pots I'm using for bias balance adjust, or the input stage may be oscillating. Some quality time with a scope will sort things out. No Burning Amp for this puppy ths time around - too bad, as it's a purty and compact package.
My big concern is that the bias balance drifts, which mat be due to the 1-turn pots I'm using for bias balance adjust, or the input stage may be oscillating. Some quality time with a scope will sort things out. No Burning Amp for this puppy ths time around - too bad, as it's a purty and compact package.
I had a Tubelab SPP amp set up with the plates at 400 to 425 volts, screens at 320 volts, cathode bias (270 ohms) and 6600 ohm OPT's. It works fine with NOS 6BQ5's (GE / RCA version, or Sylvania version) and JJ EL84's. Some genuine 7189A's work too. Every other tube I have tried, including many "7189's", EI or Chinese EL84's, several Ebay sourced Russian tubes, and some guitar amp branded EL84's, either red plate, or show screen glow. I get between 25 and 30 WPC from this amp depending on tubes and the whim of the local power company (line voltage).
B+ comes from an Antek toroid. Screen voltage comes from a simple open loop "regulator" made from a mosfet follower a zener diode and a pot. It adjusts from about 150 volts (for 6CW5's) to 350 volts.
It was my PC's speaker amp for almost 2 years in Florida......It's still in a box somewhere now and my PC has a $20 chip amp from Parts Express.........
B+ comes from an Antek toroid. Screen voltage comes from a simple open loop "regulator" made from a mosfet follower a zener diode and a pot. It adjusts from about 150 volts (for 6CW5's) to 350 volts.
It was my PC's speaker amp for almost 2 years in Florida......It's still in a box somewhere now and my PC has a $20 chip amp from Parts Express.........
That's pretty much the info I was seeking. I'll be limiting my options to either matched JJ EL84 for non-garter operation, or 6P14P-EV with garter bias equalization.
Last Night I popped the case open and did the following:
1) replaced the 1k gate stopper resistors on the input mosfet differential pair with 10k
2) Replaced the 1k single turn balance adjust pot with a much higher quality multiturn wirewound pot (surplus - new prices for the wirewound trimmers would give you a heart attack).
3) I replaced the screen dropper resistor with 3 X 2.4k, which shound keep the screens toned down to ~320V or so
When I get a chance, I'll haul the thing into work after hours and check it out.
1) replaced the 1k gate stopper resistors on the input mosfet differential pair with 10k
2) Replaced the 1k single turn balance adjust pot with a much higher quality multiturn wirewound pot (surplus - new prices for the wirewound trimmers would give you a heart attack).
3) I replaced the screen dropper resistor with 3 X 2.4k, which shound keep the screens toned down to ~320V or so
When I get a chance, I'll haul the thing into work after hours and check it out.
The amp got some bench time yesterday - no oscillation, but the bias swaps slowly from side to side on both channels. I'm going to try thermally coupling the input mosfets and see if this will fix the problem.
This amp has been sitting in a box for awhile, and as usual, my thought processes have gone and left it behind for a new concept.
I have some new driver boards in process at PCB Way, and will be revamping the power stage into something similar to Baby Huey. If I'm lucky, this effort will make it to BA 2017. More info will appear when the dust settles a bit, but in essence, the design will be DC coupled internally, with folded cascode front end and Baby Huey-style output stage using EL84 and Fisher P-P iron.
I have some new driver boards in process at PCB Way, and will be revamping the power stage into something similar to Baby Huey. If I'm lucky, this effort will make it to BA 2017. More info will appear when the dust settles a bit, but in essence, the design will be DC coupled internally, with folded cascode front end and Baby Huey-style output stage using EL84 and Fisher P-P iron.
Front end is still similar to circuit in post #17, which is diff n-ch mosfet front stage with folded cascode for level shift to drive outputs.
yeah...but a single 'FET as the current reg...let alone the amplifying 'FETs which really should be too...wish I could hear it in any case...🙂
cheers,
Douglas
cheers,
Douglas
Boards are in, mostly stuffed except for semiconductors. Will power up with 250 VDC from a bench supply to tweak in the input stage, then mate with output tubes, XFMR.
I may have to up-size the bias XFMR to deliver the type of negative supply current required by the Baby Huey output stage. This may cause me to decouple the output stage from the input negative supply, as there is no really good reason to have the extra voltage across the bias current source for the output stage.
I would like to finally hear this amplifier.
I may have to up-size the bias XFMR to deliver the type of negative supply current required by the Baby Huey output stage. This may cause me to decouple the output stage from the input negative supply, as there is no really good reason to have the extra voltage across the bias current source for the output stage.
I would like to finally hear this amplifier.
- Status
- Not open for further replies.