• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

A Leap of Faith...

I just received a pair of Hi-Fi rated 6.6k 20 VA P-P transformers from Musical Power Supplies. I plan to try them in a push-pull circuit using the 6LR8 vertical oscillator/amplifier tube, with a folded cascode differential front end and an active auto-bias circuit of my own devising. Ohter alternaives for output transformer include a pair of Hammond 1620s I have floating around and a couple of smallish P-P transformers I scavenged from a Wurlitzer organ amp chassis that used a pair of 7868s as outputs.

I'm opening the package to have a look at the transformers, and I can already smell the fresh varnish. - cute lil' buggers, they are.
Last edited:
  • Like
Reactions: 1 user
They were a lot cheaper than the Hammond alternatives. I may also try some of their Hi-Fi SE transformers.

MPS also ships faster than Edcor does these days, though that might not be the case if he gets "discovered" and snowed under with orders. I may talk to him about a custom XFMR for a tube-based headphone amplifier.
MPS has been around for a long time. 15 to 20 years ago he was in Palm Beach County Florida. I got some transformers from him that were wound in a foreign country, Haiti I believe. They were power transformers and worked fine. He left Florida some time ago and went somewhere other than Tennessee, in the Midwest maybe. I guess he moved again. He has always had a store on Ebay, that's where I found him when he was in Florida.

I plan to power the amplifier with a DC-DC converter driven by 24V, 5A switching adapter (350V B+, 175V screen). I'm expecting about 20W/channel output from the chunky 14W pentodes in the 6LR8. I just picked up some RCA types (re-marked Sylvania, most likely), and also have some Sears-branded tubes I purchased a few years ago.
I plan to power the amplifier with a DC-DC converter driven by 24V, 5A switching adapter (350V B+, 175V screen). I'm expecting about 20W/channel output from the chunky 14W pentodes in the 6LR8.
I ran PP 6LU8s (same toob, different base) in a modified clone of Norman Crowhurst's 'Twin Coupled Amp', kind of a poor man's McIntosh.
400V plates & 300V screens, It clips at about 35W per channel. I built only one for experimental purposes as is my habit.
It has run for hours with no sign of overheating or other problems running on these voltages.
The PS is all SS & regulated, The iron is Hammond.
An article covering the project was published in AudioXpress about 20 yrs ago. (y)


  • Twin Coupled Amp A Top RHS.jpg
    Twin Coupled Amp A Top RHS.jpg
    169.6 KB · Views: 128
That's fine, but what I'm doing here has nothing to do with Crowhurst's design - it is my own effort. You may want to start another thread on this instead of tacking it on here. The ultimate power in my design will be limited by the output transformers and the input SMPS. The finished project will boast tubes on top, with output transformers located below decks. That was a motivation for sourcing compact output transformers and using a DC-DC converter instead of a conventional power transformer.
I'll post a schematic when I get around to sorting out the bias currents for the output pentodes. I may add a series-pass regulator for the screen supply to cut down the screen voltage to a level that won't require lightning-bolt levels for grid bias for the output pentodes.
But you don’t get the tip-on-top vintage look with the compactrons.

If compactrons are ok, then the *JA5 is a triodeless version, rated for more watts. Same curves. I ended up using those because a wanted a lower mu driver LTP, and found it difficult to get two of the triodes to match anyway. The pentodes seemed to match up pretty well, though. Another experiment with the LU8 worked quite well, but the triodes were used completely independently so matching was a don’t care. I may build a serious version of it, with the LR8 to get the vintage look.

Of course if you want a screen regulator a 5th tube does the job, and it won’t have the spirito instability to worry about.
I was going to use a 500V mosfet in a full-pack TO-220 for the screen regulator - just a simple follower regulator with a big, fat stopper on the gate. I already have some 9-pin Novar sockets on order, and I have a footprint already made up for it in my copy of Orcad. The 6LR8s I'm using have the tip-off, so the setup should look pretty cool ( tubes on top, OTs and power supply DC-DC down under). The project will be a test for my autobias scheme rather than for the use of the 6LR8 in an amplifier, which is fairly old hat these days.
I'm using a case I bought from Marlin P. Jones, which I'll re-paint in a color that pleases me. It should be big enough for the amp PCB, with tubes on one side and all the other components hanging below on the other side of the board.
You just have to watch it with those 500V mosfets. Especially the ”big” ones. It seems the higher the current and lower Rds they are, the less power they can handle as linear amps. I look for ones only good for a handful of amps and have Rds in the ohm or two range. Those seem to survive being used as pass transistors and current sources with 200 volts across ‘em. Hell, a goddam TIP50 can do that. And at least the published SOA curve is real.

You mentioned a folded cascode. You can do that without a PNP tube? I take it the input pair will be one of the usual suspect double triodes?

I’ve been doing a little experimenting “off line” - that is not in an actual amp, for a bias servo circuit. Something to keep pairs in DC balance automatically, When multiple pairs of unmatched random sweep tubes are paralleled. I consider it a necessary first step in trying to make a 500-1k W amp. Unless you figure a way to keep from magnetizing a big toroid OPT it’s really dead in the water. Your little 20W tafos would probably benefit from well balanced currents too. The pentiodes in the L*8’s do seem to be reasonably matched right out of the boxes, but I did do conventional fixed bias with four pots on mine. They just ended up with vg1 really close when dialed into 31 mA each. Which is good, because those Fisher OPTs don’t have a lot of magnetic headroom either. Distortion is noticeable with bass heavy EDM, but classic rock is wondeful.
  • Like
Reactions: 1 user
Yeah, which one has the better “cool factor”, and which socket is going to be more reliable don’t necessarily align…. Unless Wrench had found some better 9 pin sockets than I did.

Putting all the trafos and power supply below decks would actually be better for safety.


  • A9BAF39E-9C53-45F1-B180-6693A9A1000F.jpeg
    73.1 KB · Views: 73