Converting Tube Amp To Solid State

Using MOSFET's and JFET's to replace tubes has been around since the devices were invented. There's a commercial version called a "FETRON". I have one of these, unfortunately with a blown JFET.

I can't answer your questions without analyzing the original tube circuit. That won't help because you're missing all the bypass capacitors, missing all the grid bias resistors, missing all the resistors from cathode to ground, don't have any values for the inductors, and have the wrong number for the 6SJ7. 6.3v AC on the output triode will do a fine job of AM modulating your output with 60 Hz. In other words, it won't work. It looks like a schematic created with LTspice. Have you run a simulation to see if the tube version works?

Methinks you will be better off designing your solid state circuit from scratch.

I missed a few more things. The plate voltages should not be 125VAC but rather 125VDC. All the coils will probably require some form of tuning capacitors. The approximate turns ratio of the first 2 transformers might be close, but the output stage is backwards. The tube output impedance is much higher than the load (presumably 50 ohms), so more turns would be needed on the plate side, than on the load side.

Also, why are you specifying metal can tubes? They're not very reliable, get rather hot, and are difficult to find in working condition.

Jeff, you did not notice the real source of raw hum (buzz): The plates are fed with raw AC, running the thing in self-rectifying mode.

The design is sick - forget it:

- there are no by-pass capacitors, - there is no explicit resonance or feedback for oscillation, - there is no output matching, - the pentode tubes are obviously incorrect (maybe 6SJ7GT), - the tubes were ancient already 50 years ago.

don't. There is no other good advice.

it isn't. It's a mad mess.

NT

Sorry. I didn't notice the GT suffix. It comes in both metal can and glass varieties:

Agree +1

Newsgroups need a "like" button for posts..

shift+ctrl+S kills the thread ...Jim Thompson

Thank you Jeff for your expert opinion.

Yes, I know that, by conventional standards, there are missing components, irregularities etc.

I am happy to report that I did not design this circuit. It is of historical interest relating to quack medical therapy.

I could not find the specified tubes and instead wanted to replicate it in solid state to see what the output might actually look like.

If someone can tell me, perhaps this would save me the trouble.

Then again, I am interested it how it could be converted to solid state, mostly as a learning experience. I don't expect it will perform any useful function other than that.

Bob Griffin

Well, that's no challenge at all, Bob!

There seems, perhaps, the impression that "you don't really know until you build it" -- that a device has some magic power by nature of its existence, perhaps imbued through the act of creation that made it.

Fortunately, in electronics, we have no need of that whatsoever! We can fully express the behavior of any* circuit, because that circuit must obey very well understood physical laws.

Now, that does assume having complete knowledge of the circuit. There are many electrical details lacking. What were the dimensions of the coils? Was the wiring short and optimal (as an RF engineer would wire it), or was it laid out on plywood with Fahnestock clips and bailing wire? What was the frequency range, waveform, amplitude and so on of the input? What was the load, how was it connected and used?

Taking it at face value, and assuming modest construction and coil dimensions, it would probably have some gain (perhaps enough to oscillate on its own) at some points during the supply voltage cycle, and so would be expected to output pulsating bursts of RF, at whatever frequency the coils are tuned to.

The output power would be quite weak, unless matched into a fairly high impedance, in which case perhaps a tenth of a watt might be available (and even then, probably peak, not average). In application to biological material, it would do absolutely, positively nothing, at any point in the waveform, for any input.

Perhaps if constructed in a sufficiently imposing manner, it would be capable of impressing (or scaring) the patient into some psychosomatic effect, which isn't to say it has no effect whatsoever, but such an effect could be had with much simpler (if not necessarily simpler-looking), non-electrical equipment.

Tim

Ah, well, it is quack electrical engineering, too! Note that there is no proper bias on ANY of the tubes. In general, they would conduct about full- on during the positive half cycles of the mains. Since they have quite large plate resistors, the tubes won't burn up, but unless the thing is fed lots of input signal, the tubes won't produce much output other than the 60 Hz from the AC plate supply. I see no reason it would oscillate, either, as there is no feedback path anywhere, unless there is some coupling between coils that is not in the drawing.

Looks like a total scam. Since it will NOT do anything as drawn, trying to make a non-working circuit with transistors out of a non-working tube schematic will not get you very far.

Jon

Sigh. Quack medical electronics... you came to the right place. Note that the circuit will not do anything useful as drawn. You obviously have the use of LTspice. Run the simulation and see for yourself.

I probably have both types. I'll look.

Also see eBay: If you want it to work, buy a spare as tubes that old are often dead or dying.

Several of the URL's I previously listed included some theory and same circuits for using a FET to replace a tube. Those should be a good start.

but be aware that putting -400v or whatever it was onto a solid state amp may go badly very rapidly. Valves are tough things, trannies not one bit.

NT

The schematic was done in Diptrace not LTSpice, so I have not run it. I would not know how to hanlde the probable convergence errors anyway.

As far as I know it was built upon a shielded chassis. The air coils were in cans. But due to the lack of bypass caps, etc., I suspect it was intended to oscillate. The guy who designed it would have been at least as smart to know that.

Is there any indication in the schematic that the coils are tuned to a particular frequency?

As I understand it was connected to the subject via hand-held electrodes.

Sorry but these things just fascinate me.

Bob Griffin

To answer the question as is, solid state parts do not behave like valves, especially pentodes, no matter what you do with them. If you're looking for nonobvious behaviour, you ned to use the parts specified and not change it.

NT

The supply rail is presumably of very low impedance, and freedom from mains borne noise probably not a concern, so it might not need decoupling caps.

not explicitly. But all Ls have C, and all LCs are tuned, like it or not. It looks like all 3 transformers are the same, so most likely having the same tuned frequency, give or take some tolerance.

I expect if it passed through the subject, through a neon to ground it would raise the subject's heart rate at least.

We don't know if the output feeds back to the input via some path, but it's quite likely. Even if you only get gains of 10 per valve that adds up to a gain of 10x10x10x4x4x4 = 64,000. Rather more at resonance.

NT

On the contrary - it is often much easier to make non-working pieces of electronics, I have made plenty of them in the last

I hope that there has been a transformer between the mains supply and the 125 VAC plate feed.

Why do I doubt it.

NT

No. That depends entirely on the dimensions of the coils and their parasitic capacitance, and the capacitance (including Miller effect) of the tubes.

Hence my comment about imperfect knowledge. :)

Tim