• 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.

mosfet source follower output stage.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Was told this was an idea with no merit, actually forgot about it for a while. Source followers are used to give greater current capacity to a tube driving a tone stack or even an output tube. I was wondering if you took a preamp pentode and added a mosfet to it could you drive an output transformer instead of using a power tube? I was told the result would not sound like either a triode or pentode output amp because there was no interaction between the tube and the transformer. I am not sure I agree with the conclusion, thought I would ask you guys.
 
I saw a comment once by Patrick Turner on his website about a "tube" amp he built using Mosfets for outputs. As I recall, it was deemed very "tubelike" sounding. No doubt the transformer output used also contributes to some of the traditional "tube" sound. A grounded source configuration would be the usual choice to mimic a pentode. Using a source follower will likely just give the sound of the driver/gain stage instead. The power law transfer function of Mosfets changes from square law thru 3/2 power law toward linear law as the current thru them increases. So by moderate idle current biasing one can get them centered on the 3/2 power law section to mimic tubes quite well.

Another trick you can try is to put some TV damper rectifier tubes below the source terminal of the Mosfets (like cathode bias resistors). This gives curves on the curve tracer that look very much like pentode tubes. To get triode results, also include some local "Schade" feedbacks (resistive loop from "plate" back to "grid") for each device.
 
Last edited:
Was told this was an idea with no merit, actually forgot about it for a while. Source followers are used to give greater current capacity to a tube driving a tone stack or even an output tube. I was wondering if you took a preamp pentode and added a mosfet to it could you drive an output transformer instead of using a power tube? I was told the result would not sound like either a triode or pentode output amp because there was no interaction between the tube and the transformer. I am not sure I agree with the conclusion, thought I would ask you guys.

If you are going to use a power mosfet, why bother with an output transformer. The mosfet can drive the speakers. If you did use a transformers you'd need a MUCH lower impedance one because the mosfet would be operating at lower voltage than the tubes (I assume you don't use 400 volt rails with your mosfet)

There are plenty of these amps around. we call them hybrid amps. Yu really do get a tube sound out of a tube preamp and you can save much $$$ using the power mosfet output stage because you don't need the expensive output transformers.

Some curent production Bass guitar amps have tube preamps and SS power sections because bass players need LOTS of power and kilowatt (or even 500W) tube amps are expensive. So they compromise. The same concept would work for stereo HiFi.

That said some early solid state amps did use output transformers.
 
If you are going to use a power mosfet, why bother with an output transformer.
Just curious.

The mosfet can drive the speakers. If you did use a transformers you'd need a MUCH lower impedance one because the mosfet would be operating at lower voltage than the tubes (I assume you don't use 400 volt rails with your mosfet)

About 300V I would guess.

There are plenty of these amps around. we call them hybrid amps. Yu really do get a tube sound out of a tube preamp and you can save much $$$ using the power mosfet output stage because you don't need the expensive output transformers.

I have transformers already.

Some curent production Bass guitar amps have tube preamps and SS power sections because bass players need LOTS of power and kilowatt (or even 500W) tube amps are expensive. So they compromise. The same concept would work for stereo HiFi.

Don't need much power, thinking around 5-10W.

That said some early solid state amps did use output transformers.

I have one.
 
I've been interested in a Mosfet output stage for a while, but I really only understand tubes and how to bias them. Never read up on solid state. I don't like the idea of a big electrolytic cap in the output - surely that degrades the sound? And transformers are expensive. So I'd be interested in a design with no transformers. Trouble here is I don't know how to bias the tubes properly - almost all the schematics for solid state have input devices which I wouldn't want. I just want a complete self-contained output stage. The circlotron is interesting.

Andy
 
I've been interested in a Mosfet output stage for a while, but I really only understand tubes and how to bias them. Never read up on solid state. I don't like the idea of a big electrolytic cap in the output - surely that degrades the sound? And transformers are expensive. So I'd be interested in a design with no transformers. Trouble here is I don't know how to bias the tubes properly - almost all the schematics for solid state have input devices which I wouldn't want. I just want a complete self-contained output stage. The circlotron is interesting.

Andy

film types can be used instead for the coupling cap, and as far as the OPT, with the lowish impedances used, such as 400ohms to 8 ohms makes design of opts very easy.....interleaving can be dispensed with so much so that power transformers can be used with good results.....

i am building one myself.....

wrt the capacitance issues with mosfets, the solution there would be to operate the driver tube with higher that ussual plate currents imho....:D
 
film types can be used instead for the coupling cap, and as far as the OPT, with the lowish impedances used, such as 400ohms to 8 ohms makes design of opts very easy.....interleaving can be dispensed with so much so that power transformers can be used with good results.....

i am building one myself.....

wrt the capacitance issues with mosfets, the solution there would be to operate the driver tube with higher that ussual plate currents imho....:D

Keep us posted on your progress.
 
I built a SE tube + MOSFET follower parafeed amp with a fixed ratio of feedback to the tube' plate so it can interact with the load in a scaled way, as if the load was scaled up relative to the tube's plate resistance. It acts exactly like a SE amp using a power tube.

It also uses a transformer, but rather than use an expensive SE OPT it uses a low ratio matching transformer from 70V line to speaker.

In other words, the MOSFET follower outputs 70V line voltage and impedance directly to the matching transformer. This allows the tube to interact with the load through a real transformer. It's also a convenient voltage gain for a tube front end (2vrms=>70rms is Av of 35, doable with a D3A in triode or 417A, etc.)

The whole thing is very SE-like with predominant 2nd harmonic and soft asymmetric limiting.
 
Last edited:
I put this stuff on my web site back in 2007. Nobody expressed any interest except for the usual "Transistorlab" comments so I never posted any further info to the web site. I think I posted some of the experiments on this forum somewhere, but it was probably in some loosely related thread. I don't have time to look for it right now.....tropical storm Isaac is coming:(

Super Tubes

The tubes and solid state parts are wired like darlington pairs.

Since then I have experimented and built a few of these. The bottom circuits have not worked out too well usually resulting in smoking sand. Both of the top circuits work quite well and do sound much like pure tube outputs.

The BJT circuit seems to sound the best, yet measures the worst. The distortion numbers and profile look much like the tube alone. SOA limitations on the BJT limit the fun factor and plate voltage.

The mosfet design could put out all the power the little 600 ohm Chinese OPT's could eat. Over 50 watts at 1 KHz, about 8 at 40 Hz. The distortion numbers were very low up to OPT saturation and mostly second harmonic.

All experiments were done in SE using a hacked SSE board. Several tubes were tried with 6V6 working the best for the mosfet since I could run 300 volts and 6Y6 at 200 volts on the BJT since more voltage blew the transistor.

The only P-P experiment using a hacked SPP board with 6CW5's blew up on the launch pad, so I abandoned it without knowing what went wrong.
 
Sounds like the N parts had better voltage ratings than the P parts. These current boost approaches do seem like the "bee's knees" for using power xfmrs or 70V speaker xfmrs as OTs.

Another attractive version is to use a pentode driving an N type SS follower with its output fed back (at somewhat reduced voltage from the plate; using a zener or R divider) to the pentode screen. Optionally a CCS pullup on the plate. Effectively a triode with super low Zout, high current ability, and constant Mu operation.
 
Last edited:
I built a SE tube + MOSFET follower parafeed amp with a fixed ratio of feedback to the tube' plate so it can interact with the load in a scaled way, as if the load was scaled up relative to the tube's plate resistance. It acts exactly like a SE amp using a power tube.

It also uses a transformer, but rather than use an expensive SE OPT it uses a low ratio matching transformer from 70V line to speaker.

In other words, the MOSFET follower outputs 70V line voltage and impedance directly to the matching transformer. This allows the tube to interact with the load through a real transformer. It's also a convenient voltage gain for a tube front end (2vrms=>70rms is Av of 35, doable with a D3A in triode or 417A, etc.)

The whole thing is very SE-like with predominant 2nd harmonic and soft asymmetric limiting.

My favorite, 70V transformers. A bit unsure how you fed back the tube transformer interaction thing sound cool. See, I knew you guys were a creative bunch.

Tubelab, absolutely no surprise you played around with it. Is there anything you haven't done?
 
Keep us posted on your progress.

sure thing, i am commited....i am building the power traffo and a pair of opts in the coming week....

i am using the 10EW7, a fat bottled 9pin tube....

i am using led bias in the first stage, fixed bias on the second stage.....

i am having second thoughts about regulating the B+, my reasoning is that since all stages are run classA and CCS used as loads, then the pssr should be high....

your thoughts are welcome, anyway i will have taps on the power traffo just in case....
 
Got a 4p1l mosfet hybrid breadboarded , yet to measure it but it sounds very good . Low parts count but built tube-style with choke input , choke smoothed ps and choke loads for the mosfets . Currently using (and really abusing) IRL2703 mosfets , there is a cooling fan being used at present as the mosfets are puny TO220 devices disspating 30 watts each .

316a
 

Attachments

  • mosfet4P1Lamp.jpg
    mosfet4P1Lamp.jpg
    203.3 KB · Views: 555
I, for one, would like to know more about an FET source follower driving 6GK6 (actually 10GK6) output tubes using the triode of a 6GH8 in a cathodyne configuration. Is this worthwhile?

I put this stuff on my web site back in 2007. Nobody expressed any interest except for the usual "Transistorlab" comments so I never posted any further info to the web site. I think I posted some of the experiments on this forum somewhere, but it was probably in some loosely related thread. I don't have time to look for it right now.....tropical storm Isaac is coming:(

Super Tubes

The tubes and solid state parts are wired like darlington pairs.

Since then I have experimented and built a few of these. The bottom circuits have not worked out too well usually resulting in smoking sand. Both of the top circuits work quite well and do sound much like pure tube outputs.

The BJT circuit seems to sound the best, yet measures the worst. The distortion numbers and profile look much like the tube alone. SOA limitations on the BJT limit the fun factor and plate voltage.

The mosfet design could put out all the power the little 600 ohm Chinese OPT's could eat. Over 50 watts at 1 KHz, about 8 at 40 Hz. The distortion numbers were very low up to OPT saturation and mostly second harmonic.

All experiments were done in SE using a hacked SSE board. Several tubes were tried with 6V6 working the best for the mosfet since I could run 300 volts and 6Y6 at 200 volts on the BJT since more voltage blew the transistor.

The only P-P experiment using a hacked SPP board with 6CW5's blew up on the launch pad, so I abandoned it without knowing what went wrong.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.