FETs vs Tubes - exploding the myth

[gtudoran]

Hey guys... yesterday i was bored like hell and i was thinking ..." i have a oscilloscope and i do have some spare time, why not to make some tests regarding FETs and Tubes"

So that is what i did:
- took my signal generator and used a 1Vpp / 1kHz as a test signal
- made a fet preamp like this (biased @ 4.5v , the fet was J201)

- made a tube preamp (only one triode) useing the same schematic as above (only the fet was changed with a triode and the anode resistor was 100k (no trimmer used)), anode voltage was about 120v

So here we are:

FET preamp:


Tube Preamp



I let you guys to make some conclusions and i will answer to any tehnical question (i didn't write all the settings that i used on my oscilloscope bc they are not relevant, but if someone would like to know i will post them, no problem about that).

Best regards,
Gabriel Tudoran
Analog Sound

PS: as my friend Frq.central sugested i will try and record some audio clips with both of the preamps and post them here as a little help for those that wanna hear the results.

[GP]

Those two pictures are exactly the same image file! same filename, same link...

EDIT... well, they were first time i looked! now they're not.

[gtudoran]

done .. sorry for my mistake.

Best regards,
Gabriel Tudoran
Analog Sound

[GP]

i have to say i'm not sure what this really shows. does it show the difference in nature between valves and FETs or does it perhaps simply show the effects of differing headroom?

[gtudoran]

i have to say i'm not sure what this really shows. does it show the difference in nature between valves and FETs or does it perhaps simply show the effects of differing headroom?

Is just a test showing the difference of headroom. And just to prove a thing: FET will not act like a tube ... at least that is my opinion... there are ppl that thinks this (they are both voltage controlled devices... but they have a total different way of work).

Hope i will not put up a fight regarding this topic. And i will also make a spectrogram for both amplification stages useing the same test conditions (i will make them as soon as my Spectrum Analiser probes are comeing form USA)

Best regards,
Gabriel Tudoran
Analog Sound

[petemoore]

  To compare waveshape deviations, the tube needs to be made to deviate the wave somewhat closer to the amount the Jfet is.
  The Jfet chosen for optimal 9v operation ?
  try it at 18v ?

[gtudoran]

Jfet is a J201 biased @4.5v and useing a 9v supply... i will try to make a test with 18v supply and post the results.

Best regsrds,
Gabriel Tudoran
Analog Sound

[GP]

don't misunderstand me - i'm not intending to cause a fight! i was merely wondering what the purpose of the test is.

in one way you are clearly showing how so called "fet emulations" are utterly imperfect under the operating conditions most often used but it might be reasonably argued that your test is of the operating conditions rather than the devices themselves.

[B Tremblay]

From the "A closer look at the Fetzer Valve" article:

"While the [Fetzer Valve 1.1] has its merits, we decided to further explore it to determine if it was a truly faithful solid-state version of the familiar input stage used in many tube amps."

Over four years ago, the specific FET circuit you chose to test was deemed inadequate for not being sufficiently similar to tubes.  I would like to see your scope photos of the Fetzer Valve Revisited (using different FETs and supply voltages), which was developed as a result those findings.

[frequencycentral]

To l'il old non-technical me, the 'scope shows that the tube circuit amplifies the waveform without changing it's characteristics - the output waveform is still a sine. Whereas the specific FET circuit used amplifies the waveform but also alters it to be a new waveform, almost square, not just a sine with clipped peak and trough. Just sayin'.

[gtudoran]

No offence taken GP, just trying to show a point, but i will take any of the schematic or topology that will be proposed (just one amplification stage) and compare it with a tube. I'm not saying that jfets are bad (hell no i'm useing a lot of FET pedals and i enjoy the best of them) but in the terms of dinamycs and headroom a tube (at least in my humble opinion) will be on top.
I will also make a test with a 12AX7 with a 18v anodic voltage and see how it will behave.

Best regards,
Gabriel Tudoran
Analog Sound

PS: i'm not trying to prove that FETs are better or tubes are better, but rather to put a line between them and put in Bold Characters that: Tubes are not Jfets and Jfets asre not Tubes (there are a lot of ppl that will believe that Jfet=Tube)

[gtudoran]

From the "A closer look at the Fetzer Valve" article:

"While the [Fetzer Valve 1.1] has its merits, we decided to further explore it to determine if it was a truly faithful solid-state version of the familiar input stage used in many tube amps."

Over four years ago, the specific FET circuit you chose to test was deemed inadequate for not being sufficiently similar to tubes.  I would like to see your scope photos of the Fetzer Valve Revisited (using different FETs and supply voltages), which was developed as a result those findings.

I will make a FvR and i will also post the results. What do you recomand to use? J201 or 2N5457? What schematic should i use? This one:
?

Best regards,
Gabriel Tudoran
Analog Sound

[slacker]

I think what would be interesting would be to attenuate the input signal on the Fet circuit or increase the input signal on the tube circuit so they are both about the same proportion of the supply. That way you would be seeing the effect of the Fet versus the tube, at the minute, in my opinion your results are mostly just showing the difference in headroom in the circuits. If you clip a signal enough with anything you will end up with something pretty square.

[GP]

To l'il old non-technical me, the 'scope shows that the tube circuit amplifies the waveform without changing it's characteristics - the output waveform is still a sine. Whereas the specific FET circuit used amplifies the waveform but also alters it to be a new waveform, almost square, not just a sine with clipped peak and trough. Just sayin'.

Following on from what slacker says... it depends a lot on where you clip a sine wave. For me, the first trace is almost certainly a sine but clipped very early on, i.e. if you extrapolated the waveform beyond where it clips, it would look distinctly more like a sine wave.

[B Tremblay]

I will make a FvR and i will also post the results. What do you recomand to use? J201 or 2N5457? What schematic should i use? This one:

Yes, that's the circuit I'd choose.  Use either a J201 or 2N5457, measure the Vp and Idss as outlined in the article, and try it with various supply voltages (9, 18, 24).  Use the calculator at the bottom of the page to bias appropriately.  Then repeat the tests with the other FET type.

[gtudoran]

I think what would be interesting would be to attenuate the input signal on the Fet circuit or increase the input signal on the tube circuit so they are both about the same proportion of the supply. That way you would be seeing the effect of the Fet versus the tube, at the minute, in my opinion your results are mostly just showing the difference in headroom in the circuits. If you clip a signal enough with anything you will end up with something pretty square.

I've made that test too (with a pot on the input) and the clipping begun above 100mV. But wil make a test with the FvR below the clipping zone.

Best regards,
Gabriel Tudoran
Analog Sound

[Derringer]

I recall taking a J201, 10K on Drain (9V PS), 1K on Source, 1M Gate to Gnd and I got a picture on my scope that looked more like this picture



You can definitely futz with the fetz to make it do different things

[slacker]

On the scope shots is the output waveform scaled so it's the same size in both pictures? If so, how big compared to the input are they really.

[Earthscum]

I know I just woke up, but I am so utterly confused. What exactly are you trying tp prove here? It's all about clipping characteristics. Clip that tube! Cram that signal right up in there! It's not about how MUCH signal each can pass without clipping, it's about HOW each device clips the signal. Want to see a difference? Scope a transistor booster circuit as well for comparison.

[gtudoran]

On the scope shots is the output waveform scaled so it's the same size in both pictures? If so, how big compared to the input are they really.

Tube output was about... 30v with a 1V input. Now that some of you want some other tests, i will take in account all the request from you and put the tube in overdrive and also make a J201 amplifier stage with the data from Derringer and post the results.