Fetzer Valve Vs. Fender Input

[mutant_guitar]

I posted this on another forum, but I figured it would be good here, too, since the idea for this project comes from the DIY community. This is part of my final project in one of my electronics courses this semester. The other half of this project involves running a Discrete Fourier Transform on the output of these circuits to see what kind of harmonics they really are producing. Credit goes to ROG for creating something I can test in class for a grade.

The Test:
Below are four sound samples. One of them is a tube circuit, the other two are solid state. The first clip is straight from the looper. See if you can pick out which one is the tube circuit.

The first part is played through the bridge p/u, the next two are through the neck.

http://home.pct.edu/~cumnic87/three  -- this is straight from the looper
http://home.pct.edu/~cumnic87/five
http://home.pct.edu/~cumnic87/six
http://home.pct.edu/~cumnic87/seven

don't ask about clips 1-4.

Post which one  you guys thinks is the tube circuit.

Also, excuse my crappy playing.  I think I spend more time tinkering with circuits than practicing.

The Theory:
I'll  cut this section from here, since alot of you guys already know it, it's pretty redundant.

The Procedure:
I recorded one clip through my Boss RC-2 loopstation, and played the same clip through each circuit. This way, there is no variation in playing from circuit to circuit that could alter the results.

The setup goes like this:
Boss RC-2 -->Tube Reamer (from ROG)--> Circuit under test -->Behringer mixer -->Behringer UCA202-->Acoustica Mixcraft

The guitar used was a set neck Douglas LP copy from Rondo with a GFS crunchy pat in the bridge and the stock p/u in the neck.

I added a cabinet simulator to each recording in Mixcraft.

[yeeshkul]

seven?

[aziltz]

five and six sound pretty similar.  seven sounds warmer but also buzzier, and I don't associate buzzy with tubes so its a tough call.

through an amp the differences should disappear.

[cheezit]

My guess is that the third one is tube, some type of preamp-drive buzzy circuit, certainly not my favorite tube tone. 

Which gets to the heart of the issue...it's like driving three cars and picking out which one has the super-expensive tires.  If all three drive the same way, what does that say?  Maybe the car with the expensive tires also has worn shocks.  It certainly doesn't say that the expensive tires aren't worth it, just that their influence couldn't be detected due to other factors---if all other factors are controlled, then the influence of the expensive tires is isolated and can be judged.

In this case, the situation is even worse...the Fetzer/ROG designs make it seem like a circuit can be "converted" by swapping in JFETs in for triodes...but IMHO it's only a happy accident that the conversion sometimes "works".  JFET gain stages can look much different than anything seen in a tube circuit, and triodes can be configured in ways that JFETs won't be happy about.  And we've all heard tube circuits that blow chunks and display none of the qualities that anyone associates with good tube tone. 

Also, if you ask which of your samples is better, that's one question; which sounds most like tubes, that's another; and "one of these is tubes, guess which one" is a third.

So that's my guess, plus a little opinion on the side.

[petemoore]

  Ok I'll bite...
  Seein' how I keep typin' about the first gain stage @9vdc being a Jfet has a lot of the same characteristics as a HV first tube gain stage [ie, takes a dinky singnal and amplifies it...mostly]...and there's a 'difference' but nothing near like when you get to the second stage [where the input signal has been boosted up enough to where the stage will tend to exhibit 'character' or non-linearity or [distortion].
  Six

[newfish]

To my ears, the clipping on 'Six' sounds a little softer - so that's my guess.

...and it is a guess.

Maybe we should ask Neil Young?   

Great thread though - good luck with the course.

[mutant_guitar]

thank you guys for participating so far. In another day or two, I'll let you guys know which is which. till then, I got lots of work to do. 

[jrc4558]

I hear more low end in the 7, but I wouldn't attribute that to a tube circuit. There's also an unpleasant distortion in all of them. If you were using the clean signal it would be easier to tell.

I'll have to guess, so let it be 5.

[mutant_guitar]

I would think using the clean signal would make it harder. In each circuit, just one gain stage isn't enough to push any of them into the non-linear "tube character" zone.

I understand they all sound bad. If I had ran them all through an amp, it would definately sound better. Unfortunately, I can't bring my amp in the lab, and I can't bring the lab equipment to my dorm.

[blanik]

i'd say the last one is tube

[mutant_guitar]

I  guess I can tell you guys the answer now.
5 is the revised fetzer valve
6 is the  old fetzer valve
7 is the tube

The input to the circuit is taken from an Agilent 33120A Function Generator.
The output waveform is captured on a an Agilent DSO3102A Digital Oscilloscope.
The scope sends the waveform data to a C# program where a DFT is performed.
The results get plotted on a pretty graph that I can coppy into MSPaint, upload to photobucket, and share with all you guys.

So basically:
Function Generator-->Circuit-->Scope-->Computer-->DFT algorithm-->Graph

These are spectrum graphs. The x-axis represents frequency components within a signal, and the y-axis represents the magnitude of the frequency components. The red dots are the circuit output and the blue dots are the input from the function generator.

What these graphs are telling us is that both circuits produce harmonics. I would consider anything below -20dB inaudible, but I'm not quite sure how "inaudible" is generally defined. This is just one test between one JFET and one tube, but I think it shows that JFETs can produce harmonics very similar to tubes. Next week I should have 100 of these done for each circuit at three different frequencies. Then we'll have a really good view of what's going on.

note, the first graph should say 766mVpp, not 7.66mVpp.

[Uma Floresta]

6 is the  old fetzer valve
7 is the tube

Interesting. I think the old Fetzer Valve sounded closer to the Tube clip than the revised version.

[~arph]

+1

[newfish]

Thanks for doing this - it's been an interestig experiment.

Does this mean that there's now a way of comparing Ge and Si transistors for 'mojo matching'?

[B Tremblay]

Well done!  Can you provide more information about the original and revised Fetzer Valve setups?  Such as FET type, supply voltage, bias voltage.  Some information about the tube circuit would be helpful too - schematic, supply voltage, tube type.

[Uma Floresta]

You know, the clipping on all three was actually pretty similar in character - the difference might have disappeared with some extra post-overdrive filtering - rolling off some highs.

[mutant_guitar]

all the schematics I used are taken right off of the Fetzer Valve article, with the only difference being they all have a .047uF output cap. It was the only cap I had that could  handle the high voltage in the tube circuit, so I put it on the Fetzers for fairness.

For the both Fetzers, I used a 2n5457 with a 9Vdc supply. I used the calculator in the article to set the bias for the revised after testing for Vp and Idss. The old Fetzer is biased normaly at 4.5Vdc with the standard 1.5K drain resistor.

The tube circuit is based around a JJ 12ax7 and running off of 350Vdc. It's a standard fender input, and the schematic is in the fetzer valve article under the frequency response section.

[Ben N]

I also figured it was either 6 or 7, had ruled out 5. (But I wasn't confident enough about it to post .) Either way, the notion that a standard triode based preamp circuit and a well-designed jfet analogue aren't very different in sound isn't all that surprising, especially when the test covers a limited dynamic range. (That's not a criticism, just an observation.) You might get more variation if, say, you hit the inputs with a pre-boosted signal.

It gets harder when you try to capture the saturation/dynamics/distortion of an entire tube amp, with the contributions of PI, power tubes and PT factored in. IMHO, there aren't many MV-type distortion preamps that sound any better than a well-tuned dirt box into a ~clean tube amp, which is why channel switching amps hold no appeal for me.

[yeeshkul]

yeah! i was right

[cheezit]

It gets harder when you try to capture the saturation/dynamics/distortion of an entire tube amp, with the contributions of PI, power tubes and PT factored in. IMHO, there aren't many MV-type distortion preamps that sound any better than a well-tuned dirt box into a ~clean tube amp, which is why channel switching amps hold no appeal for me.

Exactly.  So this it confirms that a particular application of tube distortion---driving a preamp tube until it spits out buzzy distortion---can be matched by a FET.   That's good news for fans of certain genres that rely on that sound, but it doesn't really extrapolate to much else.  I know FET's are capable of better sounds than that, for one.

I don't mean to bash this effort, I think it's great, but I just hope too much isn't read into the results.