New FET distortion schematic...sensation?

[Mnemonic]

Hey, dudes, don't laugh at our names! You'll be surprised how funny some American names sound for a Russian ear :wink:

[Gilles C]

Hey, thanks for the English translation.

Now, everything is clear... heu, almost. And I'm not talking about the translation, I'm talking about the formulas...  :oops:

Gilles

[Phorhas]

thnx for the time and effeort man... looks like an interesting read

[DDD]

I have to say the following on behalf of the author:
- No any permission is necessary to use his ideas or to publish them
- The subject is PAF
- Good luck!

About the author: his nickname is XBANANOV, one can find him at
http://gtlab.net/forum/index.php?showforum=2

or on the site where he publish his articles under the name "А.Дисторшен" or "Б.Овердрайв"
Regretfully his English is a little bit better than your Russian :-)

[WGTP]

I can't really understand all the different stuff in the schematic, but as someone who grew up with all that cold war crap I think it is too cool to be talking distortion boxes with dudes in Russia.  

I also have tubes in my amp from over there somewhere.  

[ragtime8922]

Thank you! I am sooooo glad I lost my job!!!...lol. Really! I have a new job that atarts in March which means 3 months of electronics and unemployment compensation! Apparently I'll need it for research purposes. I had the whole 3 months laid out until you through this Russian wrench in the mix. Oh well, there is never enough time anyway...for ANYTHING!!!

Thanks again!!!

[markusw]

DDD, could you probably check the english world lingo version for major bugs  :wink: ........besides the :idea:

you'd do us all a reeeaaaally huge favor,

markus

[stm]

Thanks for the translation.  It took me about 2 hours of slow reading to understand 75% of the article.  Very thorough and informative.

Now I would like to share some thoughts:

1) Using 4 FETS, of which 2 are of one kind and 2 of another, is quite a hassle.  Moreover, the need for 2 additional FETs for the output buffer adds additional complexity.

2) I agree the results look very promising for simulating the pentode power amp saturation, but I wonder how good would a simple CD4049 / CD4069 be for this task in comparison.  These IC's are known to sound "very amp-like".  As a matter of fact, I built an overdrive with a CD4069 about two weeks ago, and was very pleased with the clipping characteristics.  This is something to further investigate perhaps with an oscilloscope, and definitely with your ears.  Also, the push-pull of P and N channel FET alternative is mentioned in the paper.  The CD40x9 falls within this category, and it is an off-the-shelf device that would need no extra matching or whatever!

3) If you want to be very purist, you can include the two-FET voltage follower after the 4-FET stage (remember the two FETS need to be of different kind).  On the other hand, if you want to keep it simple, I see no reason why you couldn't use an op-amp in voltage follower configuration.  In particular, a rail-to-rail FET op-amp like the TLC2262 should do just fine here.

4) I do not mean to diminish the interest or the author's merit in this apparently new circuit topology.  I only want to study its pros and cons carefully.  There are some very good sounding FET amp emulators, and an ALL-FET design is *very* appealing; I just want to make sure the extra effort is really going to pay-off.  Just take into account the simple tube screamer topology just rocks.

I invite you to post your comments.

STM

[xbananoff]

Day kind, expensive(dear) friends.
I the writer of this paper also can comment its some positions.
Nobody negates beauty of limitation of numeral invertors such as CD4069. But the invertors a low amplification factor, high output resistance, often have problems to conditions on a direct current owing to some asymmetry.
4FET is dispossessed first and third of the indicated deficiencies. Also gives a not less beautiful sound, than invertor.
TLC2262 It is known MOSop-amp. There is a judgement, that the common closed loop aggravates a sound. I disjoint this judgement, therefore I shall not use any op-amp whenever possible. Certainly, you can have the judgement on this problem.

[stm]

Thanks for your answer, xbananoff.  I think this is a good opportunity to learn from each other and increase our understanding of FETs.

You mention some interesting points which I will reproduce here in italics (using my own words), and then my comments in normal text.

1) CD40x9 inverters have:

a. Low gain.  Yes.
b. High output resistance.  Yes.
c. Assymmetry when subject to DC loading.  Yes, as a consequence of high output resistance.

There is still a chance to use THREE inverters in series to increase gain.  I don't know if they will exhibit the soft and rounded clipping characteristic, so this needs additional testing.

Usually you use two or three inverters of an IC.  You still have several free inverters which can be used in parallel to have increased drive capability (lower output impedance).  In fact, some CD4049 designs like the Way Huge Red Llama don't use an output buffer at all.

2) The 4-FET topology has:

a. High gain.  If I understood the paper correctly, up to 3000 times?
b. High output resistance.  Hence the need for a buffer.
c. Assymmetry when subject to DC loading.   As a consequence of high output resistance.

3) Regarding OpAmp buffers you mention:

Negative effect on the audio signal due to the high gain and closed loop involved.

In this respect modern OpAmps, and especially "audio" OpAmps, are very transparent and accurate, with THD well below 0.1%.  In fact, one may dislike them due to the lack of coloration in the sound (some people may describe this as "lifeless" or "without character").

On the other hand, the proposed 2-FET buffer will certainly introduce some coloration to the sound, which may be desirable and positive to the overall sound.

4) Summary

a. I like your comments in favor of the 4-FET topolgy, and as such I think some experimentation and A/B testing with the inverters is worthwile.

b. It is also worthy to explore alternatives with CD40x9 inverters to get closer to the sound and curves obtained with the 4-FET topology due to the simplicity of inverter circuits.

Best regards,

STM

[xbananoff]

Hellow, STM
“I think this is a good opportunity to learn from each other and increase our understanding of FETs. “
I support such judgement completely!

”There is still a chance to use THREE inverters in series to increase gain. I don't know if they will exhibit the soft and rounded clipping characteristic, so this needs additional testing. “
I have conducted such trials. Each following invertor makes waveform by more square. The sound becomes more and more similar buzzness.
I am convinced, that for deriving a good sound it is necessary to try to receive major amplification in one stage. Such approach will eliminate necessity of usage of several invertors.
4FET really gives amplification in 3000 times and even more. And it only using one stage.
The amplification depends from mu used JFET.

“The 4-FET topology has assymmetry when subject to DC loading.”

We (I and You) abnormally have understood one another in a problem assymmetry.
At usage of the invertor to catch a linear condition it is complicated. And when it is possible is a major happiness. It is very similar to the equilibrist in a circus. It is possible certainly to poise an input of the invertor for this purpose. But I do not like the circuit(scheme), which one should be attuned. All should work at once.

About buffers on OPamp.
About good sound-OPamp it is known even in Siberia, where I dwell.
The circuits(scheme) on OPamp buffers are known for all. I would like to illustrate an opportunity to apply here JFET also.
They dye a sound low frequency harmonicses really a little. The sound in effect(result) becomes closer to a sound tube. And we you see like all a vacuum-tube sound?

4FET the circuit(scheme) is not composite. And places takes no more half-digital-chip.
I want to pay attention. 4FET gives a beautiful sound without an obtruncating of low frequencies up to the device. This indirect affirming of a small amount of intermodulation contortions of a high frequency. It is good for a sound.

Best regards

xbananoff

[Peter Snowberg]

Welcome xbananoff!

Thank you for joining with us and thank you for your great article.

I am eager to construct your circuits.

Best regards,
-Peter Snowberg



----------------------------

Big thanks also to lovekraft0 for your translation!

I'm glad that math and the love of distortion are universal languages.

[xbananoff]

It is a pity, that in English there is't a word, which one is similar to Russian "xyeBo" (khuyovo). Such performance to the full meets to that translation. The text is transferred(translated) abnormally in many places. It hampers exact comprehension of ideas.
I wish successes in recurring those circuits(schemes).
Xbananov

[ragtime8922]

Has anyone built this yet?

[stm]

I would like to try it, but I don't know if it would be possible with the FETs I have.  This is what I've got:

J201
2N5457
2N5458
MPF102

I have about 10 units each, except for the J201 which I have 20, so I don't have too many to do some serious sorting/pairing.

My bet would be to use 2N5457 / 2N5458.  Of course it would be best to use 2N5457 / 2N5459, which I don't have.

Another possibility that might work is J201 / MPF102, but I don't feel like mixing two different families.

If I understood correctly, ideally you need to have the Idss in a 3:1 ratio?

Xbananoff, your comments are appreciated!

Regards,

STM

[xbananoff]

Hellow to all.

It is not necessary very carefully to select details.
This device will work after installation of any properly functioning builders.
Versions of possible substitution are reduced below.
Russin KII303A,X-->J201
KII303I' -->2N5457,2N5458

Russian KII303A and KII303I ' have a low noise and allow to receive fine result.
The application J201 and 2N5457,58 will give hardly the greater noise probably (but it will be noticeably only in a position "MIN" of a level regulator on a guitar)
In speakersim circuit:
Russian p-channel KII103 -->J2N5020
KII303X-->J201
I hope, that this information will be useful.
I wish success to all.
Xbananov

[xbananoff]

To Peter Snowberg

On this site it is interesting. Many interesting ideas and people.
Specially impress some devices operating one transistor!
To me is to that studies here.

[Somicide]

great to see such collaboration!

welcome aboard, xbananoff!  And I'm inclined to agree about the people here being interesting (and helpful!), not to brag, of course!

You're circuits may soon be occupying bench space!

PnL,

Jeff

[RedHouse]

xbananoff,

Your english seems quite good in your posts here, would you consider making another webpage but in english?

[xbananoff]

RedHause.
The good impression about my English is bastard.

But the problems, which one preset here in a ratio of a paper, allow to make the inference about poor translation.

Natural translation to me to make complicatedly, unfortunately. The following version is possible: me send html version of that paper, translated on English, then I shall correct a little this translation to reach(achieve) the greater clearness and comprehension of a material of a paper.
Who is ready to such cooperation?