Fet version of the Sharp channel of the Bogner Triple Giant

KMG · Post by **KMG** » 11 Feb 2010, 15:09

Schematics (pdf)

Stage by stage frequency response in comparison with the original tube version

Ready to transfer PCB (pdf)

Upper side assembly (pdf)

Bottom side assembly (pdf)

3D model

Assembled board

KiCad files
http://milas.spb.ru/~kmg/files/projects ... fet/kicad/
Sample (gain control only by volume knob on gutar)
http://milas.spb.ru/~kmg/files/projects ... _drive.mp3

bool · Post by **bool** » 11 Feb 2010, 21:53

Pretty ambitious and good sounding ...

I have a question: have you thought to use a different biasing mechanism?

Perhaps using a "ring of two" CCS-like "stabilizers" in the fet source's network instead of in gate biasing networks (as usual) - due to this preamp circuit biasing peculiarities. Or are the drain voltages (as marked out in the schematic) so critical for the sound that a "stabilizer" with component tolerances wouldn't be precise enough?

I mean the "CCS-like" stabilizers should ideally set the mosfet source bias by the sources current-sensing (so that the correct voltage-drops should be formed across the drain resistors effectively).

as a reference - the "gate" biasing patent:
http://www.google.com/patents/about?id= ... dq=6683500

bajaman · Post by **bajaman** » 12 Feb 2010, 00:01

Hi KMG
very nice - I have seen this on your website too.
How does it compare with the tube noise wise?
Is it hissy like a lot of high gain solid state circuits or just as quiet as the tube.
It looks nice - I think I will have to build one to satisfy my curiosity.
Good work - thanks

Bajaman

salocin · Post by **salocin** » 12 Feb 2010, 02:37

Very impressive!

KMG · Post by **KMG** » 12 Feb 2010, 09:33

I have a question: have you thought to use a different biasing mechanism?
as a reference - the "gate" biasing patent:
http://www.google.com/patents/about?id= ... dq=6683500

You can connect source circuit to ground and bias gate circuit to regulated positive biasing circuits idividual for each stage.
Drain voltages are important because they define mark-space ratio during limiting of signal.
This ratio has a stronger effect on the sound than the shape of limiting, because it defines the composition of the even and odd harmonics.
Schematics given in patent stabilizes drain current.
Look on the following oscillograms and you will see why stabilization is bad for guitar preamp stages.
This is the response tube stage on burst of input signal
K1 - Output
K2 - Imput
K3 - Grid
K4 - Cathode

All oscillograms you can find there:
http://milas.spb.ru/~kmg/files/projects ... compl/dyn/

As you can see (by cathode voltage line) average plate curret changes it time although input signal amplitude remains constant.
The mark-space ratio of limited signal also changes in time. This allows sound to "to breathe".
This is caused by changing DC component on interstage decoupling capacitors due to grid current on positive halves of signal.
I call this "the dynamic working point"

bool · Post by **bool** » 12 Feb 2010, 11:30

This is precisely why I asked if it would be feasible (reasonable) to introduce stabilizers into the source circuit (in place of the present bias circuit).

This ciruits would in theory sense the source current (approx. drain current minus the MPSA42 base current leak) and set the source bias voltage so that the DC operating point would be ~~~optimal.

Time constants so that it wouldn't kill the tone ...

OK nevermind.

KMG · Post by **KMG** » 12 Feb 2010, 12:42

Another one note about schematics
In this oscillogram you can see the difference in limiting upper half of sine.

Tube limiting rounder than fet.
If you replace germanium diode in source by four ones waveforms practically will coincide.
In latest versions I always install four germanium diodes in source of fet.
More about tube emulatiom by fet there:
http://milas.spb.ru/~kmg/tubeemulation_en.html

bajaman · Post by **bajaman** » 12 Feb 2010, 13:47

Thanks for all the information Mike

Most of it will go over the heads of the majority of members here, but there are a few of us that REALLY appreciate you sharing these impressive designs with us.
My only negative comment is the parts count required ( a tube amplifier with similar sound characteristics is so much simpler to implement)

Now all we need is an enterprising chip manufacturer to come up with a quad custom chip to realise this schematic with fewer components (like a tube amplifier), then we can finally put the jfet ROG designs and Dr Boogie circuits to rest - they do not sound anything like a real tube amplifier.

cheers
bajaman (tube lover)

KMG · Post by **KMG** » 12 Feb 2010, 14:43

The main disadvantage in the tubes the long term instability of the parameters.
It is very unpleasant to begin to rehearse and to reveal that the sound is poor or tubes are completely dead.
It is only one advantage of my designs in comparision with tube amplifiers.

soulsonic · Post by **soulsonic** » 12 Feb 2010, 15:14

I am learning some good things here! Thanks!

borislavgajic · Post by **borislavgajic** » 12 Feb 2010, 19:53

soulsonic wrote:I am learning some good things here! Thanks!

uciti uciti i samo uciti!!!

learning is nice for mee too

thanks

zdravo drug
Borislav

bool · Post by **bool** » 12 Feb 2010, 21:00

On the page you linked ... the opamp e-recto totally anihilates sansamps and line6 (although I pretty much love their modulation effects).

But, back to your design. Another question: why germanium diodes? For the knee curve or voltage drop? Could it be substituted for a silicon? or a led?

Also, I attached the "auto-bias" stage (that's what I meant in my previous comments). Ideally, the current-setting resistor in the source "CCS" circuit would be chosen so that the voltage on the drain resistor would be the "right one". Again ideally, different fet's Vgs spread would be compensated by the source-biasing circuit.

I don't have a clue how would it perform sonically. Perhaps you could check it out if you have some spare time.

B Tremblay · Post by **B Tremblay** » 12 Feb 2010, 21:06

bajaman wrote:...we can finally put the jfet ROG designs and Dr Boogie circuits to rest - they do not sound anything like a real tube amplifier.

soulsonic · Post by **soulsonic** » 12 Feb 2010, 21:33

oh snap!

bajaman · Post by **bajaman** » 12 Feb 2010, 23:31

Come on Brian

No need to act so surprised - you have heard a REAL tube amplifier I am sure - the harmonic content is entirely different to the endless jfet substitutions that proliferate the internet - toys yes, but accurate substitutions for tube amplifiers

, I think not

cheers
Steve

B Tremblay · Post by **B Tremblay** » 13 Feb 2010, 00:56

bajaman wrote:Come on Brian
No need to act so surprised - you have heard a REAL tube amplifier I am sure - the harmonic content is entirely different to the endless jfet substitutions that proliferate the internet - toys yes, but accurate substitutions for tube amplifiers , I think not
cheers
Steve

I'm not surprised at all. As I've written before, runoffgroove.com does not make any claims about amp-inspired FET-based circuits sounding "just like the real thing." We don't use the terms "emulation" or "simulation" because we don't want to make false claims. Please keep in mind that we are freely sharing those circuits. I could understand the criticism if we were fleecing unsuspecting buyers after we'd issued some "we-make-FETs-sound-just-like-tubes" claims, but we've done neither.

There's no dispute in matters of taste. If a circuit sounds good to an individual, then it has value. If you find no value in it, then don't bother with it. The three of us at runoffgroove.com are simply enjoying a hobby and then sharing our enjoyment with others.

RnFR · Post by **RnFR** » 13 Feb 2010, 02:02

bajaman wrote:Thanks for all the information Mike
Most of it will go over the heads of the majority of members here, but there are a few of us that REALLY appreciate you sharing these impressive designs with us.

yes, please keep them coming! i'm not going to pretend like i understand everything going on with this design, but i think that's the first step to clarity, right?

bajaman · Post by **bajaman** » 13 Feb 2010, 03:42

The three of us at runoffgroove.com are simply enjoying a hobby and then sharing our enjoyment with others.

Hey - I can relate to that

No offense meant - runoffgroove is a great site with some very interesting projects - I know folks that have built many of your freely shared designs and have been very happy with the results

These Russian guys are taking the whole fet thing to a different level and they are sharing there findings with us here and I guess I am just excited to see these developments that go further towards mimicking that elusive tube sound

I just love the true international flavour of this forum and it is always nice to see new ideas shared here, whether they are from Chile or Russia or anywhere else on our planet.
Please accept my humble apologies if I have offended you in any way.
cheers
and thanks to all who share the love of guitar (and bass

) pedals here on freestompboxes
bajaman

RnFR · Post by **RnFR** » 13 Feb 2010, 07:00

i think this translation is a good start to figuring it all out.

http://babelfish.yahoo.com/translate_ur ... =Translate

KMG · Post by **KMG** » 13 Feb 2010, 09:29

But, back to your design. Another question: why germanium diodes? For the knee curve or voltage drop? Could it be substituted for a silicon? or a led?

For the knee curve of course. this is why four diodes give better results than one. Silicon gives less dynamic resistance dV/dI near zero currents. Maybe leds, but leds have higher voltage drop.

Also, I attached the "auto-bias" stage (that's what I meant in my previous comments). Ideally, the current-setting resistor in the source "CCS" circuit would be chosen so that the voltage on the drain resistor would be the "right one". Again ideally, different fet's Vgs spread would be compensated by the source-biasing circuit.

I already wrote that current stabilization is bad for sound in this case.
We need only set up initial current (drain voltage). Don`t forget - we already have DC local negative feedback by source resistor.
Latheral mosfets have less Vgs spread than other types due to their structure.
If you prefer constant resistors - replace trimmers after adjusting.
I think it will be enough to do it once for each stage design.
I use trimmers because constantly experimenting,