Help ME design Mosfets stages (as tube stages of amps)

[richon]

Hi,

hopefully noone gets angry because I show this picture from www.runoffgroove.com


I build guitar tube amps and i found myself very interested in building a good mosfet OD/Distotion, so please be kind to understand some "tube amp" concepts in the way i explain myself.


if you see the first stage of the schematics (the mosfet stage):

-the voltage divider to make the mosfet work is the 4,7K + 1M ??

-if i want this stage to have more GAIN, should i increase the 1K resistor o decrease it?? (i guess i shoud decrease it)

-can i add a "cathode bypass cap" parallel to the 1K  resistor to ground to have the clasic boost to some frecuencys like tubes??

-what would happend if I increase both 1M resistors to 10M besides a higher impedance input??? would this broke the voltage divider making the mosfet work bad or not work at all?

-can i build a MODELING of an amp with mosfets (like Professor Tweed, Eighteen )


Ricardo,  Viña del Mar,   CHILE.

[CLAAS]

Check it out..

This is a diy-forum!

Sorry for that..

Hopefully Thilo

[dschwartz]

what kind of reply is that?

Check it out..

This is a diy-forum!

Sorry for that..

Hopefully Thilo

[R.G.]

Have you read http://geofex.com/Article_Folders/mosboost/mosboost.htm?

[richon]

thanks,

this explained very detailed the voltage divider thing , the bias, the value of the "cathode" of the mosfet  and the high impedance..


still, would the bypass cap work in the same way as FETs and Triodes?


Besides, does a MOSFET OD have "tube Color" or should i mix MOSFETs and FETs to get the COLORED sound?

[R.G.]

still, would the bypass cap work in the same way as FETs and Triodes?

If you mean bypassing the source resistor to increase gain, yes, it does.

Besides, does a MOSFET OD have "tube Color" or should i mix MOSFETs and FETs to get the COLORED sound?

Here you have a problem.

No device, not a JFET nor a MOSFET distorts in just the same way that a tube does.

Think about it - if you could just replace a $5 tube with two $0.05 JFETs, would there still be any tube amps made? Obviously not.

MOSFETs and JFETs do distort, but they do it in their own way, not just like a tube. An emulation of an amp made by substituting JFETs or MOSFETs for tubes is NOT going to sound exactly like the tube version. It may sound good, but not the same. Once again, think about it. If it was that simple, wouldn't everyone already be doing it?

[richon]

No device, not a JFET nor a MOSFET distorts in just the same way that a tube does.

Think about it - if you could just replace a $5 tube with two $0.05 JFETs, would there still be any tube amps made? Obviously not.

MOSFETs and JFETs do distort, but they do it in their own way, not just like a tube. An emulation of an amp made by substituting JFETs or MOSFETs for tubes is NOT going to sound exactly like the tube version. It may sound good, but not the same. Once again, think about it. If it was that simple, wouldn't everyone already be doing it?

yes i know (i build tube amps ).... i wasn't trying to invent and discover all over again the WHEEL 

i was only trying to get as good as i can with FETs and MOSFETs to some of the character and tone of some amps... and get finally a good OD sound.

so the question would be:
what would saturate closer-likely a tube OD in an amp when pushed hard by a previous stage:  FETs (j201) or MOSFETs(BS170)¿¿¿???

[mojo_hand]

Since you're asking an opinion question, I'll offer an opinion answer.  Others will answer differently, I'm sure.

I find that either sort of FET sounds more like a FET than it sounds like a tube.  When you use either a MOSFET or JFET in an ultra-squeaky-clean amp stage, e.g., a hi-fi preamp, they will be pretty hard to tell from tubes.  The dirtier you get, the easier it will get to tell them from tubes.  Telling a JFET from a MOSFET, at any distortion level, is a bit harder.

The most satisfactory arrangement I've come up with yet is to use FETs as clean boosters which overdrive a tube.  You get some of the best of both worlds that way, since the early, clean stages are compact, cost-effective and reliable, but you still get some real tube sound where it counts.  You may even be able to use 12v for everything, from FETs to tubes to heaters.  Or you could try 12v for the heaters alone, and 15-20 volts for the rest, for more gain, headroom, and better sound.  If you don't know what to do with the other half of your 12A*7, you could even do the tube part as a push-pull output stage.

There are a lot of possibilities with FET-tube hybrids, and in my opinion, that's the surest way of getting truly tube sound while using FETs.  A well thought out compromise.

As for which kind of FET sounds more tube like, I don't think I could declare either to be a definite winner.  The topology is going to have more impact than the FET type.

Again, all of the above is strictly my opinion, YMMV.

[petemoore]

  1rst gain stage I'm inclined to think if I could easily exchange tube relatively exactly for Jfet, I might have trouble deciding which is truly preferred.
  2nd stage, while working with the 1rst stages output [let's say it is many times the voltage swing of the 1rst stages input], at that point 9v of power supply and whatever is left for headroom [power supply minus what the Jfet needs to stay mostly linear..like tube would] may cause the limited headroom to have it's effects shown on the linearity of v2 signal output, the basic theory of many Jfet distorters, and tube distortion preamps.
  If the power supply allows sufficient headroom, and signal swing is big enough to make the output tubes distort, me guesses tube amps with clean tube preamps could have the clean preamp part replaced by jfets and mosfets.
  Taking the 5e3 as example, the preamp is clean but pushed the power tubes to distort, having not tried it I can only imagine it will sound different by a noticable amount, ...and not microphonic .
  With mosfets and Jfets of gain variety, and sockets...
  I don't know what MM amps used for the....amp with SS preamp and 2x6l6 output had for actives in the preamp, amp sounds fine when it was working, the toastedly crisp looking 6l6 I found in the bottom of the cabinet at dudes gig kind of put that one on the shelf, tube failure in this case was because of downhanging tubes, no tube retainers and loose tube sockets with hot tube falling out of them.
 

[richon]

so, finally no-one really likes MOSFETs and FETs as replacement of Tube stages

and it would be only advisable to use them just as super clean BOOSTERs?

[CLAAS]

Leaning by doing

[ambulancevoice]

Hi,

hopefully noone gets angry because I show this picture from www.runoffgroove.com


-if i want this stage to have more GAIN, should i increase the 1K resistor o decrease it?? (i guess i shoud decrease it)

yep decrease

[mojo_hand]

so, finally no-one really likes MOSFETs and FETs as replacement of Tube stages

and it would be only advisable to use them just as super clean BOOSTERs?

If I had to pick any solid state, analog device for modeling tubes, FETs would be it.  I'd choose JFETs over MOSFETs most of the time, both because they're easier to work with in some ways, and because I'm more familiar with their theory.  But they have their limits.  You've found runoffgroove, which is probably the best collection of designs like that which there is, complete with instructions on using that same design process yourself.  Want to do a '62 Deluxe in FETs?  The answers are all there.  You might want to check out Doug Hammond's site, too, now that he's gotten most of it moved to its new location.

But there are things you simply cannot do.  With everything set up just right, a JFET can sound a lot like a generic triode.  But can you make it sound just like a Sylvania 6L6GC?  How about a Mullard KT88?  I don't think that anyone would claim that.  Yet the sound of a particular amp is usually very dependent on the tubes used, and few argue that one cannot tell the difference between brands of tubes, at least some of the time.  We're nowhere near to being able to nuance FET sounds that completely.

And there are problems in theory, too.  Some combination of the Doug Hammond/ROG school of FET design (replicate old tube schematics) with Danyuk's theory would seem like our best shot at sound like that, but with most schematics that cannot be done.  If the tube gain stage has a cathode resistor of 3.9K, gain of 35, and will not clip with any input signal less than 3v, you are out of luck, there is probably no FET made which can do that under any circumstances, and those which would come closest could not do so within the triode mode of operation.  You can follow a schematic like that with FETs -- the "Supreaux" is proof of that, and I happen to think it's a great success at what it tries to do.  But there is no FET which will work right and stay in triode mode with the 3.9K resistor called for.  Danyuk also says that bypassing cathode resistors with caps is a no-no if you want to stay in triode mode, yet most old schematics require that, and FETs usually have so much lower gain than most tubes that you almost cannot reconcile theory with a 1950s tube topology.

So I'm not saying that it can't be done, the sound samples at runoffgroove are proof that it can be, up to a point.  But there are limits you're going to run into, including ones that none of us have yet figured out a satisfactory way around.  If we had, we'd be spending 25 cents on materials to replicate $300 NOS tubes, and we'd be retiring next year.  But we just aren't there.  If you can get there, more power to ya.

[richon]

do you have the link to the NEW   Doug Hammond page to read about his "replication of old tube schematics"


i'm not giving up.... i've still have faith 

at least i'm learning about fets and mosfets a lot!!!

[mojo_hand]

He has been moving his stuff here: http://www.aronnelson.com/gallery/v/DougH/

But be patient, it's taking a little time.

[petemoore]

so, finally no-one really likes MOSFETs and FETs as replacement of Tube stages

and it would be only advisable to use them just as super clean BOOSTERs?
  3 mosfets and 4 jfets are doing work for me right now, this work could be done using tubes.
  Jfets can easily be made to distort in desirable fashion, exhibiting many similar characteristics of tube distortion, an important character is the roundness of the clipping 'knee'...how sharp are the curves where the clipping occurs during +/- wave excursions of signal swing, simply pressing a Jfet near it's power supply limit can cause frequency gain shifts and signal 'squashing'...run two high gain stages series @9v supply, or drop supply voltage with a single stage, press the Jfet to distort.
  Output tube distortion...another matter altogether. Rather than try to re-create this without tubes, I have created it with tubes. Expensive, but tried and verified. If I were to try non-tube-tube output type distortion, it would most likely be with Jfets as sort of an 'addendum' to the power supply and...everything else required as output stage [Ac mains/PT/rectification/Chassis/OT/etc.].
  But every time I've gotten that far [frame/PS/etc.], I've used tube type signal circuits. Not that Jfet or Mosfet couldn't do fine work and sound great, all the designs I've read about [that's some googling too] for SS outputs puts headroom well above clipping levels, clipping outputs are consistantly regarded as undesirable and ka ka.
  For clean signal voltage boosting...get headroom, 9v seems plenty to do this to a guitar pickup source.
  Want more add more, raise headroom/power supply voltage as necessary, make a clean preamp...
  For distortion...info a plenty on that, try Mu Amps or seriesed Fet Stages, Fet @ low voltage, CMOS...just get the + or -,  or both halves of the signal swing to 'press' on a parameter outside of linear amplification of your fet [of other] device and distortion results.
  A next big step might be using a transformer to couple Jfet to speaker, but I havent' found much as far as Jfet output amps, I made a Mosfet amp, works great, but is designed to be linear.
  My take on it is that Jfet/Mosfet preamps, boosters, OD's, and distorters are where the real mileage is to be wroght from the devices, tubes can't really be made to distort like SS devices...lol...[haven't really tried..].
  I'm about to order some IRF820's, I feel they could easily replace 'V1, 12a_7' [first two tube gain stages] in many of the amps I'm using now, and would make great utility as source followers, probably could be used to good result as 'V2' [gain stages 3&4] in clean preamp design to push tube output stage hard.

[R.G.]

  I'm about to order some IRF820's, I feel they could easily replace 'V1, 12a_7' [first two tube gain stages] in many of the amps I'm using now, and would make great utility as source followers, probably could be used to good result as 'V2' [gain stages 3&4] in clean preamp design to push tube output stage hard.

Instead of the IRF820s, try some of the Zetex and Supertex TO-92 high voltage MOSFETs. They have 500 and 600V MOSFETs in TO-92 package. Mouser stocks them.

Supertex has *depletion mode* MOSFETs, which do bias in the same way that pentodes and triodes do.

For those of you who have not yet, go read "MOSFET Follies" at GEO, http://www.geofex.com