Jones Calling DougH and JoeD! and ROG - Pentode preamp simulation!!!

[mjones99]

Guys, I have been thinking about this for a while, wouldnt it be great to really get close to the sound of a 6sj7 or ef86 in solid state?  I dont know how possible it is, it seems that a 12AX7 half can be simmed pretty close (90%+) with jfets, but the pentode is another animal.  The screen compression is a major part of that sound, the problem it seems to me is how to get that effect without going into an overly complicated design.  I saw Joe's Vulcan and speech clipper snippet (awesome work Joe), Doug you have more experience than anyone I know at simulating tubes in solid state, ROG stuff is awesome, especially Thunderchief, both Doug and I have a ton of experience using and designing real pentode preamps, I was thinking between us and anyone else who'd like to join in, maybe we could get to 90%+ of the pentode preamp sound.  How about some sort of variable local negative feedback that increases with signal strength?  Is it possible to vary the V+ inversely to signal strength on one device without involving others?  I am solid state retarded, you know this Doug...  I want to sim the Octal fatness, the whole thing, in silicon, I used the matchbox as a model, so far it aint even close, it sounds like an overdrive yes, but not like an amp with pentode in preamp.

Jones

[stm]

Hi,

Exactly what do you mean by "screen compression?"
Is there a place where this effect is described either sonically or physically?
I understand the voltage of the Second Grid in a pentode affects gain as well.  Do you think it takes an important part here?

I'd like to think that as JFETs are good for triodes if surrounding components are chosen properly, maybe a MOSFET would be good to implement a pentode.  After all, pentodes have higher gains and I understand are able to clip harder.

[mjones99]

Hi,

Exactly what do you mean by "screen compression?"
Is there a place where this effect is described either sonically or physically?
I understand the voltage of the Second Grid in a pentode affects gain as well.  Do you think it takes an important part here?

I'd like to think that as JFETs are good for triodes if surrounding components are chosen properly, maybe a MOSFET would be good to implement a pentode.  After all, pentodes have higher gains and I understand are able to clip harder.

Essentially what happens with the screen compression is that the harder the input of the tube is driven the more the screen voltage sags, causing gain loss, other wise the tube is a high gain amplifier, so it compresses, and depending on the screen resistor this can be a little or alot.

[Skreddy]

Yeah; I think the ROG circuits adapted for MOSFETs would sound much closer to the real thing than JFETs do.  They'll still sound MOSFETy, but IMO that's better than JFETy 

There are lots of typical LED/LDR compression schemes, plus the Vulcan diode compression scheme, that can be put to good use...

[mjones99]

I am thinking maybe using the Davis diode compression with a mosfet instead of jfet might be closer than just the mosfet, I am going to try that next.  I am out of passive parts so I have to wait till next week to get more resistors and caps from futurlec.  probably try using different types of diodes in the circuit as well.  I have just finished design on a new clean boost pedal using one tube and one mosfet, so I will be prototyping that next week or so as well.

[Skreddy]

I am thinking maybe using the Davis diode compression with a mosfet instead of jfet might be closer than just the mosfet, I am going to try that next.  I am out of passive parts so I have to wait till next week to get more resistors and caps from futurlec.  probably try using different types of diodes in the circuit as well.  I have just finished design on a new clean boost pedal using one tube and one mosfet, so I will be prototyping that next week or so as well.

Sounds interesting! 

[MartyMart]

You have to watch out with multiple Mosfet stages, I recently built Joe's Mosfet Obsidian
and tried "N" channel Mosfets, which introduced quite a bit of noise, P channels work better
as specified, BS250's.
Interesting idea .....

MM.

[Doug_H]

Sounds like fun, I don't have time for another design project right now. I tried the diode-compression op amp on the breadboard a while ago and it sounded pretty nice. I keep thinking, just build a compressor circuit and mod it for whatever distortion/amp/etc sounds you want. For the OF I wouldn't worry about trying to fit it into this jfet/mosfet "amp topology" stuff, but that's just my approach.

[mjones99]

Sounds like fun, I don't have time for another design project right now. I tried the diode-compression op amp on the breadboard a while ago and it sounded pretty nice. I keep thinking, just build a compressor circuit and mod it for whatever distortion/amp/etc sounds you want. For the OF I wouldn't worry about trying to fit it into this jfet/mosfet "amp topology" stuff, but that's just my approach.

Thats actually a good idea Doug, I think I may have been stuck in a RunOffGroove ;-) not that thats bad, btw this new tube boost is massive sounding, I am putting a 6SJ7/12SJ7 tube preamp in a pedal, my prototype was done in a loaf pan, but it will get put into a 4x6 hammond steel chassis, I will try to have the schem up soon, its on paper but I need to clean it up and redraw in DW4.  Its similar to a schem thats on my freewebs crap page but better sounding, and only very slightly more complex, I just hope AES keeps those small power xfrms rolling (saves like 3 bucks a shot and .8" inside the enclosure) every guitar player in town is going to want one here in Zanesville, not that theres that many, but building OFs can keep me fairly busy if I want it to, seems like everyone that hears my OF wants me to build them one, Have a great day Doug, and everyone.  Oh yeah, I am designing a new active direct box and preamp for bass, hybrid blendable tube/SS as well.

[Eb7+9]

Guys, I have been thinking about this for a while, wouldnt it be great to really get close to the sound of a 6sj7 or ef86 in solid state?  I dont know how possible it is, it seems that a 12AX7 half can be simmed pretty close (90%+) with jfets, but the pentode is another animal.

From a "characteristics" point of view a Pentode has similar transfer to a NPN Bipolar transistor on account of the saturation region knee - ignoring the participation of the sceen these are closest inter-device approx ... Triodes and jFET are not "that" similar since jFETs also exhibit Saturation (pinch-off) kneeing while Triodes obviously don't, and vice-versa Triodes have an input loading current that's exponential in respose to input voltage which jFET's don't have - what they do share otoh is soft cutoff ...

Something tells me it would be interesting to port the Fatness to a hybrid Bipolar-jFET design - up until now it's been pretty much either-or ... as for the output stage of the Fatness one could probably get away with a simple Bipolar stage and added some filtering as Doug has done in the Meteor and others ...

~jc

[Joe]

I think Jack Orman's Minibooster is a good bet for the input stage. I would run an extra 10M from gate to drain on the lower transistor and use an input cap. This keeps the gate from sticking, and the slight negative-feedback smooths the sound a bit. The diode trick also works with the minibooster.

Regular JFET stages would probably work for the triodes, except I would do the same thing with the 10M resistor, just sounds better that way. The simple-sim might be a good fit for the output stage, and provides a line-level output.

[mjones99]

Right on JOE!  You da man!  Great idea, isnt the minibooster alot like a cascode tube circuit?

[Joe]

Yeah, it's a similar type of circuit. But I thought of it more for the higher gain. Try the diode trick with the minibooster stage. If it lowers the gain too much, then raise the source resistor value (& use a proportionally smaller bypass cap.) Same for the second stage.

Both stages sound better with the diode IMO, it's not the magic answer to tube emulation, but keeps the transistors out of saturation.

Don't know much about output stages. I've read that having some DC in a transformer along with AC produces even-order harmonics. But I'm unsure about how to apply the idea to a pedal, or which winding to put the DC through, or how much. If there's enough gain, possibly a simple low-gain stage could be used, mainly to invert the output.

[mjones99]

I was actually thinking of using a small audio Xfrmr 3W or so, with a mosfet hooked to it like a tube would be, then take the signal off the secondary, at first, but that really adds some cost and real estate, I will probably try your simple sim for the output, If I can get like 90% of the sound I'll be really happy.  I HAVE been having great success with hybrid circuits using high voltage mosfets with a pentode tube, but that still requires a HV power supply, but is certainly less costly than a full tube circuit, thanks Joe!