Mosfet OT saturator!! Yarrrr!!!

[lowell]

I have an idea that I want to try. Basically I've found a circuit in Teemuk's awesome book "Solid State Guitar Amplifiers" that I want to build and use to saturate an OT. If anyone has this book it's Figure 3.10-A on pg. 66.

The circuit:
It is a uncommonly-used solid state power amp with a center-tapped OT in push pull configuration. The input is a split load PI that is capacitively coupled to 2 totem-poled output transistors.

What I'm after:
I think I'd like to use MOSFETS for the power transistors cause I believe they exhibit more tube-like characteristics... though maybe I'm wrong there.
I want the output to be 200mv-1v for feeding into an amp input or power amp in - not a speaker. So we're not talking a lot of power here.

I've never designed or built a solid state amplifier, so I apologize for the questions as I pretty much don't even know where to start:

Being that I'm not after volume or power, will I need a VAS before the split load or will a 400mv humbucker signal be sufficient?
What transistors would be good for this project?
What should my Vcc be?
What OT should I use?

[loudthud]

A good place to start is with the OT. Knowiing the primary impedance and power rating will help define Vcc and the current the output transistors need to operate at.

[J M Fahey]

Hi Lowell.
Whenever you refer to something with words like

If anyone has this book

and

Figure 3.10-A on pg. 66

, please always provide a link, so we all talk about the same.
Different versions may not match at all.
Thanks.
PS: providing a working link might encourage someone who does not have this book yet, to download and read it

[lowell]

My bad JM,
If found a link: http://www.thatraymond.com/downloads...ttala_v1.0.pdf. I also figured the solid state gurus would catch the drift of the basic circuit from my description.

[teemuk]

please always provide a link, so we all talk about the same.
Different versions may not match at all.

They do in this case. At least for now.

There's really not much to see, though: The figure depicts a basic power amp circuit similar to ones used in countless tube amps - only the tubes have been replaced with NPN transistors and a required biasing scheme is added.

It's really just a crude example of how the overall circuit would be constructed. Seen a SS radio schematic from early 1960's? That's it.

The magic of the circuit doesn't really lie in using the output transformer but more in the overall difference to generic solid-state power amps:
First of all, the output stage is configured as a voltage amp loaded by the OT's primary - this kind of circuit will behave very differently than the typical emitter follower unity-gain buffer output in many aspects. Secondly, the voltage amp driving the final output stage is push-pull, instead of being single-ended. Thirdly, the OT limits the amount of global negative feedback one can use, which results into higher output impedance and increase in overall distortion. The clipping distortion will also be less sudden.

Trace Elliot (and according to 1979 patent Fender as well) have built circuits like this with MOSFETs, few others with basic BJTs. Peter Hartley constructed a low current version with JFETs in 1979 and used it in the preamp of his Hartley Thompson amps. High voltage is not mandatory (the OT's turns ratio defines that) but finding a proper output tranformer will be a problem.

[lowell]

Thanks for the info Teemuk.. and for writing the book, it's awesome. The little I know of transistors, or jfets specifically, is that if Vgs is 0v then it is fully on. Therefore, it's much like a tube in that either a negative gate bias is needed, or the source voltage raised to get some negative bias voltage and reduce quiescent current. I also believe that the gain of a jfet is Rd*gFs... I may be wrong here. So the bigger the drain resistor the more gain, to a certain extent.

Let say we have a jfet with a gFs of 1000mS. We want a gain of 10. Our Rd would then be 10/.001=10k ohms. So then with this power amp would I want the DC resistance of the primary of the OT to be 10k from center tap to one end of primary?

I should also note that the purpose of the circuit that I'm building is not clean power, it's to saturate the OT when cranked, and get a max of 1v p-p on the secondary.

[Steve Conner]

Didn't Gallien-Krueger do this in the 70s? I'm sure they made a solid-state amp with an OT, and it was based on the kind of "tone through topology" school of tube amp modelling. The schematic has been posted on this forum before.

I've seen a hi-fi amp with a trifilar wound 1:1:1 OT.

I've made circuits myself using an EL84 driver and Champ OT, driving the bases of a couple of bigass power transistors that drove the speaker directly. (There are a few tricks I added that I don't want to publish, but they don't affect the general argument.) My prototype made 60 watts and sounded very similar to the tube driver circuit "barefoot", just a hell of a lot louder.

An output stage like this, or the kind you saw in Teemu's book, gives current drive to the speaker, just like a pair of pentode tubes. Then you can stick global NFB around it, and it reacts just like NFB does in a tube amp. That's probably the most important thing about it, tone-wise. And obviously that's null and void when you use it as a line-level effect, it needs to interact with the speaker impedance.

To a first approximation, no transistors exhibit tube-like characteristics. Those that have useful characteristics are so variable with manufacturing and temperature that you have to stick lots of NFB round them to stabilize the operating point. But that also removes the non-linearity you were trying to make use of. This is another important issue (annoying Catch-22, even) in solid-state amp design.

The most "tubey" soft clipper I've seen to date is Marshall's old "diode inside a bridge rectifier". They chose high voltage, high current diodes because these have soft knees at lower current.

[teemuk]

Didn't Gallien-Krueger do this in the 70s? I'm sure they made a solid-state amp with an OT, and it was based on the kind of "tone through topology" school of tube amp modelling. The schematic has been posted on this forum before.

Yes they did. The output devices in those were BJTs. The PI stage was the typical LTP you find from 90% of amps - both tube and solid-state. The trick was: GMT - Gallien-Krueger used that circuit as a PI (unlike most SS amps) and coupled the collectors straight to the output devices without the single-ended voltage amp in between. Run that way, and with little NFB, the circuit actually clips quite softly. Furthermore, when the push-pull circuit sums the two (clipped) out-of-phase signals together it also "error-corrects" and softens clipping.

Trace-Elliot amps had the basic OpAmp -based PI - similar to that of Music Man hybrid amps. Their trick was to fit Zener diodes into the feedback loop of the first OpAmp to get gradual soft-limiting.

MOSFETs don't really behave like most power tubes. That's mostly a myth. You can get them to clip moderately softly when you operate them in common source and with little global negative feedback - preferably even open-loop. This is practically how most MOSFETs in generic SS amps are never operated. The generic transformer coupling style, on the other hand, forces you to use them that way and the device characteristics to clip more gradually then ordinary BJTs can shine out.

[J M Fahey]

Nice discussion, that's why I love this Forum.
I think that Steve Conner's approach was the best, from a practical standpoint, and it does not even need an output transformer , because the "floating" or "ground-independent" drive provided by separate secondary windings puts the load always in series with the transistors.
As a side note, I have built and sold MosFet push-pull ; real output transformer (connected drain to drain) in "tube" topology .... and got crazy with non linearity, horrible crossover distortion, etc; which could not be sufficiently feedback corrected because of the low instability threshold caused by the output transformer.

[lowell]

Ok Steve can you further explain your circuit? Were your power transistors connected in parallel and was the output from the collectors as in common emitter arrangement? Then the collectors tapped via caps directly to the speaker? Or were they in push-pull? If they are in push-pull wouldn't you need an OT to combine the signal back together? Why the OT before the power transistors, more current drive? Also, if the circuit was push-pull I'm not sure how the champ OT would provide 2 out-of-phase drive signals..?

[lowell]

Question: Can I incorporate this transformer into my circuit? It's a 1:1 ratio so what goes in voltage wise is what goes out correct? Does a 1:1 ratio still increase current? My guess would be no but maybe I'm wrong. From what I've read of these they are simply a way of transferring the AC and ridding of ground loop issues. I'd like to try running some jfets in push pull common source into this if possible... or is this pointless and I should just stick with my first project being transformerless.

Transformer 1:1 For Direct Boxes

1 other q:
Can you have a push pull circuit w/o a transformer? How do the signals recombine and not phase-cancel if so?

[J M Fahey]

Hi Lowell.
1) That direct box transformer is just that, it won't change your sound precisely because direct boxes are supposed to be "transparent", pass sound through without change.
Besides, for that price it can't be good; I used some Jensen Transformers in my mic preamps (for AKG C451, used in my home-made speaker measuring chamber ) and paid around $50 each.
2) Yes you can, search for the Peavey patents, they do just that.

[Wilder Amplification]

Question: Can I incorporate this transformer into my circuit? It's a 1:1 ratio so what goes in voltage wise is what goes out correct? Does a 1:1 ratio still increase current? My guess would be no but maybe I'm wrong. From what I've read of these they are simply a way of transferring the AC and ridding of ground loop issues. I'd like to try running some jfets in push pull common source into this if possible... or is this pointless and I should just stick with my first project being transformerless.

Transformer 1:1 For Direct Boxes

1 other q:
Can you have a push pull circuit w/o a transformer? How do the signals recombine and not phase-cancel if so?

First off...why would you want OT satuation? There was a thread on here long ago about OT saturation and from most peoples' opinions on that thread as well as my own, OT saturation doesn't sound very good. It was also stated that "OT saturation" was up there with "Tube watts vs solid state watts" and all the other sales pitch schemes commonly used at Guitar Center as well as countless other places.

Most modern day SS amps are an output transformerless design (OTL). They call it "complimentary symmetry push pull". It's been done one of two ways that I've seen. The less common way is to use a single rail supply with the collector/emitter joining point (the output) biased at 1/2 the rail voltage and the speaker is capacitor coupled in this circuit. The more common approach is to use NPN/PNP current carrying transistors driven by two pairs of transistors arranged in a darlington pair with a dual rail supply (equal voltages of opposite polarity).

[Steve Conner]

http://music-electronics-forum.com/t16762-2/

That was a good thread on OT saturation. I contributed a bunch of measurements and scope shots starting at post #54.

The project is still worth trying, if only so you can see saturation with your own eyes and get a feel for what it is and what it's good for.

Solid-state amps can be push-pull without a PI or output transformer, because of the existence of PNP transistors, which are the mirror image of NPN ones. There were never any "PNP" vacuum tubes, to make one you would need an "anti-cathode" that emitted positrons.

[Wilder Amplification]

http://music-electronics-forum.com/t16762-2/

That was a good thread on OT saturation. I contributed a bunch of measurements and scope shots starting at post #54.

The project is still worth trying, if only so you can see saturation with your own eyes and get a feel for what it is and what it's good for.

Solid-state amps can be push-pull without a PI or output transformer, because of the existence of PNP transistors, which are the mirror image of NPN ones. There were never any "PNP" vacuum tubes, to make one you would need an "anti-cathode" that emitted positrons.

Don't you mean protons?

Maybe if we...oh nevermind.