MOSFET source follower for output stage of AB763 blackface

[R.G.]

I don't understand how the cathode follower is set up so it is at the normal negative bias of the output tube. Unless this is done by choosing a cathode bias resistor that will do so by resisting the -100V supply??? Again, in the case of the MOSFET, a source resistor instead of a cathode resistor from source to the -100V supply.

Let's take a step back and think a bit.

Everything needs certain inputs of the right kind and amount to function correctly. For instance, a gasoline-engine car needs a supply of proper-octane gasoline. If you feed it diesel oil, it doesn't work no matter how much diesel fuel it gets. Likewise for diesel engine trucks and such. Gasoline is not a good fuel for them.

A power pentode or power beam tube grid needs to be fed a voltage somewhere between the voltage at the cathode and tens of volts more negative. In typical fixed bias amps, the cathode is very near 0V, so the grid needs fed between 0V and tens of volts negative. At some point, a tube like the EL34 or 6L6 will go into cutoff when its grid voltage is way negative, perhaps -60V for an EL34 and -50V for a 6L6.

The useable range of grid voltages is then 0V to maybe -60. In a normal amp, they are held there in no-signal conditions by a resistor to bias supply, which is set just a little more positive than the cutoff voltage.

If we want to drive that grid with the source of a MOSFET, we have to arrange the MOSFET so its source covers that range of voltage that the grid needs. And the MOSFET itself needs fed what inputs IT needs to work right.

So the MOSFET source has to be able to go to maybe -60V compared to ground. The only good way to do this is to make a negative supply bigger than that, maybe -80 to -100V, and put a resistor between the source of the MOSFET and that new power supply. The drain of the MOSFET also needs to be supplied with a voltage higher than the source under all conditions. So the drain needs a voltage significantly more positive than 0V.

When these are set up correctly, the source of the MOSFET will *follow* the gate voltage on the MOSFET, but a few volts lower because of the threshold voltage the MOSFET needs to operate. So to make the grid of the power tube which is being fed by the MOSFET source be at -37V, the MOSFET gate will be biased to maybe -33V, the source will be maybe 4V (a guess at a threshold voltage for the MOSFET) lower than that, and the tube grid will be happy.

There has to be a power supply more negative than the MOSFET will need to pull the grid because the MOSFET doesn't pull the grid negative, it lets the source resistor pull the grid down. When the grid needs to go higher than the bias voltage, the MOSFET conducts more current and pulls the grid up.

Is this making sense?

[SixStringBender]

HAHA I was thinking positive would be going to ground if bridge rectifying the line voltage, but really hadn't figured out weather or not it would be a problem. Like duh! lol

So I've looked for transformers and I'm having a hard time finding one. The closest thing I've found from mouser is 553-F7-120. I guess that'll be enough voltage for JJ 6V6Ss and probably 6L6GCs if I ever put 6L6GCs in. I'll keep searching though.

Edit: okay, after reading R.G.'s post I missed above, I understand the 120V transformer won't work because it doesn't have double the needed voltage. Maybe not even center tapped too. I have found some step up control transformers that are 115V primary and 230V secondary. If they have a center tapped secondary it should be 115V-0-115V. I think the ones that have a dual secondary in the description that are 115V-230V are center tapped. I have not a clue what current rating I need.

About the source resistor. What value is it? My wife didn't see it on the schematic. Like I said, my helper is clueless to this stuff, lol and I'm almost blind. Unable to see to read for 10 years. Thankfully screen readers exist.

I'm gonna gather what knowledge I have of this, have her draw a schematic with my description and put it in text so you guys can tell me if I have this correct.

I'll probably do the octal socket and plug in SS rectifier idea and put a small board in the amp with the filter caps, current limiting resistors and bleeder resistors on it. I think it should all fit between the PT and main board inside the amp.

Moving to the bottom of the chassis now. I'll have to move the power supply board and cap pan cover over to make room to move the OT to put the new PT for the MOSFEts.

I'll probably mount the mosfets on heat sinks to the chassis (insulated) and wire everything to them between the board and output tubes. I'll lay it out as neatly as possible. Sounds like fun.

See anything wrong with the plan?

[SixStringBender]

R.G., sorry I didn't catch your last reply before I typed that last message.

Yes sir, that sure explains a lot. That is how I was thinking it was working after martin's description last night and after I reread your article a few times last night. That certainly paints an even clearer image in my mind. I very, very much appreciate you taking the time to give me a lesson. The description of the negative supply enlightened me.

I assume the source resistor value will have to be calculated once I know my negative voltage.

What is the current rating the transformer needs to have?

Is 47uF at about 200V a good smoothing cap for the power supply to these MOSFETs? I may or may not include bleeder resistors. I assume the tubes will drain the caps when the amp is powered down with hot output tubes.

martin manning, I appreciate your lesson as well. You have been a great help

Kagliostro, I haven't forgotten your reply. I haven't had a chance to go over it yet. Can you give a description?

I'm finally enjoying school and there aren't even any gals here. lol

[martin manning]

I put a new version of the schematic below showing a bipolar power supply for the MOSFET's and revised values.

For a transformer, you are probably limited to voltages of either 120V CT or 230V CT. 120V CT would yield about +/- 80V DC, which would be cutting it pretty fine for a -37V power tube bias requirement. I'd go for 230V CT, which will give ~+/-160V. This Triad flat-pack looks pretty good: http://www.mouser.com/ProductDetail/Tri ... hTyFgQA%3d
These are pretty compact, 2" x 2.5" x 1", and it could be mounted to a small eyelet board along with the MOSFET's and other parts. It's rated 50mA at 230V, but I believe that would go down if it is connected as a bipolar supply; you are using the whole winding full-time here, like in a FWB rectifier.

The source resistor value is not too critical, but t will set the current required to hit the desired bias point, which I would keep to 3-4mA. I suggest 33k at 1W.

This circuit should work for larger amps, should anyone else want to try it.

[Kagliostro]

I'm not so skilled as to explain much about the circuit, however I try to give a descriprion

The schematic is relative to an old Geloso Public Adress amp, model G.1/1110 A, that uses 3 12a*7 preamp tubes and 4 EL503 power tubes

V1a and V1b (a 12ax7 tube) with catode connected directly to ground, are devoted to the 2 input (it is a microphone amplifier) followed by a volume pot

between V1 and V2 there are phnono intakes followed by a volume pot

V2a (a 12ax7 tube) is used as an anode follower to drive a James Tone Control

V2b (the second triode of V2a) is an anode follower gain stage

V3a (a 12at7 tube) has an NFB connection to the cathode from a dedicated secondary winding on the Output Transformer and is an anode follower gain stage

V3b (to me) is arranged as a cathodine PI

and here there is the difference respect the usual arrangement, the signal from each branch of the PI didn't feed directly the power tubes grids

the signal feed via a 0.1 capacitor in series with a 22k resistor the base of a BC147 (NPN transistor), the collectors of the two transistors are joined and supplied with a

positive voltage obtained from the filament winding of the Power Transformer, with a single diode followed by a 12v electrolytic capacitor (I think there are near 8.4v

there)

on the opposite side of the 22k resistor who is connected to the base of the BC147, at the junction with the 0.1 capacitor that feed transistor with the signal, there is

connected a 150k resistor (one for each BC147) that is connected to the Bias circuit

The power tubes are 4 EL503 and the catode of each pair is connected to gronund with an element formed with 3 in series parts, the first seems an inductor, the second is a

5ohm 2W resistor and the third is a fuse

I tried to give the description at my best, if someone can explain better he is welcome

K

[SixStringBender]

martin, I very, very much appreciate this. We'll try to look over it tonight so I can get a part list and try to get everything on order.

Kagliostro, I'm searching for a huge cleanish hard driven sound. Ultimately I want a Dumble SSS clone, but I imagine the MOSFET source follower will help my amp get closer. The 64 Vibroverb Custom mods took it a long way, but as always I want more. The mid 80s tone from SRV is my favorite tone of all. A lot of people like his early 80s tone, but me, not so much. About any healthy blackface amp can get that tone with a few tweaks. Pull V1, disconnect the tremolo, install a SS rectifier and stiffen the power supply and your there. In a DR change R67 and R68 for the Super Reverb and Vibroverb values and install a larger OT and even a DRRI is there. The preamp mods are what takes the blackface more toward his mid 80s sound from that point. It takes some of the extended bass and all the honk out of the tone. The honk is what I don't like about his early tone. Most people rave about El Mocambo, but his tone sounds honky at El Mocambo to me. However, his playing is great in that show, as it always was.

[tictac]

If you really want SRV-like tone you'll need to use 6L6GC power tubes; the 6V6 is a nice tube but it won't have the clarity and headroom of a 6L6....

If you're talking about changing OT's in a Deluxe Reverb then you may as well change PT's too since the 6L6 has twice the filament current of a 6V6 and the Deluxe PT runs hot as is....

That being said, I've been interested in direct-coupled amps for a long time and am really interested how your project will turn out. I have a schematic of a mosfet drive circuit that was on the web about 12 years ago but haven't had the situation to actually use the circuit. Martin's looks like it will work well, hope everything works out for you....

TT

[SixStringBender]

TT, there is something I like about the JJ 6V6S. I've tried Tung-Sol 5881s and the JJ 6L6GC in my BillM modded Blues Junior and I just don't like them as much. Too loud mostly. The JJ 6V6S just sing at a volume I can handle. This changing OT taps to reflect the proper primary too.

I want to eventually try KT66 in this amp (remove tube retainers). Before I do it'll have an even larger Super Reverb size Classic Tone OT with 4, 8 and 16 ohm taps (40-18001) and most likely the Allen TP35 Heyboer. I'm gonna put the classic tone 40-18003 choke in it if I do that too. BTW, 8 ohm cab, 4 ohm tap yields 8K primary for KT66s. I have an 8K primary with my Vibroverb size OT (40-18006) for the JJ 6V6S.

Martin, we went over the schematic. I have the MOSFET install figured out. However, I need help with the bipolar power supply itself. How to rectify it and filter/smooth it. I'm trying to study it, but I'm having difficulty learning it.

Also, what is going on from R57? The description my helper gives me of R56 to the bias circuit sounds the same as stock, but the path from R57 sounds different. I thought all that remained unchanged.

Here is the description of what I do understand. I'll leave out V6B through V8 since it all connects the same way as V6A through V7.

Here goes. V6A's plate and R56 to C27 (as it is stock).

Through C27 to a fork in the road. One fork to R56 (where the other side of R56 goes to the bias supply that I assume remains unchanged).

edited for zener: Other fork goes to our new 220R 1W gate resistor, through the 220R 1W resistor to the anode/blank side of the 12V 0.5W Zener and then to the MOSFET's gate.

Source of MOSFET to cathode/band side of the 12V 0.5W zener, then to R60,. Other side of R60 to control grid of V7.

Source of the MOSFET, before the R60 connection, also goes to a 33K 1W resistor. Other end of 33K 1W to the -160VDC bipolar power supply.

+160VDC, from bipolar power supply, to drain.

Now how to build the VDC bipolar supply from the transformer? lol I know how to connect it to the MOSFETs once I have it set up.

Again your help is very much appreciated. I know it takes time to do this, so I appreciate it.

[SixStringBender]

I've been studying. I think this does the bipolar power supply.

First I'll describe the build of the bridge rectifier.

Put the leads of the band side of two rectifier diodes together in a V pattern (leaving the blank side of each apart).
Starting with two fresh rectifier diodes, put their blank side leads together in a V pattern (leave the band leads apart).
Now having two sets of rectifier diodes, connect the leads to finish a diamond pattern.
Blank lead to band lead on one side and do the same for the other side. Connecting the blank lead to the band lead.
That completes the bridge rectifier.
The band junction is the VDC positive output.
The blank junction is the VDC negative output.
The blank/band junctions are the inputs for the VAC.

Now this is how I think I connect the secondary of the new power transformer to the bridge rectifier.

The center tap of the secondary of the power transformer goes to ground.

One AC secondary tap to the AC input (band/blank junction) on one side of the bridge rectifier.
The other AC secondary to the other AC input (band/blank junction) side of the bridge rectifier.

Lets protect the bridge rectifier with two inrush current limiting resistors. I'll call them R80 and R81

R80 will go to the band junction of the bridge rectifier (positive).
R81 will go to the blank junction of the bridge rectifier (negative).

Smoothing the VDC with polarized electrolytic caps. I'll call the caps C50 and C51.

C50's positive lead will connect to the loose lead of R80. C50's negative lead will connect to ground.
C51's negative lead will connect to the loose lead of R81. C51's positive lead will connect to ground.

I think installing bleeder resistors in parallel with the electrolytic caps may be a good idea for safety since we have so much power in reserve. I'll call them R82 and R83.

R82 will go in parallel with C50.
R83 will go in parallel with C51.

Connecting to the MOSFETs.

The positive side of C50 goes to the MOSFETs' DRAIN.
The negative side of C51 goes to the 33K 1W resistor for the MOSFETs' SOURCE.

Is this right? If so I am amazed! lol If this is correct, I now need help selecting rectifier diodes, and help calculating resistor and cap values. For me, the hard part is over if this does it. Understanding it and getting a mental image is the hardest part.

Can the transformer center tap and filter caps go to the chassis for ground or should I put them to the PCB ground?

[Kagliostro]

@ SixStringBender

I posted the schematic only for the presence of the BC147 transistors between the PI and the grid of the power tubes that remember what you are trying to do with the mosfet

not because you must have an interest on the whole schematic of the amp

Ciao

K

[martin manning]

SSB, your description of the bridge rectifier circuit looks correct to me. IN4007 rectifier diodes will be fine, and I don't think you really need a current limiting resistor.

Note that the transformer I linked above has two 115V secondaries, which will be connected in series, and the series connection point becomes the CT. The phase relationship is important too, so consult the spec sheet.

The description of the MOSFET's looks good too, except it wasn't clear how the Zener protection diodes are connected. They are simply placed across the gate and source, with the cathode (banded) end to the source. A 1A Zener would be fine.

On the bias supply, I duplicated the existing setup so that there would be separate bias controls for each FET (and output tube), each connected to the 220k resistor going to the output side of the coupling caps. That may not be necessary, but it's nice to have.

[R.G.]

A 1A Zener would be fine.

Yep. In fact, even a smaller current zener works for the gate protection. There's not a huge amount of current in most transients, so even the little half-watt, $0.10 zeners are OK.

On the bias supply, I duplicated the existing setup so that there would be separate bias controls for each FET (and output tube), each connected to the 220k resistor going to the output side of the coupling caps. That may not be necessary, but it's nice to have.

MOSFETs have varying threshold voltages, so you might get two that were not from the same wafer and one is a volt or two higher or lower than the other. I think your caution in providing a separate bias adjuster per MOSFET is more of a necessity than a nice-to-have. Good work.

[martin manning]

Thanks for looking over my shoulder, R.G., that's exactly what I was thinking with the dual bias supplies. And thanks for the Follies page too, BTW... Good food for thought and experiments by many folks!

If I were doing this I'd be inclined to put the two bias supplies on the same board as the FET's using trim pots, making a nice compact and self-contained unit.

[tictac]

Here's something I can across about ten years ago or more....

It was around the time RG wrote the Mosfet Folllies and I think he helped this fellow get this circuit working...

It has an alternative bias supply that's not as refined as Martin's but may work for some who want to use the existing supply....

TT

[SixStringBender]

WOW, I'm so glad I have you guys helping me with this. I can't believe I learned that and got it right that quickly. lol I read a few sources after my wife went to sleep and there was an article by R.G. that helped me understand how the caps connect in series to ground to form the DC center tap ground reference. Or at least that is how I interpreted it for myself because the article just said the caps are in series to ground.

martin, yes thank you for the verification of the 12V Zener across the GATE and SOURCE. I wasn't sure about that and I was going to ask about how it was oriented. I'll edit my post in case there are text/verbal learners in the future. It makes more sense what it's purpose is now.

I figured it didn't matter witch secondary lead, other than CT, goes to either AC common input (band/blank joint) of the bridge rectifier so long as the center tap goes to ground.
I thought the bridge rectifier did all the separation, so to speak, of the positive and negative by blocking positive and allowing negative on one side and by blocking negative and allowing positive on the other.

I'll study the bias supply and put two 10K 25 turn trimmers in it. I hate touchy bias pots. I had been looking for a chassis mount one to replace the stock one for a while and couldn't find one.

What value electrolytic caps and bleeder resistors do you-all recommend for the dual power supply?