MOSFET as Cathode Follower?

18watt-specific Tech Talk - Building, Fixing, Parts, Mods...

Moderators: CurtissRobin, colossal, zaphod_phil, Daviedawg, Graydon

Post Reply
User avatar
Shottky
Occasional poster
Occasional poster
Posts: 204
Joined: Sun 12/14/08 2:00 am
Location: Waterloo, ON

MOSFET as Cathode Follower?

Post by Shottky »

Have any of you had any experience using a MOSFET as a cathode follower replacement? I was thinking of dropping one of these into a Superlite TMB, before the tone stack. Some may consider it heresy to do such a thing to a tube amp, but if it doesn't colour the amp's sound and provides the necessary current boost, then I'm all for it.

Here's the link:
http://www.geofex.com/Article_Folders/m ... tfolly.htm

Thoughts and comments welcome.
0 x

tkdrew
Newbie
Newbie
Posts: 80
Joined: Mon 08/07/06 2:00 am
Location: Mountains of East TN
Contact:

Post by tkdrew »

I have used them in several amps, they work great for a buffer type stage. I used IRF820's and I honestly can't tell they color the tone at all. Here's a link to some clips from a couple of my amps with one as a cathode follower.



http://www.smittytoneamps.embarqspace.c ... 4533294005
0 x

User avatar
zaphod_phil
Builder, Admin
Builder, Admin
Posts: 14991
Joined: Wed 03/19/03 2:00 am
Location: YYZ

Re: MOSFET as Cathode Follower?

Post by zaphod_phil »

Shottky wrote:Have any of you had any experience using a MOSFET as a cathode follower replacement? I was thinking of dropping one of these into a Superlite TMB, before the tone stack. Some may consider it heresy to do such a thing to a tube amp, but if it doesn't colour the amp's sound and provides the necessary current boost, then I'm all for it.
I don't actually see any need to buffer the SuperLite's tone stack, as the amp already has tons of gain, and the insertion loss of the tone stack is used as part of the design, so that the PI doesn't get driven harder than it should.

Also while a MOSFET follower doesn't colour the amp's sound, a true DC-coupled cathode follower actually does colour the sound by soft-clipping the waveform - see the relevant section in www.coastin.com/AESPaper.pdf In the SuperLite design, I warm-bias the 2nd preamp stage to give the same kind of effect.
1 x
Nature abhors a clean tube amp

User avatar
Shottky
Occasional poster
Occasional poster
Posts: 204
Joined: Sun 12/14/08 2:00 am
Location: Waterloo, ON

Post by Shottky »

tkdrew wrote:I have used them in several amps, they work great for a buffer type stage. I used IRF820's and I honestly can't tell they color the tone at all. Here's a link to some clips from a couple of my amps with one as a cathode follower.



http://www.smittytoneamps.embarqspace.c ... 4533294005
Excellent! It's good to hear that you have had success with this approach. I'm just interested in finding affordable solutions to keep product costs down. I'm going to consider putting them in some future designs. At the very least, they're exceptionally easy to test.

zaphod_phil wrote:I don't actually see any need to buffer the SuperLite's tone stack, as the amp already has tons of gain, and the insertion loss of the tone stack is used as part of the design, so that the PI doesn't get driven harder than it should.

Also while a MOSFET follower doesn't colour the amp's sound, a true DC-coupled cathode follower actually does colour the sound by soft-clipping the waveform - see the relevant section in www.coastin.com/AESPaper.pdf In the SuperLite design, I warm-bias the 2nd preamp stage to give the same kind of effect.
Hi Phil, thanks for your input. I mentioned the superlite partially to get your attention as the main designer, and partially because I was curious as to how the design of the tone stack interacted with the first gain stages. It's good to hear that it is unnecessary in this particular design.

Thanks for the link to the paper - it looks like it contains some pretty interesting information, so I'll go give it a read.
0 x

User avatar
Shottky
Occasional poster
Occasional poster
Posts: 204
Joined: Sun 12/14/08 2:00 am
Location: Waterloo, ON

Post by Shottky »

tkdrew wrote:I have used them in several amps, they work great for a buffer type stage. I used IRF820's and I honestly can't tell they color the tone at all. Here's a link to some clips from a couple of my amps with one as a cathode follower.



http://www.smittytoneamps.embarqspace.c ... 4533294005
Did you heatsink the FETs to the side of the chassis? I'm planning the layout for my new octal preamp design and need to order parts.
0 x

tkdrew
Newbie
Newbie
Posts: 80
Joined: Mon 08/07/06 2:00 am
Location: Mountains of East TN
Contact:

Post by tkdrew »

Yes, I attached it to the chassis dead bug style. The IRF820 has voltage on the tab so be sure to use an insulator and a nylon screw/nut so as not to ground it out to the chassis. Be careful also if you use the thermal transfer grease, most types are fairly conductive. Just use a tiny amount in the center of the tab/insulator and don't coat the whole insulator.
0 x

User avatar
zaphod_phil
Builder, Admin
Builder, Admin
Posts: 14991
Joined: Wed 03/19/03 2:00 am
Location: YYZ

Post by zaphod_phil »

I have to say I still see little or no benefit with using a MOSFET as a substitute for a CF in a tube amp. In reality TMB tone stacks don't usually *need* to be driven from a low impedance. I'm now convinced that the benefit a real CF provides is actually in how it asymmetrically soft-clips the waveform, which is something a FET simply won't do. See the scope traces in http://www.freewebs.com/valvewizard1/dccf.htm
0 x
Nature abhors a clean tube amp

User avatar
JamesHealey
Newbie
Newbie
Posts: 96
Joined: Mon 04/14/08 2:00 am
Location: Unknown
Contact:

Post by JamesHealey »

Guess it's just another tool in the box of tricks for amp building, im sure in something like a Fender Champ converted to a 5W Plexi design the Mosfet would reduce the need to add another tube socket.

This is something Im considering this idea for anyway.
0 x

User avatar
zaphod_phil
Builder, Admin
Builder, Admin
Posts: 14991
Joined: Wed 03/19/03 2:00 am
Location: YYZ

Post by zaphod_phil »

Yes, it's a useful tool, but that was actually a poor example. The overdriven cathode follower is a crucial part of the classic Plexi tone, adding crunch and removing harshness. A MOSFET simply won't give either of those, since it doesn't have any grid current.

But as you said it is indeed a useful tool in some places. For example the 6SJ7 pentode boost pedal that I've bought all the parts for but haven't got around to building for the last couple of years. :lol: When I get round to building it some day, it will have a MOSFET follower to buffer the high impedance output of the pentode stage down to a nice low impedance to drive the guitar cable.

Or if you want to add a TMB tone stack to a single stage preamp. As I said before, TMB tone stacks don't actually need to be driven from a low impedance. However, the effect of the TMB stack's low impedance on the output of a typical preamp stage is to reduce its stage gain. Placing a MOSFET source follower between the preamp stage and TMB stack will allow that preamp stage to maintain its full gain level.

HTH
0 x
Nature abhors a clean tube amp

User avatar
Shottky
Occasional poster
Occasional poster
Posts: 204
Joined: Sun 12/14/08 2:00 am
Location: Waterloo, ON

Post by Shottky »

I was planning to use a MOSFET follower for a 2-channel 12ax7/6sl7 amp, since with two channels and a tube phase-inverter don't leave me any room for a cathode follower on my chassis.

What's the typical gain loss between the first gain stages and the Tone Stack output, when a MOSFET/Cathode Follower is present?

What is the typical gain loss when I plug the first gain stages into the Tone Stack WITHOUT a MOSFET/Cathode follower?

Basically, I don't know how to account for the standard Marshall Tone Stack Rload for calculation purposes.

I'm worried about pushing the 6SL7s too hard (as they seem to go microphonic), but I'm also worried about not getting enough gain through a 6SL7 preamp and 6SL7 phase inverter if I DON'T include the MOSFET follower. I decided that it may be best just to install a switch to swap it in and out, and see which I prefer.
0 x

User avatar
zaphod_phil
Builder, Admin
Builder, Admin
Posts: 14991
Joined: Wed 03/19/03 2:00 am
Location: YYZ

Post by zaphod_phil »

A two-stage 6SL7 preamp will have plenty enough gain. In the SuperLite TMB the 12AX7 actually gives too much gain, so we have to attenuate the signal between the two preamp stages using the 470k/470k resistive divider. Using a 6SL7 you may just want to tweak that divider to attenuate a bit less so you get the same overall gain as the 12AX7 channel. You can certainly put in a MOSFET follower, and it won't hurt, but IMO it's not really needed. In a SuperLite type amp part of the challenge is not too hit the PI too hard, and the tone stack really helps with that.

BTW you can also model various tone stacks, and how they behave when driven from different impedances, if you download Duncan's free Tone Stack Calculator from duncanamps.com
1 x
Nature abhors a clean tube amp

User avatar
zaphod_phil
Builder, Admin
Builder, Admin
Posts: 14991
Joined: Wed 03/19/03 2:00 am
Location: YYZ

Re:

Post by zaphod_phil »

Shottky wrote:
Fri 08/21/09 2:32 pm
tkdrew wrote:I have used them in several amps, they work great for a buffer type stage. I used IRF820's and I honestly can't tell they color the tone at all. Here's a link to some clips from a couple of my amps with one as a cathode follower.



http://www.smittytoneamps.embarqspace.c ... 4533294005
Excellent! It's good to hear that you have had success with this approach. I'm just interested in finding affordable solutions to keep product costs down. I'm going to consider putting them in some future designs. At the very least, they're exceptionally easy to test.

zaphod_phil wrote:I don't actually see any need to buffer the SuperLite's tone stack, as the amp already has tons of gain, and the insertion loss of the tone stack is used as part of the design, so that the PI doesn't get driven harder than it should.

Also while a MOSFET follower doesn't colour the amp's sound, a true DC-coupled cathode follower actually does colour the sound by soft-clipping the waveform - see the relevant section in www.coastin.com/AESPaper.pdf In the SuperLite design, I warm-bias the 2nd preamp stage to give the same kind of effect.
Hi Phil, thanks for your input. I mentioned the superlite partially to get your attention as the main designer, and partially because I was curious as to how the design of the tone stack interacted with the first gain stages. It's good to hear that it is unnecessary in this particular design.

Thanks for the link to the paper - it looks like it contains some pretty interesting information, so I'll go give it a read.
The latest Rev Superlite TMB''s tone stack (with the 56k slope resistor) is optimized to work efficiently when anode -driven. So it doesn't need a cathode-follower or emitter-follower stage.
1 x
Nature abhors a clean tube amp

User avatar
crgfrench
Occasional poster
Occasional poster
Posts: 494
Joined: Fri 04/27/18 3:02 am

Re:

Post by crgfrench »

zaphod_phil wrote:
Sat 09/26/09 7:35 am
BTW you can also model various tone stacks, and how they behave when driven from different impedances, if you download Duncan's free Tone Stack Calculator from duncanamps.com
That is a great calculator. If you don't want to download software there is also a web version of this Tone Stack Calculator, at http://www.guitarscience.net/tsc/marshall.htm . Bonus: works on Macs also.
1 x

User avatar
colossal
Moderator
Moderator
Posts: 337
Joined: Thu 05/10/07 2:00 am
Location: In and around the lake

Re: Re:

Post by colossal »

zaphod_phil wrote:I don't actually see any need to buffer the SuperLite's tone stack, as the amp already has tons of gain, and the insertion loss of the tone stack is used as part of the design, so that the PI doesn't get driven harder than it should.

Also while a MOSFET follower doesn't colour the amp's sound, a true DC-coupled cathode follower actually does colour the sound by soft-clipping the waveform - see the relevant section in www.coastin.com/AESPaper.pdf In the SuperLite design, I warm-bias the 2nd preamp stage to give the same kind of effect.
I am reattaching the paper cited in this reply, as it was lost a long time ago.

AESPaper.pdf
You do not have the required permissions to view the files attached to this post.
1 x

Post Reply