Question about small caps on preamp tube pins

[Steve A.]

Adding small caps (20pf-50pf) to tube pins to help with oscillations and other problems. I always wondered what the difference was if you added a 20pf cap between the grid or plate to ground, or from the plate to B+. I would guess that the effect would be more pronounced from the grid to ground because of the gain in the stage. So depending on the gain you might need a 120pf cap across the plate resistor to have the same effect as a 20pf from the grid to ground. Is that true?

It seems to me that running a small cap between the plate and the grid would have a different effect as it would be a small local feedback loop. The Dumble .047uf cap and 22M resistor between the plate and ground of the 2nd stage in the non-HRM ODS's it seems to compress the signal a bit. So how would all of that apply to a 20pf (or 120pf) cap from the plate to the grid in a preamp gain stage?

Thanks!

Steve Ahola

P.S. With chassis mounted tube sockets I would just tack-solder the caps and observe the results but this is for the clean channel of my Mesa Boogie Maverick amp that I am rewiring as an AB763 SR and the pcb is packed to capacity (with there being 3 or 4 resistors or caps inside each of the preamp tube sockets so as to not waste real estate.)

[soundguruman]

From the plate to B+ will cause a sharp high frequency roll-off in the frequency response.
Frequently, as in marshall and fender,a 7pf 1000 volt cap is added between the plate and the grid. This is as you said a feedback loop. It tends to help stop the V1 tube from ringing and squeeling.

plate to ground will tend to create a resonant point, you already have an electrolytic cap from the power supply to ground, or you should.
Grid to ground will tend to kill RF and high frequency noise on the input, as in PV designs. PV uses 39 pf.
No, they all have different effects, not the same.

In a boogie, all these caps are already installed...

[Steve A.]

In a boogie, all these caps are already installed...

Not any more- they are all on the cutting room floor!

I ripped them all out in rewiring the clean channel of my Mesa Maverick as an AB763 SR.

Thanks for your very informative post! It is a slight ringing that I want to get rid of so plate to grid sounds like the way to go. (Yikes- I forgot to ask about plate to cathode!)

Steve Ahola

[soundguruman]

Not any more- they are all on the cutting room floor!

I ripped them all out in rewiring the clean channel of my Mesa Maverick as an AB763 SR.

Thanks for your very informative post! It is a slight ringing that I want to get rid of so plate to grid sounds like the way to go. (Yikes- I forgot to ask about plate to cathode!)

Steve Ahola

You ripped them out, that's why it rings!
Yes plate to grid 7.5 pf 1000V (MUST be 1000V or higher) on V1A input stage. This is Fender Marshall method which I like the best. I would like to use silver Mica cap, but ceramic disk was used by Fender Marshall and it's OK.
Real trivia, as long as I am on the subject: Marshall used a piece of shielded wire, the shield was connected to the plate (plate voltage). The center conductor was the grid wire. And so (think about it) this forms a capacitor between plate and grid. Clever devils.
Plate to cathode- this was first introduced at the dawn of Mesa Boogie, when people complained about the screeching ringing of the V1 preamp tube. Actually it was used in amps that go back to 1930. Mesa did not invent that, but they will tell you they did.
It sound like mud to me, but Smith installed a 250 pf cap from plate to cathode. It did not work as well as plate to grid cap, just made the thing muddy.
Tube Dampers: These are very effective. There are expensive ones, and cheap ones made out of foam, made by Fender for Hot rod Deluxe. The high density fire proof foam will help stop the ringing, works pretty darn well.
Mahalo Nui Loa, dude. (Kailua Beach)

[Steve A.]

You ripped them out, that's why it rings!

I went ahead and put 33pF from plate to grid on V1a (BF AB763 channel) and 68pF on V2a (the Mav Lead channel). I won't have to worry about ringing in the New Year- Ho, ho, ho!

The stock Mav "clean" channel*** used a .005uF cap in series with a 330k resistor between the grid and the cathode of the second gain stage. And I seem to remember the ODS HRM schematic showing a 150-250pF cap from plate to cathode on the two stages in the OD section.

Thanks!

Steve

P.S. Ken Fischer reportedly used the "plate as shield" in some of his amps. And the Peavey Classic 30 layout shows an interesting interleaving pattern between grid and plate...



*** Muddy as hell in my book but I guess that was as clean as Mesa got in 1993! I like a little character (and growl!) in my amps- to hell with the violin-like sustain that Randall Smith craves. I'd rather sound like a saxophone than a damned violin!

[soundguruman]

Upon arrival of the first Mesa amp, I called randell and chewed him for a solid hour, and I was not polite.

[Chuck H]

Marshall started using plate to grid capacitance in it's designs early. How early? Not sure. But I had a '77 JMP with the V1a plate and grid leads spooged together against the chassis.

It's my understanding that the grid to plate cap will increase in effect (and decrease in frequency IIRC) relative to gain. As in, low input signal, less effect. High input signal, more effect. You can see where this makes good sense for a high gainer. Very small caps go a long way in this circuit. You can hear the effect as low as a couple of pf. I use a 4.7pf on V1a in one of my designs and this seems to keep the crisp highs but roll off hard before ice pick territory and squelch instability. I did it because in the prototype for this design there was a seven inch lead from the input jack to the first grid. I used shielded lead for that. Then when I changed the layout I had only an inch and a quarter between the jack and input grid. The lack of capacitance from the shield caused some instability. Rather than use a corse "bleeder" cap or trying to get shielded lead in that small run (which may or may not have worked since the short distance wouldn't provide much capacitance) I tried a 10pf cap from grid to plate. This worked but was too intrusive on the high end too early in the amp and harmonics got lost. So I went with a 6pf and that was good. Figuring I was on a roll reducing that cap size I tried a 4.7pf and that was better (for this amp). So that's what I used.

As I said. caps used in this circuit can be deceptively small. Anything over 6pf is a very audible reduction in treble. I typically run about 200Vp and use high quality, high temp 1kV ceramic caps. In some early design work I've done this with 500V caps with no problems.

[exclamationmark]

Connecting a capacitor between grid and plate increases the miller capacitance of the tube, and depending on the tube type the value of this capacitor will do different things. Basically the effective capacitance of the capacitor is multiplied by the gain of the stage (plus the stages initial miller capacitance). From here, you can view the miller capacitance as essentially a capacitor from grid to ground, and it forms a RC filter with the grid stopper.

Putting a cap across the plate resistor (or plate to ground/cathode), forms an RC filter with the output impedance of the stage. Putting a cap from grid to ground is probably my least favourite method of removing treble because it inexplicably sounds different (worse...) compared to the other methods. It's particularly noticeable if you have a somewhat dynamic pre-amp where you can go from sparking cleans to fairly saturated distortion based on how hard you hit the strings. The cleans aren't so sparkling, and are in fact a little dull, and the distortion still contains too much high end. The problem isn't so bad with the other bypassing methods. No idea why.

[dai h.]

as I understood, (for a typical plate follower stage as for the 1st stage in a 5F6-A Bassman as example), cap from plate to ground and across cap are the same. Plate to ground = plate to DC ground. Across plate R = plate to AC ground (grounded by local filter/decoupling cap). Plate to grid is out of phase, so small cap = phase cancels higher freqs. which increases effect as output increases (so useful for squealing that gets worse as vol. is increased).

As for grid to ground cap sounding worse, maybe something to do with the output impedance of the tube (one capable of more drive might sound better?).

[Steve A.]

I just put the Maverick back in its cabinet and then came back to my computer room to find 3 new posts on this thread. Darn! So I have the 33pf cap from plate to grid on V1a and a 68pf cap from plate to grid on V2a and will try that out for a few days. (The stock Lead channel has the 1M grid load for the initial gain stage bypassed with a .001uF cap which I removed so the 68pf cap is very insignificant compared to that.) I also replaced the 220k resistor to ground from the output of the reverb springs with a 470k resistor. I'd like more reverb on tap and I thought I would give that a shot. If it starts squealing like a stuck pig I think I know what to do now!

I guess I didn't mention in this thread that I just rewired the Rhythm channel of the Maverick as an AB763 Super Reverb, and it is a million times better than the stock channel which is basically the same as the Subway Blues- a decent enough amp but nothing to write home about. The Lead channel is considered by many to be a poor man's Dumble- that is really nice, especially with the few tweaks I did to it (like removing the .001uF cap to ground from the grid of V2a). But I found the Rhythm channel to be very boring- it sounded clean, it sounded good, but it was not at all exciting.

I was checking out the schematic for the Lone Star and noticed that it was practically identical to the AB763 Super Reverb- "Damn- that is what I wanted!" So I think that I now have the best of both worlds- the Blackface magic of the Lone Star Clean channel and the Dumblesque vibe of the Maverick Lead channel. Plus it only cost me a 1/3 of what the new Lone Star Special heads are selling for ($1800.) I'll be drawing up the mod so that other Maverick owners can join in on the fun!

Steve

[pdf64]

'The stock Lead channel has the 1M grid load for the initial gain stage bypassed with a .001uF cap which I removed so the 68pf cap is very insignificant compared to that'
From Aiken Designing Common-Cathode Amplifiers
for a stage gain of say 60, a 33pF grid to plate cap is roughly equivilant to a ~2000pF grid to ground cap.
Pete.

[jmaf]

'The stock Lead channel has the 1M grid load for the initial gain stage bypassed with a .001uF cap which I removed so the 68pf cap is very insignificant compared to that'
From Aiken Designing Common-Cathode Amplifiers
for a stage gain of say 60, a 33pF grid to plate cap is roughly equivilant to a ~2000pF grid to ground cap.
Pete.

I was talking about a filter with a forum member, when another member stepped in and asked if Miller capacitance had to be taken into account...I forgot and was totally messing up. Everything at the grid is multiplied, so picofarad caps shouldn't be ignored. a few pf won't do much but 2nF will really shape the tone.

[soundguruman]

Resistors contain smoke. If you hook up a resistor wrong, it will let the smoke out, and it won't work right anymore.

[Steve A.]

'The stock Lead channel has the 1M grid load for the initial gain stage bypassed with a .001uF cap which I removed so the 68pf cap is very insignificant compared to that'
From Aiken Designing Common-Cathode Amplifiers
for a stage gain of say 60, a 33pF grid to plate cap is roughly equivilant to a ~2000pF grid to ground cap.
Pete.

So much for my intuitive grasp of electronics! You are absolutely correct- the clean channel sounded a lot better with no cap on V1, more lively and responsive. With the 33pf and 68pf caps it was starting to sound like a Mesa Booger again! :-( I will try something much smaller than the 33pF cap there, like the 7.5pF value recommended by soundguruman for Fender/Marshall preamp (I don't have anything that small but two 15pf caps in series result in a capacitance of 7.5pf).

So having a 68pf cap from grid to plate on V2 would be like a ~4000pF grid to ground cap. Hell, that is much worse than the .001uF cap to ground that Mesa used.

Thanks!

Steve

EDIT I just put 10pf mica caps on V1a and V2a. Damn, they make both channels sound very nice. More thoughts on grid to plate caps: I believe that the equivalent grid capacitance to ground is more pronounced when there is a strong signal coming from your guitar. When you are playing softer it sounds like the effect is much less. So in that sense it seems to be very responsive to your playing. With a .001uF cap to ground on the grid it is always there, and is even more noticeable when you are playing soft (as it swallows up a lot of your high frequencies). In any case with the .001uf cap from grid to ground, the Lead channel was not very responsive to playing dynamics and backing off the guitar volume control did not clean up the signal nicely- IMO it was too muddy to be usable.

[Steve A.]

Resistors contain smoke. If you hook up a resistor wrong, it will let the smoke out, and it won't work right anymore.

Electrolytic caps are even worse!