Monkeymatic Penta (Matchless Clubman inspired)

[martin manning]

I agree with you, Mark. I have had very good results with JJ tubes too. This particular one, however, was just not functioning properly as-received. The best thing to do is test and reject (send back) any that are defective. This puts pressure on the manufacturers to improve.

[xtian]

I think JJ are the Toyota Camry of tubes. Reliable, predictable, good quality, inexpensive but not cheap.

[xtian]

Amp is back. Owner is very happy, but hoping to squeeze out more clean headroom.

Using my scope, with 1v p-p 1k sine wave at input, I see an incredible 250v p-p totally clean output from the EF86 second stage. Amazing. This goes into the volume pot, then to the PI. So I want to make sure my PI and power section are doing all they can with the clean input source.

1. The input cap for the PI, C11 on the schematic, is 0.1u. More often in LTP, I see a value one tenth that of the mirroring cap, C15. Good place to tweak?

2. The power tube's 220K/220K grid ref (R16/25 on the schematic). I think I want to hit the power tubes harder, so change to 470K/470K?

3. Bias: The schematic says 270R cathode resistor for each power tube, with 420 on plates and 380 screens, and shows 25v on the cathodes. I calculate this as 160% plate dissipation!!! What? On my amp, I have 406v plates, 397v screens, 360R cathode resistors, 25.2v on cathode, giving me about 120%. Still seems too hot.

[ChrisM]

1. Yes it is oversized. To calculate the 3dB rolloff frequency use fc = 1/(2*pi*R*C) where R is the output impedance of the PI + the input impedance of the output tubes (220K) in your case.

A 10nF-22nF would probably make more sense. Same can be said to the other 100nF cap used for decoupling. Too big a value there can lead to block distortion.

2. You could do this. This will increase the distortion though, though you will be increasing it in the output stage. Look at EL34 datasheets, it states a maximum grid leak value there. IRRC it is around 500K.

3. This amp is notorious for running the tubes really hot and killing them quick. Use good quality tubes and bias them a little cooler. I think some of this amp's tone can be contributed to the hot bias but I would set them a little cooler.

[Darkbluemurder]

All good points.

Those 270R cathode resistors on the power tubes are definitely too small. I use 390R on each power tube in my Chieftain build with 410V on the plates and 27V on the cathodes. That puts them close to but under 100% plate dissipation.

If you want to increase the headroom of the PI, lower the 47k tail resistor to about 15k.

Cheers Stephan

[xtian]

All good points.

Those 270R cathode resistors on the power tubes are definitely too small. I use 390R on each power tube in my Chieftain build with 410V on the plates and 27V on the cathodes. That puts them close to but under 100% plate dissipation.

If you want to increase the headroom of the PI, lower the 47k tail resistor to about 15k.

Cheers Stephan

Super, thanks! I'll try the PI tail mod.

I installed 430R cat resistors, which gives me 100% diss.

[xtian]

Hmm. About the PI tail resistor, Aiken says,

This will give the best balance to the PI outputs. This resistor has little effect on gain, but a major effect on balance and headroom.

http://www.aikenamps.com/index.php/the-long-tail-pair

[matt h]

(deleted)

[xtian]

Many thanks, Matt. In addition to reducing the tail resistor, is there more that can be done to increase voltage swing out of the PI?

[matt h]

(deleted)

[Darkbluemurder]

But the PI values in that amp are lifted straight from a Vox AC30. to put it in relative terms, a power amp plate supply of 300V with an 8v bias voltage. That's a third (or a fourth) of what you're contending with in this output stage.

Exactly. That's why I suggested to lower the PI tail resistor. With the Vox values the EL34s may not be driven into heavy overdrive. Instead you will get most of the overdrive from the PI.

Cheers Stephan

[xtian]

Sorry, forgot to follow up. I ended up making the PI match Bassman values and using a 12AT7. I also changed the power tube grid leaks to 470K/470K, and the input to the PI to 0.01uF. Finally, upped the cathode bias resistors to 430R each, making the dissapation about 105%.

The changes were subtle, but pleased the owner greatly. A bit less compression, more headroom, and better clarity. And now I know the power tubes will last a lot longer. I measured about 25 watts clean, 35 cranked.

[SilverFox]

I'm considering building only the preamp portion of this amp for use with a MOSFET power amp.

The other possibility would be to use a Fender 6L6 output which sounds like what you ended up building.

Will that radically change the sound since much of the character, as discussed, comes from pi distortion?

Would this be a good preamp design?

I didn't have time tonight to go through all 7 pages of the build but will do that tomorrow.

Thanks,

Silverfox

[xtian]

@SilverFox: I'm not sold on this design. My original goal was to experiment with EF86 pentode in the preamp (but not in V1) to learn what color it could give to the circuit. Instead, as I found with this amp, it does not distort at all, but pushes the PI into clipping very easily. It's a good sounding amp, for sure, but not for the reasons I was after.

[matt h]

(deleted)