Monkeymatic Penta (Matchless Clubman inspired)

[Tillydog]

Higher plate voltage now: 158v, but still reading zero at pins 1,8. So weird.

...

I'm using 10K in place of the 8K2 dropping resistor. So, 9.2v dropped across this resistor gives me 0.92mA. Can that be right?

Also, 169v drop across the 220K plate resistor gives me 0.77mA.

That's stupid. It should be eating a few milliamps, right? Maybe I should take the next step toward the Vox design and change the cathode resistor to 1K5. What do you think?

The currents through the EF86 look credible, so no worries there.

The real puzzle is why you're not seeing a volt, or so between pins [EDIT] 3/8 [/EDIT]& ground.

Give the 1k5 a try ....I have played around with an EF86 as 2nd stage, and it is interesting to see/hear the effect of changes. The worst thing that happens is that you drive it into cut-off, because then it sounds like XXXt. To guard against this, you need to bias it on the warm side of centre and arrange the preceding stage to compress the signal on peaks. The screen voltage also needs to be pretty low IMHO. When you get it all right, it sounds superb. Don't be afraid to put a big grid stopper on the EF86 (even up to 560K) - it doesn't knock off the highs like it does with 12AX7s as the Miller effect is much less, but it can really smooth out the odd harsh transient if you push it hard.

Experiment!

[xtian]

Experiment!

Excellent advice--thoughout the build--t'dog. Thank you very much!

[martin manning]

Those EF86 voltages don't make any sense... There has to be something wrong there. Can you measure the voltage to ground on all of the pins and post them?

[M Fowler]

Does it matter if you use 2&7 rather then 3&8? I ground 2&7 and use 3&8 for grid on EF86.

Mark

[martin manning]

Pins 2 and 7 are connected to the shield, so one of these two should be grounded. Pin 8 is g3, which should be tied to the cathode, pin 3. Pin 1 is g2, and pin 8 is g1.

[M Fowler]

Layout shows two 68k grids tied to pins 2&7 of the EF86 is why I was wondering.

Mark

[xtian]

Layout shows two 68k grids tied to pins 2&7 of the EF86 is why I was wondering.

Mark

You're looking at my V1, a 12ax7, two halves paralleled as my first stage(s).

V2 is the EF86. Sorry I didn't label it!

The amp sounds superb. I have no idea why there is no voltage across the cathode resistor. I've got 150v at the plate, 36 at pin 1 grid. 2 and 7 are tied together and grounded (shied). 3 and 8 are tied together and connected to ground via parallel 2K7 and 25uF. 4 and 5 are filament. 9 is my input grid.

What can I say? There are plenty of pentode power sections where the cathode is directly grounded, and thus has 0 volts. I guess it's not that weird.

I have a signal generator and o-scope now! I could actually do some testing! I have no experience, tho. Like, apply a 1v, 1kHz sine wave at pin 9, and observe waveform at pin 6?

[martin manning]

But those power sections with grounded cathodes have a negative bias voltage applied to the grid, and your cathode isn't supposed to be grounded! You have an error somewhere. The screen voltage should be about twice what you have and the cathode should be ~1.7V.

[Firestorm]

The 0.1uF on the screen: leaking?

[M Fowler]

You're looking at my V1, a 12ax7, two halves paralleled as my first stage(s).

V2 is the EF86. Sorry I didn't label it!

Well that makes sense now

[Tillydog]

The real puzzle is why you're not seeing a volt, or so between pins 1/8 & ground.

D'oh that should read pins 3/8 & ground, sorry for any confusion - original post modified.

I've got .... 36[V] at pin 1 grid

Wow, that's low - when I was talking about keeping the screen grid low, I was thinking ~90V.

[Tillydog]

Daft question alert (and I apologise in advance for asking.....):

You are measuring voltage at the correct end of the 2k7 resistor, aren't you? Just that the resistor connections are 180 degrees out in the build from your layout because of where you've chosen to ground it.

Just sayin'.....

Like I said - I had to ask, sorry!

[xtian]

Yes, you had to ask. With one DMM probe clipped to circuit ground, I probed both sides of the cathode resistor. No voltage visible, to the three digit limit of my DMM: 0.000. And of course, as sanity check, everything else I touched gave me proper readings.

Had my buddy play it for an hour yesterday, with his Burny paul-style. Sounded great! Though I wouldn't mind more gain out of the EF86. So I still plan to try the 1K5 cat resistor. Interested to see how that will change the voltages. Will also likely move the cathode components off the tube socket, if only to help visually clear up the layout.

[martin manning]

If there is no current flowing through the Rk you have in there now then I can't see it making any difference. It's not functioning the way the schematic has it drawn.

[xtian]

What I see on the scope, that's peak-to-peak AC, right? To convert to RMS, take half the p-t-p and multiply by 0.707.

If I'm reading right, I see 250v p-t-p on my OT secondary. This gives me 88vac RMS. Am I on the right planet? I was expecting something around 30 volts. Will it be different because of the non-reactive dummy load?

Also, I'm seeing that the voltage, post PI, is LOWER than post V2:

input: 1v peak-to-peak
post V1: 3.8
post V2 (EF86): 136
post V3: 27.5, before clipping, not much higher after that
post EL34s: 250

The spec sheet for EF86 says 50vrms max output. Well, my 136p-p/2*0.707= 48vrms. OK, then.

Seems like I want to hit my PI far less hard. Decrease the value of the voltage divider after the volume pot?