Can't get tube cricket to bias

[MikeH]

I built a tube cricket- the schematic says to bias the annodes to 1/2 the V+(mine is around 11.5v), but I can only get them down to around 10V.  I checked and I am using the right value trimpots; do I need to try a different value for the trims?  Or do I need to maybe adjust the 4k7 cathode resistors?

http://www.beavisaudio.com/Projects/TubeCricket/TubeCricket_Final_2008.gif

[earthtonesaudio]

You could adjust the 4k7 resistors, but as you're doing it, note whether 1/2 supply on the anodes is the sound you like.  Sometimes biasing them away from the midpoint sounds better, sometimes not.  The main reason to use 1/2 supply biasing is to get more signal before distortion.

[PRR]

Are the tubes lit-up?

[MikeH]

Yeah, the heaters are running fine.  Bias is definitely not where I want it- with the trims turned to maximum resistance I get the most output, but the output is still lower than it should be. I measured the trimmers, and they are around 100K (95 or so)- should I try a bigger resistance?

[MikeH]

I suppose I could also have a problem with my layout; I've been over it several times but here it is.  Maybe someone can see something I missed.

*note- I realize the pin 5/R5 junction is not connected to ground in this picture- I caught that and fixed it before I etched.  Besides, my heaters wouldn't be lit if I hadn't.  Also I'm using a switch instead of the gain control, obviously.



Thanks!

[earthtonesaudio]

I would experiment with lowering the cathode resistors.  You could clip-lead a pot in parallel with the existing resistors, dial it down to where it sounds good, and then calculate the parallel value and swap in a fixed resistor.

[PRR]

I don't see where heater pin 5 connects to ground.

I don't understand switch S1.

Assuming the tube is lit and the values are correct: it should pull up to 11.5V and down to 3V. Yet yours doesn't. So something IS wrong, and we can't see it from here.

> I am using the right value trimpots

100K? With power off, put your ohm meter across the trimpot and make sure it adjusts to at least 80K. Not that I doubt your work (though I often doubt mine), just to be sure the pot maker didn't make/mark it wrong.

Power on: What is the voltage at the Grid (1Meg) and cathode (4.7K)? Grid better be "zero" (less than a tenth volt). Cathode should be several tenths volt.

I don't have a 12AU7 in the house. But I asked my idiot assistant and got these numbers. First with 100K up top and a dead-short on the 4.7K, the plate should swing way down, 2V-3V. Stock un-trimmed, around 4V. Trimmed to around 50K, 5.7V. Personally I would fix the 100K and trim the cathode R, that wants 10K+ to get 5.7V. I don't trust these results too closely: they are derived from measurements at 50V-300V which are generally valid, but as we get into tenths-Volt bias they will be a little off. Still shows the trend.



If the plate is "stuck" up above 10V, even with cathode R shorted, the tube is not pulling current. Re-check your heater. Apologies if you know tubes, but I've learned that not everybody grew up with tubes. You must have orange glow at top and bottom of the internal metal bits, though it can be hard to see because of the glass silvering. You should have at least 11V from pin 4 to pin 5.

Missing grounds is my favorite mistake. Done that so many times that I've lost count.

[MikeH]

Thanks guys

I don't see where heater pin 5 connects to ground.  I don't understand switch S1.

I mentioned above- that was an error on my layout and I fixed it before etching.  And the switch was wrong too I had to fix that, but it was just there for experimenting/etc.

Tube fillaments are definitely lit.  There is a definite glow, and the tube is producing heat.  Unless there's something I don't know about heaters, they look lit to me.  Can a heater be "half lit"?  If so I would think it wouldn't work at all.

I measured the trimmers, they were 95K or so.

Voltage on the first grid (pin 2) is -.46V and on the second grid (pin 7) it's -.268V. 
Voltages on the cathodes (pins 3 and are .039V and .056V respectively.

Uhf.  I have a couple more 12au7s; I'll try another to make sure that it's not a bum tube.  If that's not it (and it's not likely to be) I'll try adjusting the cathode resistors, but man- there must be something else wrong.  I've had similar problems with FETs when I was using a 10K trimmer instead of 100K, for example, and that was my first thought here too.  I could pull those too and double check I suppose.

And I'll double check that everything is grounded too.

[MikeH]

Interesting- I tried a different tube and the bias voltages are down to 9V now.  Maybe I'm looking at a "several little things adding up to one big thing" sort of a problem.

[PRR]

Those grid voltages are awful negative. Working with more "normal" supply voltages, they would be "bad". However I have not worked much with tubes at 11V, and not in a long time. Maybe such low plate voltage does not collect enough electrons? The "average" electron energy is zero, but it's a Gaussian bell-curve, and some electrons escape with enough energy to fly right into the grid wires. Normally the plate-suck pulls most electrons between the wires, but 11V isn't much suck.

> Can a heater be "half lit"?

There's a range of heater voltages where "the cathode works". And on higher supply voltages and working at low current, it is acceptable to under-heat. High current with not-hot cathode strips the cathode, but this is a 20mA tube working at 0.1mA so we are nowhere near "damage". But heater voltage also shifts the bias curves; heater voltage variation is the worst cause of drift in DC amplifiers. "Zero" the amp at 6.3V heater, wobble it to 6.0V and 6.6V, you get several tenths Volt error.

BTW: the plan shown gets heater power after the polarity protection diode. The heater does not care about polarity. If your raw supply is less than 13V, you could move the pin 4 heater jumper over to SW1, before the protection diode, gain another 0.6V. Normally that would be negligible, but if you have 11V or less at heater, you should probably try it.

Have the tubes been idle a while? If so, can you rig a solid 12.3V-12.9V supply to hot-up the heater and leave it cook overnight or a week? Stir-up the cathode barium, hold the getter at active temp to absorb stray gas? When trying to set an exact bias on a new tube in my VTVM, I have to "cook" the tube for a week before it settles down. I have seen other old/idle tubes "clean up" after some hours on simmer.

> try adjusting the cathode resistors

The obvious thing is to jumper the 4.7K, jam the tube hard ON, and see if plate voltage falls. And incidentally where the grid voltage gets too.

> Maybe I'm looking at a "several little things adding up to one big thing" sort of a problem.

That's my impression too. Little tube geometry differences, a little underheating.

[Johan]

there is a write up on the PAIA website describing how and why tube grids go negative when the tube runs on low/starved voltages. ..try find it...PRR's idea about electrons escaping and being caught by the grids is, if I remember correctly, in line with the PAIA explanation..
also, the patent for the original BK Butler Tubedriver ( later bought by Chandler) has a note on it too, and gives it as the reason the grid resistor on the tubedriver goes to +9volt instead of ground. ...the tubes do work at very low voltages, just not exactly the same way as we're used to see them and not always in a repeatable manner. carefull tube selection may be needed
J

[MikeH]

Thanks guys - I'll look for that article.  I'll also try shorting the 4k7 resistors and see what I come up with.   I didn't measure them before I put them in but who knows, maybe what looks like a red band is really actually orange.  Maybe the yellow was running low that day.  Doubtful, although I did get a bulk pack of resistors once, and I don't know if it's the lighting on my bench, but blue and green looked exactly the same.  I couldn't tell the 1Ms and the 10Ms apart.  I had to meter them.

[PRR]

> maybe what looks like a red band is really actually orange.  ... don't know if it's the lighting on my bench, but blue and green looked exactly the same.

Yes, popular problems.

Seems like old A-B used zingier ink than we find on modern parts.

You need a WHITE light near your bench. Fluorescent, tube or screw-in, is NOT white, it is 2 or 3 narrow colors in white-like proportion. Sunlight and incandescent are still the best handy color-check light sources.

Color-blindness is also popular. Total color blindness is rare, but a lot of guys have trouble with a couple of colors. Have you had a color-blind test? They often do that for a Drivers License: a lot of dots with a number hidden in them. However you might "pass" for driving yet have trouble with the subtle shades on modern resistors.

[MikeH]

Ha ha, no I'm not color blind.