A new twist on micro-phonic tubes

[nickb]

OK we've all come across microphonic tubes - tap 'em and they make a tinkling sound come of the speaker.

Now, here's a new twist -"speakerphonic" tubes. I'm struggling to find an appropriate antonym. These are tubes that make an audible noise when you apply a signal. Yup. You heard it first time and right here on MEF. Jensen -look out!!!!!

Here's what happened. I built a tube tester that does 18 at once. it applies the power by switching the cathode with a FET on for 80uS. It's all happily working away when I notice a faint tinkling sound and it's coming from the tubes. The bigger the voltage/current the louder it is. I haven't come across this in any of the literature. I haven't looked very hard either to be honest. I guess the electric field exerts a tiny force on the elements and they move slightly a bit like an electrostatic speaker. The precise converse of being microphonic. If I blow the dust of some old textbooks I should be able to estimate the forces.

I should point out that these were not what I would consider microphonic tubes as I had pre-screened them to avoid testing those that are.

I can't possibly be the first person to stumble onto this. Someone else out there will know more. But it does beg the question that this might be a more controllable way to test for microphonic tubes.

[Enzo]

Usually we wouldn't hear it over the other noises, including transformer singing.

So you are essentially hitting it with an square wave? Turning that 80us period into a sine wave, do you still hear it? My first thought was maybe impulse from heating as current turns on and off.

[The Dude]

I've heard it, but it usually subsides after the tubes warm up. Did you leave the tubes on your tester long enough for them to warm up a bit?

[nickb]

I've heard it, but it usually subsides after the tubes warm up. Did you leave the tubes on your tester long enough for them to warm up a bit?

I think I know what you are thing of - the expanding heating metal sound - right? This is nothing like that. It's 100% the electrical stimulus that doing it.

Here's a schematic if it helps the understanding of what I'm doing.

Schematic_BC.pdf

[The Dude]

Gotcha and thanks for clarifying. Yes, that is the sound I was thinking of.

[nickb]

Usually we wouldn't hear it over the other noises, including transformer singing.

So you are essentially hitting it with an square wave? Turning that 80us period into a sine wave, do you still hear it? My first thought was maybe impulse from heating as current turns on and off.

The heaters are well and truly warmed up so it's not them. As you say, I don't think you would hear this over the transformer if this were in an amp. I suspect that if it were possible to change this to a sine wave you would not hear anything.

[Enzo]

I understood you completely, I wasn;t thinking of warm up expansion sounds. Look inside an EL34 in an amp, you can see into the insides through holes in the plate assembly. You can see grids and the cathode surface. As the signal flows, changing currents can be seen modulating the brightness of the cathode sometimes. I was thinking maybe a tube is loose enough that that changing warmth might make it sing. Only some elements loose make a tube microphonic.

[nickb]

I understood you completely, I wasn;t thinking of warm up expansion sounds. Look inside an EL34 in an amp, you can see into the insides through holes in the plate assembly. You can see grids and the cathode surface. As the signal flows, changing currents can be seen modulating the brightness of the cathode sometimes. I was thinking maybe a tube is loose enough that that changing warmth might make it sing. Only some elements loose make a tube microphonic.

Ah! Stick? Wrong end? I've got it!

I see what you are saying. This is a very fast phenomenon, way outside any thermal time constant. I did try to mic it up to record it but failed due to lack of a decent mic. Just as well 'coz I can't sing either.

But that is interesting about cathodes getting hotter -that's a whole new topic

[R.G.]

Electrostatic attraction between tube elements ... might ... do this if the plate structure or cathode structure flexed a bit under the (small) electrostatic attraction. The same force that pulls electrons to the plate tries to pull the cathode, and space charge, anything with a negative charge on it, over to the plate, or vice versa.

Thing, is, this is small, and in a tube with a smallish signal on it, the plate-cathode voltage doesn't change all that much. Hitting it with full B+ in an 80uS pulse is probably much worse than any tube actually amplifying audio. Moving elements will change the gain of the tube as well as introducing a capacitive signal, and both are asymmetrical. For a sine-ish input, it's unlikely to be audible in music any more than the other slight distortions. Probably less than the voltage coefficient of carbon comp resistors, although we'd have to have more data to calculate it.

[nickb]

Electrostatic attraction between tube elements ... might ... do this if the plate structure or cathode structure flexed a bit under the (small) electrostatic attraction. The same force that pulls electrons to the plate tries to pull the cathode, and space charge, anything with a negative charge on it, over to the plate, or vice versa.

Thing, is, this is small, and in a tube with a smallish signal on it, the plate-cathode voltage doesn't change all that much. Hitting it with full B+ in an 80uS pulse is probably much worse than any tube actually amplifying audio. Moving elements will change the gain of the tube as well as introducing a capacitive signal, and both are asymmetrical. For a sine-ish input, it's unlikely to be audible in music any more than the other slight distortions. Probably less than the voltage coefficient of carbon comp resistors, although we'd have to have more data to calculate it.

To test the electrostatic force between elements idea, I simply turned off the heater supply and it was silent so that is ruled out. So, it does seem that space charge is very significant. It also reduces the effective distance between the charged 'surfaces' so increasing the force.

Then there is a ballistic effect. At 700mA plate current worst case something like 3 x 10^18 electrons/sec are smashing into the plate. With an applied voltage of 450 (pulse testing remember) the momentary power input is 315 watts. I suspect that this is the real mechanism. And yes, this not something that would be audible in any amplifier. It was very tongue-in-cheek when I mentioned speakers. That's my dry sense of humor for ya'.

I tried to record this and this is the best I could do with a crappy mic. It really sounds much more 'tinkly' than this in reality. Remember a rapid series of pulses is being applied.

tube noise.mp3

Finally, I don't think this is of any value is assessing how microphonic a tube is in general since different elements are involved. It will remain a fun observation of no practical value.

BTW: Happy Easter All!

[Mike Sulzer]

What does it sound like with just one tube under test?

[nickb]

What does it sound like with just one tube under test?

That was just one tube

To take one measurement it actually takes 3 sets of 16 x 80uS measurements, the three sets being around the point of interest. Each set is averaged and a plane fitted so that the wanted values at the POI can be interpolated. As a consequence there a many pulses for a single test and this was running a long series of tests.

[R.G.]

I always thought a good microphonics tester would mount a batch of tubes on something like a 1/4"/6mm plywood panel with a speaker under the panel. Play pseudorandom pink noise through the speaker and measure the output of the tubes for correlation to the pink noise. If it correlates, it's microphonic, as the internal noise will not correlate to the speaker output.