Is this normal?

[daz]

I never really noticed this before, tho maybe it's always done it but i never noticed. I have a dual bias setup and i usually bias el34's to about 35ma, and as i watch the meter the value keeps fluctuating even after the amp is well warmed up. This is with no input by the way, or a guitar with the volume all the way down. Usually only varies by about a millivolt, maybe even less, but it never stops. I feel as tho it didn't do this in the past but i may be wrong and why i just happened to worry about it now I don't know. It does it with any tubes. Could it be a cap, or is this typical? If it IS a cap, any way to figure out which of the 4 might be causing it?

[loudthud]

Is the line voltage, B+ or bias fluctuating?

[Enzo]

And your mains AC voltage? I bet it is moving around too. That happens all the time, and everything in the amp will follow along with it.

[daz]

Is the line voltage, B+ or bias fluctuating?

Doh ! Never thought to check that. Just did, and the mains fluctuated alright. About 1/10v about ever 10 seconds, then it went up about 2 volts and started fluctuating a few 10ths. At times it stabilizes for as much as 20 seconds, but then since the meter only registers minimum .1v and we're measuring millivolts it was probably fluctuating in smaller increments that would affect the bias reading even when the mains looked stable. Thanks, guess thats it. I was only concerned because i don't recall it always doing that, but maybe our mains has been more stable in the past and thats why i just noticed it.

[R.G.]

What size is the signal you're measuring? 10mV? 100mV? What are you using to measure it? On what scale? And what are the resolution, accuracy, precision and repeatability of the measurement device?

1mV is hard to read on a scale that's 10V max. It's a piece of cake on a range that's 10mV full scale.

If it's a digital voltmeter, and you check the specs, chances are that it's got "yada, yada, yada, x% of full scale +/-1 digit". Or two digits. Many DVMs have 200mV scales as the native internal range, and this is exposed for the lowest full scale range; on top of that, they often have 3 1/2 digits, the half digit being the "1" that comes on to separate 999 from 1000.

The simple process of digitizing requires something to make a decision that the measured quantity is above or below the cutoff point for the next/previous count. In practice, no digital instrument, even if perfect, can get closer than +/- 1/2 of one count. If you're seeing count-flipping in the last digit, it could well be the real, true, analog voltage wandering over the threshold and back. Real-world digital instruments also tend to new conversions on a regular basis. So a few times a second, it starts over. There is some wobbling in that.

In general, you need an instrument with a resolution that's 8 to ten times finer than the smallest interval you want to squint to be able to trust the lowest digit or binary reading. And on top of that, a precision and repeatability to match. Then you get to needing it to be accurate. It's possible for a super-precise, high resolution device to be grossly in accurate, like reading 5.345V when the actual voltage is 7.

And that's just in the meter.

There is another issue - 1/f noise, or flicker noise as it's sometimes called. 1/f noise gets worse as frequency drops. It's responsible for a good deal of "wander" in readings, getting worse at low frequencies. You could just be reading flicker noise.

The moral of this rant is that it's really, really hard to tell what you have if you have a change of one or two least significant digits on a DVM.

[hunfyuss]

I have never come to this before

[soundguruman]

I never really noticed this before, tho maybe it's always done it but i never noticed. I have a dual bias setup and i usually bias el34's to about 35ma, and as i watch the meter the value keeps fluctuating even after the amp is well warmed up. This is with no input by the way, or a guitar with the volume all the way down. Usually only varies by about a millivolt, maybe even less, but it never stops. I feel as tho it didn't do this in the past but i may be wrong and why i just happened to worry about it now I don't know. It does it with any tubes. Could it be a cap, or is this typical? If it IS a cap, any way to figure out which of the 4 might be causing it?

The line voltage from the power company is changing up and down all the time.
That's what you are seeing, more likely.
If a capacitor is going bad, which is possible, change them one at a time until you find it.
Capacitors can test good, and fail under power and heat, true. But not all the time and not very common.
It's more rarely than often, that a capacitor behaves like this.

[Tom Phillips]

Had a similar problem once. The signal pulsing could be seen on the scope. The amp did not have a tremolo circuit. Turned out that the amp was plugged into the same breaker branch circuit as a washing machine that was operating at the time. One this was figured out it was obvious that the amp's signal pulsing was exactly in sync with the agitation motion of the washing machine. This is just one example of the tings that can trip you up.