plate current variation with fixed bias

[thehoj]

I am wondering about two slightly different situations with in my 845 / 211 amp which is fixed bias. The amp also runs a mosfet source follower which can source current to the grid when it swings positive.

Firstly the question about plate current variation.. With fixed bias is it normal for plate current to rise when approaching maximum swing on the grid? While running 845 tubes, I have my bias set such that idle current is 80mA. When I input a 2Vp-p 1Khz sine wave signal on the input, and increase that signal through the driver to nearly full drive, the current rises.. Up as high as 95 - 100mA when I'm at nearly full power (into positive grid region). I realize that music is not going to be a constant steady signal, but it does still seem odd to me.

My second concern related to this, is about one of the 845 tubes I have. It seems to act this way, but to an extreme amount. The plate current seems to want to rise much more aggressively with the amount of swing on the grid.. I can bias this tube even as low as 40mA, and much earlier than the other tube, the plate current starts to rise significantly.. It will rise up to 110mA as the swing on the grid goes up to nearly full power.

So on one tube, with idle current set at 80mA my plate current seems to vary between 80mA - 95mA, so a variation of 15mA.. On the other tube, the sky is the limit with current variation. This follows the tube. So if I swap tube locations, it doesn't seem to have anything to do with supporting circuitry for that particular channel.

So, does it seem like maybe I've got a bad tube?
And, with regard to the other tube, with the smaller variation, is that normal?

I'm new to the fixed bias world, I've only used cathode bias up to this point.

 

[Groucho2004]

How and where are you measuring the current through the valve?

 

[Merlinb]

With fixed bias is it normal for plate current to rise when approaching maximum swing on the grid?

Yes, it's called the self-rectification effect. Valves are non-linear -they tend to amplify positive input swings more than negative ones. The average over a whole cycle therefore increases slightly. The increase is even greater in class AB amps where positive signals are amplified but negative ones just push each valve into cutoff. You didn't say what your examples were.

 

[thehoj]

How and where are you measuring the current through the valve?

Measuring current by measuring voltage drop over a 10 ohm resistor from cathode to ground on each tube

 

[thehoj]

Yes, it's called the self-rectification effect. Valves are non-linear -they tend to amplify positive input swings more than negative ones. The average over a whole cycle therefore increases slightly. The increase is even greater in class AB amps where positive signals are amplified but negative ones just push each valve into cutoff. You didn't say what your examples were.

Sorry, should have given more info about amp. It's single ended, class A, capable of going into A2 region.

Thanks for the info, I was pretty sure the variation is normal. Though With the tube that is doing this variation to the extreme, is it likely that there's an issue with this tube?

 

[Merlinb]

Your 845 sounds abnormal, but perhaps it is a false reading, depending on what kind of meter you're using.

 

[thehoj]

Your 845 sounds abnormal, but perhaps it is a false reading, depending on what kind of meter you're using.

Using a fluke multimeter measuring the voltage drop over a 10 ohm resistor between cathode and ground. Same method of measuring bothe tubes, but different results for each tube, regardless of tube location.

I'm pretty sure it's something specific to the tube.. But I'm kind of curious what it would be with the tube causing it to act this way.

 

[kevinkr]

Measure the 845 grid current as you drive each tube into class A2, there should be a clue there.

What you describe behavior wise is about what I would expect for class A2 operation.

 

[thehoj]

What you describe behavior wise is about what I would expect for class A2 operation.

Even with the tube that is jumping from 40mA up to 110mA when crossing into A2?

The other tube doesn't vary nearly this much.. Unfortunately I don't have a 3rd tube to test at the moment.


Also, measuring grid current.. I wonder what the easiest way to do this would be?.. Meter set for current measuring, inline with mosfet source and tube grid?
Or a temporary small resistor between source and grid, measuring voltage drop over that resistor?

 

[jean michel]

hello ,

so possible short circuit to G1 up to Anode ... replace the triode or control the triode to lampmeter ( ? ) , other solution , look and control the bias résistor and capacitor to connection for Cathode , so possible deficious ...

sorry , not idéa for help

 

[Groucho2004]

Measuring current by measuring voltage drop over a 10 ohm resistor from cathode to ground on each tube

Several things you have to consider - cathode current != plate current with A2 drive. Also, you really need to use a scope to see the waveform, in particular to see the DC and AC components. A simple multimeter won't be of much help.

 

[thehoj]

Several things you have to consider - cathode current != plate current with A2 drive. Also, you really need to use a scope to see the waveform, in particular to see the DC and AC components. A simple multimeter won't be of much help.

Right.. So if I'm measuring cathode current, that's a combination of plate and grid current?.. That makes sense, if I'm measuring a combined current, then the amount that the cathode current increases by when going into a2 is the amount of grid current I'm drawing.

So that seems to indicate that one of the tubes is drawing a LOT of grid current..

 

[Groucho2004]

So that seems to indicate that one of the tubes is drawing a LOT of grid current..

Possibly. As I mentioned, without using a scope you're flying blind.

 

[DF96]

For the 'normal' valve the amount of current increase is a rough guide to the amount of second harmonic distortion, as both the DC shift and H2 are caused by the same thing (valve second-order nonlinearity).

 

[thehoj]

For the 'normal' valve the amount of current increase is a rough guide to the amount of second harmonic distortion, as both the DC shift and H2 are caused by the same thing (valve second-order nonlinearity).

Does it make sense that if I'm actually measuring current from cathode to ground, I'm not actually talking about purely plate current, but rather plate + grid current..? The current in reading doesn't actually seem to change until I get near / into positive grid region.. On both tubes, it's just one of the tubes is drawing more than the other.

 

[BinaryMike]

Does it make sense that if I'm actually measuring current from cathode to ground, I'm not actually talking about purely plate current, but rather plate + grid current?

Correct. And grid current angles up sharply as soon as you cross into positive grid territory -- much sharper than the anode current increase, which simply continues on its previous trajectory. Your grid driver is essentially working into a forward biased diode load when conduction begins.

 

[thehoj]

Correct. And grid current angles up sharply as soon as you cross into positive grid territory -- much sharper than the anode current increase, which simply continues on its previous trajectory. Your grid driver is essentially working into a forward biased diode load when conduction begins.

Alrighty.

I still feel like something isn't quite right with the one tube, as it seems to draw a lot more current than the other 845 tube.

I didn't mention something that I'm sure is related, because I was curious about this type of operation in general.

At one point I had an issue with the regulator for the filament on that tube, A part of the raw DC power supply that feeds the regulator was shorting to the chassis. This was when I was running the amp fairly differently, and was actually cathode biased. Without knowing exactly how it occured, I believe what ended up happening is that for a short period of time the tube ran with no bias basically.. It was maybe for 10 seconds or so.. No red plating.. And it seemed to work fine after that when the amp was cathode biased (no class A2 operation).

After switching to fixed bias and drawing grid current now, I'm noticing this issue, and I'm thinking it's gotta be an issue with that tube..
Could running the tube for a brief moment with the grid at the same potential as ground (in my previous cathode biased configuration) have diminished some aspect of this tube, causing it to draw excessive grid current?


As an interesting side point, when operating the amp with 211 power tubes, also biased around 80mA, even when pushing the amp right up to the point of clipping (48Vp-p over an 8 ohm load -- way into the positive grid region) the current increase over idle that I measure from cathode to ground is lower than even the good 845 tube.. Where with the good 845 I'm seeing maybe a 15mA current jump from cathode to ground at full power, with either of the 211 tubes I'm seeing less than 10mA current increase from cathode to ground, and this seems kind of odd to me now too.. I would expect the 211 to draw a similar amount of grid current.

 

[BinaryMike]

I'm no expert in this area, but I think the more subtle consequences of tube overcurrent are cathode/heater stripping and gas emission due to overheating. Gas can definitely cause excessive grid current. Perhaps another member can elucidate.

Be careful about inferring grid current by cathode metering. Grid current occurs in sharp pulses that defeat the 'True RMS' capability of all but the best multimeters. Judging by the kind of questions you're asking, I would say that you're ready for an oscilloscope.

 

[SpreadSpectrum]

Be careful about inferring grid current by cathode metering. Grid current occurs in sharp pulses that defeat the 'True RMS' capability of all but the best multimeters. Judging by the kind of questions you're asking, I would say that you're ready for an oscilloscope.

Mike, you said what I was going to. You would really have to look closely at a multimeter specs to see if it is going to be able to do a measurement like this with much accuracy. I would guess not unless I saw data to the contrary.

 

[euro21]

Try to measure at 100Hz. Average DVM is cheating at 1kHz.