Gas test function in a valve tester

[Skywave]

Valve testers I've seen often have a button that you push to initiate a 'gas test' of the valve under test (VUT). A common method to do that is to switch in a 1 megohm resistor between the grid bias voltage source and the control grid of the VUT. But how does that work? And is a 1 megohm resistor a critical value?

Al.
June 5th.

[Radio Wrangler]

Not particularly critical in value, but very important in function.

You are metering the grid current, so you want your meter on its most sensitive range (this is one of the reasons why later valve testers had such sensitive meters) and the valve under test could have internal shorts, or just a bad level of gas, so the resistor is needed to limit the fault current and protect the meter.

David

[ms660]

Originally Posted by Radio Wrangler

You are metering the grid current

In some testers yes, in some testers no.

Lawrence.

[Skywave]

I'm currently trying to develop my own design of a basic valve tester: it will not be measuring grid current. So with that in mind, what is the appropriate arrangement to check for 'gas' in a valve under test?

Al.

[ms660]

Originally Posted by Skywave

I'm currently trying to develop my own design of a basic valve tester: it will not be measuring grid current. So with that in mind, what is the appropriate arrangement to check for 'gas' in a valve under test?

Something to indicate -ve grid grid current, either a meter in series with the grid to show the grid current directly, or a meter in series with the anode to monitor any change in anode current when doing a gas test using a switched series resistor in the grid circuit (resistor in series with the grid with a test switch (gas test) in parallel with the series resistor, the latter method was common, eg: Avo MK11, Taylor 45C, Hickok 1-177 for instance.

Lawrence.

Could merrily inserting a mega ohm or two in the grid lead to an indication of gas? A negatively biased valve is essentially an infinite impedance job, inserting the resistor would make the anode current vary a lot with gas present, no need for a super sensitive meter.

[GrimJosef]

The main problem I have with gas testers is that they can declare a power valve 'good' because the tester never gets the valve hot enough to outgas what's adsorbed onto its metalwork. Of course if they say the valve is 'bad' then it is. Very. But a good test is just the start really.

Cheers,

GJ

[PJL]

Originally Posted by GrimJosef

The main problem I have with gas testers is that they can declare a power valve 'good' because the tester never gets the valve hot enough to outgas what's adsorbed onto its metalwork. Of course if they say the valve is 'bad' then it is. Very. But a good test is just the start really.

Cheers,

GJ

The Mullard tester uses DC under 'typical' operating conditions and has a grid current test mode, shame about it's other shortcomings.

[GrimJosef]

I confess I'm not very familiar with the Mullard tester. By way of an example, do you know (or know how to find) the a-k voltage and the Ia at which the Mullard tester checks a KT66 ? If it can run the valve with 25W continuous anode dissipation for 5 minutes or so then that could be long enough for adsorbed gas to start to appear.

Cheers,

GJ

[PJL]

It only goes up to 250V and anode current will depend on condition.

Details in attached.

[GrimJosef]

That's a shame. 250V isn't really typical (335V in the Quad II, 360V in the RCA 32216A, 400V in the Leak TL/12 Point One, 415V in the Williamson amp ...). My experience has been that the gas current can be exquisitely sensitive to valve temperature - just opening the window in the workshop on a warm day can change the reading. So power dissipation is critical. But I get the feeling that for a given dissipated power, the problem can also be worse at higher voltage and lower current. Perhaps faster electrons have a greater ability to impact-ionise the gas ?

Cheers,

GJ