Testing valve rectifiers

[ppppenguin]

In the absence of a valve tester how do you test a rectifier such as EZ80 or GZ34? Obviously if it works in a set then it's probably OK but I'd like to get some idea of how to measure it. I have LT and HT PSUs, plus several meters. I've tested triodes against the data sheet like this and could, if I was feeling patient, plot a set of curves.

Obviously I can check the heater, HK insulation and reverse breakdown. Though I wouldn't push that to find out how many kV it would take to make a GZ34 flash over  

That leaves emission, and balance between the two halves. I can conenct a low voltage PSU, wind it up and find the maximum current the cathode will give. Not sure that's very good for the valve.

Any thoughts?

[Murphyv310]

I'd reference against a NOS valve if you have one and pull 75% of the maximum current at a somewhat lower voltage, than the design centre maximum. Then see what the resultant output is compared with the NOS valve, perhaps not very scientific but will give an idea if the valve is ok for further use. If it's low emission I doubt it would supply anywhere near the rated current.

[Mike Watterson]

Not the maximum possible, but note voltage drop (i.e. applied HT) at rated current. If you can't reach rated current or the voltage drop is much higher than desired, then the emission is low.

A flash-gun based or similar capacitor leakage tester that uses a flashing neon (dark = OK, slow flash rate is G Ohms) for AK reverse when heater running and also H-K isolation. These usually have 1M to 2M series resistance. Some old capacitor meters had a leakage test built in.

The HK test and the PIV should be done with rated heater voltage or current and valve warmed up. Some only fail on HK leakage when hot. Not sure of effect on PIV. If vacuum is not so hard or other contamination or physical shock damage are reasons for low PIV.

I've used a 20M Ohm input DMM on volts to non-destructively measure Ge and Si diode PIV as well as Vce and Vcb breakdown (avalanche effect). You can use a select on test 1N4148 or other diode as a low current Avalanche regulator. In contrast the breakdown of K A reverse of tube is flash over and H-K is leakage. A sufficient series resistance as used on semiconductor PIV non-destructive tests is non-destructive. Most valve rectifiers ought to never light the neon even infrequently even with a 400V supply via 2M Ohm. I replace the typical flash 300uF with a 1uF 600V and 2 x 1M in series to + terminal for DUT. The return - terminal of DUT is a neon with a 0.1uF polyester 250V in parallel. The 2M and 0.1uF is a fast enough rate with a short to look solidly on.

[Murphyv310]

Hi.
I never said to use full current or maximum voltage.
I doubt you can get any where near an accurate result with too low a voltage or current, if only passing a few milliamps the voltage drop will be low and vice versa if you try to pull more current with low anode volts both current will be low and voltage drop high. One reason I stated to have a NOS valve as reference and have current around 75% of max and likewise with applied voltage a little less than the design centre maximum. I'd al as be inclined to use a DC supply.

[Mike Watterson]

I was referring to Jeffery, oddly I didn't see your post, Trevor. I'd not use a NOS but go by the datasheet ratings, and/or look at later AVO test settings. There was a lot of variation and a NOS part could be 60 years old. You'd not know how good it was then, nor how much it's deteriorated.

[Murphyv310]

I was referring to Jeffery, oddly I didn't see your post, Trevor.  I'd not use a NOS but go by the datasheet ratings, and/or look at later AVO test settings. There was a lot of variation and a NOS part could be 60 years old. You'd not know how good it was then, nor how much it's deteriorated.

Hi Michael.
The reason I'd use NOS is as you rightly state even a NOS valve will have deteriorated, let alone a used valve or pull, we need a benchmark to base our tests on and to find a good old valve will be hard, it's less likely that we'll find any old valve at 100% emission, I've bought many alleged 100% valves over the years and none are at that figure.

[Diabolical Artificer]

If you look at the Avo valve tester manual all it specs for testing most, if not all rectifiers is an anode current with the anode volts specification given as "REC", which I presume means recommended, what ever that means. Most valve rectifier datasheets state 4 parameters in simple terms - total supply R per plate (anode), max C or L, V RMS & Ia @ filter IP. With that in mind a good test setup would be a variac feeding a mains tfmr, a cap and a load or a DC supply into the rectifier + cap under load. If V RMS or Ia is off it's suspect or dud.

When testing valves with a simple setup you get a feel for a good valve,if your voltmeter or ammeter reading starts mucking about somethings not right. Give the valve a wiggle in the valve base, still not right, it's iffy.

Andy.