Valve testers & 'soak' testing.
 

It is possible to use a commercially-designed valve tester to 'soak' test a valve at its full rating for an extended period of time - say, 30 minutes or one hour?
And is such a facility available with these type of valve testers for full-wave rectifiers?

Al. / Sept. 11th., 2021

Re: Valve testers & 'soak' testing.
 

Hi Al, from the units I have seen I would say no, the transformers look as if they are designed for about a 30% duty cycle max.
You will be able to soak test your EF80 for several hours, but no chance with rectifiers or power valves.

It is simple enough to make a large PSU and a couple of meters to carry out soak testing if you wish, you can then put the valve back in your nice precision tester to check on progress

Ed

Re: Valve testers & 'soak' testing.
 

There's an emission test on the Mullard High Speed valve tester that I 'suppose' you could leave on for 30 minutes. If any damage was pending to the valve tester there's a cutout that operates.

But all valve testers are different.

Re: Valve testers & 'soak' testing.
 

Been raised several times in the past - see"Search". Definitive answer - NO.

Ed's advice is spot-on. 'Specially as just about all the known makes are approaching OAP age.

Regards, David.

Re: Valve testers & 'soak' testing.
 

Thanks: so the answer is 'No' - which was what I anticipated; just wanted confirmation.
So, a large PSU as a Test Jig for FW rectifiers is called for. Ho-hum, another project to build :-). Plus something similar for watt-ful AF output valves, such as the EL34, EL37, KT88, etc.

Al. / Sept. 11, 2021

Re: Valve testers & 'soak' testing.
 

I agree with all comments saying NO, all models I've seen and tested, even modern designs simply have too small transformers to perform any proper soak testing of high power valves.

Some modern design look very fine on paper but during heavy load they simply don't have the regulation needed built into their designs so the voltages sag when the testers get hot and rise when they are cooled down again, which happens over the course of a soak test as the room temperature as well as the cooling fans affect the thermal stability of the designs.

Nowadays I use high & low voltage power supplies from Delta Elektronika with DC-only options for all voltages and also my old KSM power supplies which have most voltages available albeit only with a 6.3V heater but they handle 500V@0.5A & -200V@0.1A and also -200V@0.01A, they are shown in the included photo, but also a setup with school laboratory power supplies from IMPO, one high voltage model and one low voltage model for AC/DC heater.

Re: Valve testers & 'soak' testing.
 

Ha! I could do with a machine to 'soak test' myself, since I am actually well into the 'OAP category'. ;D And I'm not thinking about using the bath, either! ;D

Al. / Sept. 11th.

Re: Valve testers & 'soak' testing.
 

Once a valve tester has done enough to determine whether a valve is working to spec or not, it's done its job... all of its job. There is then a delay while the next valve is fitted and the pinout, heater and electrode voltages are dialled-in, before the tester is run again. Some tests like shorts and gas are quite low power. So 30% of the time being spent doing full current test is a fair estimate for a machine working commercially, testing valve after valve.

To back up the theory with a proper practical experiment, look back in this forum. It's a few years ago but there was a chap who bought an AVO VCM and decided to use it to burn-in valves for his guitar amplifier. So was the valve testing the tester, or was the tester testing the valve? It cost him a very expensive tester. I don't know whether his guitar gently wept, but the guitarist certainly did.

So that's a no from both theory and practice departments.

David

Re: Valve testers & 'soak' testing.
 

This is what you are looking for I think. I have knocked up a birdnest version and it works fine.

https://sound-au.com/project165.htm

I will be building it into a case "soon" ( its project 30 ).

Joe

Re: Valve testers & 'soak' testing.
 

Hi Folks, this type of tester could also be adapted to give a cathode peak emission test

With a fully warmed up valve the GB volts can be reduced for a few mS and the Ia monitored on a scope. This is a comparative test but "may" act as an indicator of remaining valve life ~ discuss.

Ed

Re: Valve testers & 'soak' testing.
 

The Mullard tester is a DC tester and is capable of testing many valves under normal operating conditions limited by the 250V stabilised supply. However, the test conditions can exceed the valve dissipation ratings if the valve has good/excellent emissions so more like 'burning out' rather than 'in'.

Re: Valve testers & 'soak' testing.
 

Al, the datasheet of a rectifier often has a few 'full ratings', so you may need to select particular ratings that you want to test with. For example applying the correct heater voltage seems straight forward. Setting up a testing condition that causes each diode in the valve to alternately conduct the max rated continuous peak conduction current imho is probably the other key condition to apply.

One option is not to use a PSU with the max rated winding voltage, but rather use a lowish voltage winding that effectively provides the same peak current waveform per anode, but doesn't end up consuming the same power. The valve is not then subject to alternating PIV conditions, but is subject to the same internal power dissipation (and hence normal operating temperatures). Using a 5U4 example, the practical difference could be a DC load that dissipates 470Vx225mA=106W to meet the datasheet 'full rating', versus using a 75-0-75V secondary and just shorting the DC output (no need for filter cap), where both setups force a peak 800mA current pulse through each anode.

Another option is to use 6.3 or 5V AC or DC (whatever is convenient) for the heater, and use a bench DC supply to continuously conduct sufficient current through each plate such that the DC dissipation in each plate is equivalent to the dissipation during normal AC operation. For example a 5U4 anode conducts about 270mArms when it has an 800mA peak, so something like a 50Vdc supply would cause that level of dc current through a 5U4 anode, and therefore the bench supply would need to provide about 540mAdc total.

And then there is the issue of what you consider is a pass/fail condition. Do you only test at the end of the soak, and what do you test, or do you try and test certain parameters during the test, or do you just see if a fuse pops?

And as an aside, the vintage simple emission tester I have tests a valve like a 5U4 using a 32Vac supply through about 280 ohm of limiting resistance, so given the voltage drop across a 5U4 diode that means the test current in a single anode is quite low. And also worth noting that such a 'Merit' test requires pushbutton activation, and forces a DC offset through the tester transformer as the test circuit is effectively a half-wave rectifier.

Re: Valve testers & 'soak' testing.
 

Tektronix used to soak their valves in massive racks designed for the purpose. This document describes the racks and schematics for particular valves, and the aging cycles https://w140.com/tekwiki/images/0/0a/Tek_tube_aging.pdf

Craig

Re: Valve testers & 'soak' testing.
 

That's for new valves though Craig, like for selecting triodes and pentodes with certain parameters, Al is probably mostly testing used pulls.

As regards "soak testing" most iffy valves I've tested show up as iffy straight away within 20 seconds or so. Soak testing AFAIK means extended powering up of a device in order to show up any faults missed in short period bench testing.

For rectifier testing surely a big tfmr powered off a variac with a load should show up any duds. How do rectifier valves tend to fault? Probably just low emmision like triodes, so a 50mA load should show up any tired valves straight away.

Andy.

Re: Valve testers & 'soak' testing.
 

In the case of what Tek and others did, it's a process to weed out infant mortalities. New valves with a defect which limits them to a short life, but which doesn't spoil the parameters in the beginning, say like a poor seal on a pin, or one not outgassed properly.

Much of what they did wasn't selecting valves for better parametric performance, just improving the confidence they could have in their longevity.

David

Re: Valve testers & 'soak' testing.
 

From a large batch of 5u4 'pulls' that I tested, 'faults' included leakage at PIV, and emission level and imbalance at an on-voltage.

PIV leakage testing doesn't need an energised rig or a soak as an insulation resistance tester shows it up (even at 1kVdc for a 5U4, which is getting poor), and leakage can occur on either anode so best to test both anodes to cathode.

A minimum emission current at a given voltage drop is an acceptane and life parameter for the only rectifier datasheet with info that I've come across (6X4W and WA), but imbalance isn't a defined parameter. The 5u4 batch showed that emission current spread gets larger as the on-voltage is increased, but even with 24Vdc applied the valves could be sorted for weak and better (once you set a pass level). Imbalance in on-voltage may be a concern, and a pass/fail level probably needs to be aligned with application performance, as it may not matter.

Re: Valve testers & 'soak' testing.
 

They also had a matching and selection process following aging in the large racks. The aging process of course weeded out infant mortality, but it was more to do with ensuring that the characteristics had stabilized.

https://w140.com/tekwiki/images/1/14..._procedure.pdf

Craig

Re: Valve testers & 'soak' testing.
 

That's interesting Tim, hadn't heard of imbalance in dual diode rectifiers,makes sense though,it's the same with dual triodes.

Yep, Tek 500 scopes have the odd 6DJ8 and other small signal triodes & pentodes with Tek part numbers on them indicating certain characteristics, makes sourcing replacement valves fun.

Re: Valve testers & 'soak' testing.
 

This is interesting, thanks. Can we take it a little further please.

Leakage at PIV, valve hot or cold? If DC then is it polarity specific? Leakage similar between anodes? Why should it leak? Glass is a good insulator, does it leak between any two pins, connected or not? What, basically, is the failure mode.

Emission imbalance, normally the cathode has more than enough coating to supply the number of electrons, so would this be an indication that the valve is near end of life?

With a rectifier one assumes that each half is used, worn out, equally which is not the case with a DT. Back in the 70's our TV failed to sync lock, reading as to why it was the sync separator failing due to low anode current, fix was to swap it with another similar valve.

Re: Valve testers & 'soak' testing.
 

HI,
the Tek Tech article is very useful, thank you.
I think that designing a unit to Soak test valves safely, is actually more complex than it would seem.