Valve test question

[Phil Y]

Hello all.
I have finally got round to trying my late fathers valve tester(AVO Mk3)and have come accross something strange that i hope that one of you clever chaps can explain. The first valve i tried was on old (1960s) EL84. I followed some notes written by my father and all seemed to be ok. Insulation test fine and current test low, just off the bottom of the "green" but it is a "from the back of the drawer" valve so no matter. Next i tried some newish EI brand EL84s that were in use untill the middle of last year in my Stereo 20, until i built my KEL84. I have 8 of these valves(2 stereo 20s). Every one of these valves failed both the insulation test and the current test. On the insulation test they all show short on the G1 & A2 positions and on the current test they all have the meter needle shooting off the right hand end of the scale. The warning buzzer went off on a couple of them. After testing these i was getting confused as these valves all work in an amp! Next i tried a brand new JJ brand EL84 i bought as a spare for my KEL84 and this was fine. Insulation test OK and current test in the middle of the green (good)sector on the meter. Has anyone any ideas what is going on with these EI valves? If they did not work in an amp it would be OK but they do! Could there be an internal connection on them that does not matter in an amp but is "non standard" Any suggestions welcome.

Regards, Phil.

[Richard]

HI Phil
I'm not familiar with your machine but can confirm that used EL84 often test as weak insulation, <5M but still test and work ok otherwise in my Avo 160. It is a real reading, new ones test fine, but the insulation seems to decline before they fail. If you use ones that read low insulation it will be an indicator of age so try and pair them for similar current and watch them carefully in a fully darkened room checking for plate redness to check they are not "running away".

[Phil Y]

Thanks for replying Richard, I have had another look at the valves. Firstly, when i said they failed the leakage test i did not make it clear that they read dead short in the G1 and A2 position. I've put a DMM on the valves and the EI ones have pins 1&2 connected together but from what info i can find there should be no connection to pin 1. I metered the JJ one and no other pin is connected to pin1. I presume the valve tester is testing leakage to all other pins so it sees a dead short from pin 2(grid) to pin 1 as a fail. I imagine this could also be causing the fail on current test. As pin 1 is not connected in an amp this is why it causes no problem there i presume. Looking into the envelope of the EIs you can clearly see a wire from pin 1 going up inside the functional part of the valve, it is next to the heater filament in the center. I think this must be the explanation, the only question is why are the EI valves made this way?

Only slightly less confused, Phil.

[NickG]

Looking at this

http://www.mif.pg.gda.pl/homepages/f...010/e/EL84.pdf

pins 1, 6, and 8 are internal connections. That means they can go anywhere (or not) and still be correct.

I would guess the vester is getting confused if its expecting a second anode (I assume thats what A2 is).

[Richard]

Hi

I'm not sure why it happens with some valves and not others, for any other reason than age/use, as EL84s have been made by many factories for many years so I'd have thought they'd be identical in use. The unused pin wires may be used as supports, "a disconnected valve pin or to a pin on which an internal electrode is anchored. Such pins are marked IC." (see Avo manual.)

You mention some notes, do you have the manual?
http://bama.edebris.com/manuals/avo/vcm-mk3/

manual download,
http://bama.edebris.com/download/avo...m%20mk-iii.pdf

On pages 2 and 9, it looks to use the same 9 roller pin selector as mine so perhaps start by checking yours is set to
041 230 605

It says of 0, "this switch position leaves the particular valve pin completely disconnected", corresponding to pins 1, 6, 8, so I don't think it will be checking anything to those pins.

[Phil Y]

Whoops, my fault.
Thanks for the help both of you. You were right Richard, somehow i managed to set the first roller to 7 rather than 0. The EI valves test OK now. I didn't realise unused pins could be connected to anything. I will remember that. It was just bad luck the error i made with the setting, both the old Mullard i tried and the new JJs have pin 1 used just as a support onto the mica so they tested OK with my incorrect setting. The EIs with pin 1 and pin 2 connected together failed. If i had set something else wrong on the tester so that non of them worked i would have re-checked my settings more carefully but as some worked, i assumed i had things right.
While i have your attention, is it worth pairing up output valves for current ie the best 2 with the worst 2 in my KEL84 or is it only worth bothering if you are a supplier with loads of valves to choose from so you can get pretty near identical pairs?

Thanks again, Phil.

[Richard]

Hi Phil

It's good when it turns out to be something simple! The different construction is interesting isn't it, I wouldn't have expected that. I had a look in my Avo 160 manual and found,

"When the symbol * appears among the selector switch set-up figures it indicates that an unknown electrode may be connected to this pin internally. To obtain the complete selector switch coding test with an ohmmeter between pins marked * and all others. ...etc..."

So, this was clearly known about as the figures I gave are the ones I use, 041 230 605, but they are actually written in this manual as *41 23* 6*5. We live and learn!

Yes, pair them up whatever their age and retest in 1000 hours or so, gives lower distortion and noise and may even stop one running away. Generally pairs within 10% current are very good though you can often get away with more. I write the values on the valve as current/slope then you can see if it's aged later, nominal 48/11.