Homebrew Valve Curve Tracer using re-purposed AVO 2 panel tester.

[Electronpusher0]

This project will probably be viewed as sacrilege by some, so those of a sensitive nature should look away now.

I have been contemplating building a valve curve tracer for some time and had gone quite a way in designing it when I came across this thread on the forum:

https://www.vintage-radio.net/forum/...t=valve+tester

This describes a curve tracer using an assembled pcb from Alain Ducrocq Trading as Radiolec in France. This looked exactly what I wanted so decided to build it. I bought the pcb from Alain’s e-bay shop.
It needs at least three transformers / secondaries, one at 24V for the pcb, one for the heater supplies and one for the Anode HT. The Anode supply is simply half wave rectified without any smoothing
My wish list included multiple heater voltages and several different Anode voltages.

I have two, two panel AVO valve testers, one complete and restored in wooden box and one that is just the control panel. This one has a meter that had lost its magnetism and was the subject of my attempts at re-magnetising, reported previously.
This unit, of course, has all the transformer voltages I need so I decided to re-purpose it. The meter, centre switch and pots were removed and stored as spares.
In the original the Anode supply is selected by the “Anode” switch from tappings on both transformers, I rewired this to only use the one transformer leaving the other for the heaters. This also left a spare secondary from which I could derive 130V to use for the screen supply to test pentodes etc.
I had to re-label the “Anode” switch settings to reflect the new tappings and to indicate the peak voltage rather than the rms. (because of the half wave applied to the anode).

Alain’s board was mounted onto a sub chassis, in turn mounted onto the old meter mounting points.
There is also a 20v secondary (measured 22V off load) on the HT transformer and I used this to supply the pcb. This did not give quite enough volts to achieve the -36V needed for the grid supply so built a voltage doubler on some matrix board to supply the -36V from the 20V secondary.

The screen supply is derived from the 120V secondary as mentioned above, this is doubled and provides 360V dc for the screen regulator. The AVO screen switch had all its transformer connections removed and replaced with a resistor chain this selects a voltage from a Zener chain with a total voltage of 300V. The wiper of the switch goes to a MosFet pass transistor with a simple transistor current limit. Alain also supplied the pcb for this supply.

The 3 wire data cable emerges from the hole where the meter centring screw once sat.
I had a similar problem to Chris (see link above) with the software not recognising the serial port adaptor I was using. This problem is not unique to this project, it has been reported on the forum in connection with other projects. I am using an expressbus serial adaptor and the first (cheap) one I tried did not work.
This was because the adaptor did not fully emulate an RS232 port but appeared to the computer to be a USB port. I swapped this for a Startech EC1S952 port native adaptor that uses a UART and appears as a full native RS232 port.

The software then ran and does work well. As well as drawing the curves you can specify the current to perform measurements at and it will calculate gm, ra and mu.
As Chris reports, the messages are in French but the software is pretty intuitive. I attach plots of 2 halves of an ECC82.

However, like Chris, I had problems testing Pentodes, the curves were far too flat and the Anode currents far too low. The oscilloscope revealed that screen supply was completely collapsing during testing. I then tested using an external bench high voltage supply at 250V. This gave far better results with just a bit of wiggle on the trace, this is due to the limited response of the regulator in the external supply. The screen currents were peaking at close to 40mA on the analogue meter on the bench supply so probably much higher.
I will use Chris’s solution and leave the switch selectable screen supply in the AVO case and put a current boost supply in the valve / patch panel case.
The original transformers in the AVO hold up well during testing, you can hear the current pulses being drawn, scoping the Anode supply shows minimal droop.

The absence of the second panel means that I have had to make up a box with patch panel and valve bases. All the valve bases have ferrite beads on every pin.

Peter

[Electronpusher0]

More images

[Radio Wrangler]

The use of half-wave rectified and not smoothed AC for the anode provides the sweep for curve tracing. But when the anode voltage swings below the screen voltage, electron flow in the valve will partially switch from the anode to the screen. This gives the famous tetrode kink in the anode current and also gives a peak in screen current demand.

For this reason you need a stable, low impedance DC screen supply. It sounds like the arrangement used isn't up to the current demanded if it is getting pulled low.

David

[Electronpusher0]

Thanks David, I am sure you are right, certainly the bench psu gives much better results.
I have been following the HV power supply thread
https://www.vintage-radio.net/forum/...d.php?t=167901
and will change the screen supply to use the Sussex valve tester scheme of LR8 and MOS fet pass element.

Peter

[David Simpson]

Peter, I wouldn't worry about re-purposing an old AVO 2Panel Tester. I 've been doing similar morphing of AVO, Marconi, etc. stuff for ages for my analogue valve testing projects. And, encouraging other to do so.
Your Curve Tracer project is great, just like Chris Wood's back in 2016. I take my hat off to you both. Once again, my analogue efforts pale into insignificance. My fault, with my mindset being stuck in 1960's analogue times.
Certainly a great example to younger Forum members. I can see this as the way forward in the 21st century for valve testing. 'Specially as 1940's/50's/60's old AVO VCM's, Taylors, Mullards, etc. are falling by the wayside of late.

Regards, David

[chriswood1900]

Well done Peter, another good example of Alains design, of all the problems sorting a screen supply was the hardest, it is a pity it does not form part of the design, but I think he was mainly thinking of triodes when he did it.
I still want to make some improvements to the screen supply like the updated supplies in the Sussex. I still use varying the LR8 which supply collapsed until I put a transistor with it. It does not provide much protection if the valve fails under test, but I did add a fuse.
As David said nothing wrong with re purposing an AVO case if the guts are redundant its not a historic artifact there are 100s still around, its a pity you don't have the other half of the AVO to provide the different valve holders.

[David Simpson]

I might have the other V/H half lying spare, but have removed the Thumbwheel Switch assembly as a spare for my MK3. Will have a shufti in my storage caravan.

Regards, David

[Electronpusher0]

That would be fantastic if you do have one David, even without the thumbwheels it would give me a selection of valve bases, even better if it still has the 1.5V heater transformer in it.
Thank you all for the encouraging comments, I do not feel so guilty about using the AVO panel now.
I have ordered a transformer to make an uprated screen grid supply, the same one I used for the magnetiser project. I have a selection of HT transformers but the secondary voltages were all too high.

Peter

[Radio Wrangler]

Originally Posted by Electronpusher0

I do not feel so guilty about using the AVO panel now.

David Simpson did indeed do a bit of AVO butchery a few years back, which might have raised ire in some people, but then he used the DC rig he'd built to accurately calibrate some valves, which he distributed and circulated around the people using untouched AVOs so that their calibration could be checked. A very generous thing to do considering the tedious work in plotting point by point.

So, if anyone gets stroppy about what he did, there will be a small uprising in his defence!

I went down to Rob Rustbridge's last auction in Lancashire, made a trip from Fife specially and there were a few items that seemed unloved and went for trivial bids. Rob had said he wanted to pack the minimum amount back in his van. (The pasties were excellent, by the way). So I bought a couple of them thinking of people who I could give them to. One was a geiger counter (for a friend living in Dalgety Bay - look up news reports) and one was a little crude valve testery sort of thing with a panel of assorted valve holders and a couple of nice meters. Later last year I took it up to David, and he seems to have been having good fun with it since. Plenty of assorted valve holders, but nothing equivalent to the roller switch.

The AVO VCMs are gorgeous machines, but they are flawed. The meter is too fragile and is too easily damaged by unsympathetic users and mis-set controls. The reliance on calibrated valves to check their accuracy and in some cases requiring rare voltmeters to calibrate them is a nuisance. If you have a measuring instrument and you can't independently verify its results then you should feel a little less than confident in those results.

But demand from guitarists and hifi buffs for valve testers has simultaneously pushed their prices high, and put them into the hands of people who don't know how to operate them carefully.

Back in the sixties, I bought a VCM MkII from Jim Fish, I think it cost me a fiver. Everything was going transistorised and it had started looking like a white elephant. I tested every valve I had, and plenty from other people. I didn't know that the meters were fragile and wasn't careful enough. Even though the VCM MkII has a higher current movement then the later models, the meter eventually went. I guess I'd had my money's worth out of it. Those transformers got turned into a bench power supply. The panel of sockets with the roller switch went in the dustbin as I couldn't see any further use in them.

No doubt we all look back at things we wish we'd kept....

Anyway, I picked up a little Taylor 45D in the eighties at the G-QRP doo. I think someone was asking £20 so I didn't grumble.

This reminds me that there are a few boxes of valves the Kingdom ARS has inherited to check through and see what's worth keeping.

David

[David Simpson]

OK then Peter, PM me your contact details. Yours for just P & P. At the moment its installed into the top of a Marconi TF995A/2 empty cabinet( a redundant project), along with another made-up VB panel containing spare valve holders. You can have the whole caboodle if you want it. I've still to find the wee transformer. 'Fraid I'm maxed out space-wise with VT projects in old TF995 cabinets.
Yep, I'm extremely grateful to David (RW) for giving me the old Roberts V/T last year, & more gratitude to Ed Dinning for the Handbook & a spare Roberts.
To me, this passing-on of old redundant equipment to others for re-purposing, is at the heart of the Forum camaraderie.

Regards, David

[Electronpusher0]

Thats very generous, thank you very much indeed.
PM sent

Peter

[retailer]

Looks like good work there on the curve tracer - I've often thought about making one but I have a long list of to-do's I doubt I'll ever get to it.