anyone built a valve tester ?

[Steve G4WCS]

Im just wondering whether anyone has had a go at making a valve tester. I guess the prinicipal parts would be a filament supply, HT supply, some sort of stimuli (unrectified ac), and measuring device.

I also guess the practical difficulties would lay in
1) multiple filament supplies required
2) multiple HT voltages ?
3) a lot of miltiway switching
4) getting hold of the required transformers, valve bases ( and maybe rotary swtiches in this day and age )

another difficulty may lie in getting hold of all the required characteristic data to determine whether the valve was within spec.

It may be a good little project, on the other hand it could prove to grow into a monster that would end up costing more than buying a good one off ebay or similar !

Anyone had a go? or can offer any advice ?

[Ed_Dinning]

Hi Steve, I've built several over the years with varying degrees of success and now use commercial ones. I does however teach an awful lot about valves and is well worth the time spent.
There have been many article in RB, Radiophile, BWVS bull, the Bernards books and PW etc so you should be able to fins some circuits, let me know if you need any copies.
Heater voltages can be derived from a DC power supply if required, programmed by a series of resistors (on the prog input, not as potential dividers) ~ the LM317 could be used here. It is also easy to wind transformers. RS do a kit with the pri already wound.
When you have studied the circuits you will find the testers often used 1/2 wave rectified AC as a means of testing: again more tapped transformers. Some circuits used DC supplies and an oscillator for u (gain) indication.
Straight DC supplies will only give you the static valve characteristics, but by adjusting the grid volts by about 1% it should be possible to get a measure of gain.
You will also need a high voltage supply for testing electrode insulation, with series resistor and possibly a neon.

I hope this hasn't put you off, mail me if you need additional information.

Best regards Ed

[Duke_Nukem]

This is something I've been thinking about on and off for quite a while.

Hardware-wise I think the wiring and switches are a nightmare. However things can be simplified to some extent by running heaters from a stabilised DC supply and programming via S/W ... enter the PC. Some pin connections that might need to be swapped just might be achieved using FETS ? Also good protection could be provided (no more stuffing 37 volts up that EF80's heater!).

Use the PC as the intelligent front end, connect via serial to a PIC micro in the valve tester which in turn sets and monitors the various volts and current.

The other thing to consider is the PC's sound card. With this you could generate arbitary test waveforms and also digitise resulting waveforms for display i.e. potentially automatically plot the various characteristic curves.

Given someone already had a PC, such a system would certainly cost less than an old characteristic meter and also boast more features too.

Pity the poor person who has to work out all the settings in the first place though !

As regards mains transformers, custom wound ones can be within budget. I got a custom 400W toroid job with perhaps half a dozen windings on it for ~£40. By using programmable DC voltage regulators you wouldn't need anything like as many secondary windings as an old tester.

TTFN,
Jon

[pmmunro]

This would certainly be an interesting project, if you have enough mental stamina to see it through. Was the cost of commercial valve testers an incentive? It might be worth looking at the original Avo patents for valve testers, (downloadable on the net), which give condsiderable insight into their design philosophy.

Has anyone ever thought how much it would cost to make say, an AVO VCM163, today? I don't think Avo's present owners would be intersted because of image concerns, even if it could be sold at an economic price, which might be possible given current market prices.

[Steve G4WCS]

weel thanks for that, a few good ideas already.

Just bringing to mind what I already have lying about, Ive a couple of 5V programmable switch mode power supplies which would do the heaters with ease, and somewhere have an old solid state 1kv at about 50mA supply that could be used for insulation tests. Ive also got various meters etc that can be utilised.

Cost would be a factor in my decision to knock something together, I have a colleague with a taylor tester who can be pressed into service to do the odd valve, but I have approx 3 biscuit tins full of mainly small radio and TV valves, some of which may be new, or ex set pulls (although in the correct packaging), the odd oldie in there, none of which are worth particularly much, and I would like to go through them all at some point, whizz the bad ones and list the good ones fo my future projects, but it doesnt justify the cost of buying a tester just yet.

I had thought that if I can knock something simple up to do the miniature 9 pinners and a few octals that would do me.

If its no trouble to anyone, circuits or any other info would be gratefully received, but dont go to any bother, its just a back burner project even if it takes off. Of course it would be nice if I could do a reproducible design rather than a one off, but I suppose everyones junkbox and requirements are different.

I also quite like the PC/PIC idea, but Ive no idea how to even start with that, Im the age now that Id have been relegated to sweeping up and making tea shortly after the advent of microprocessors !!

thanks all

Steve

[XTC]

I've thought about this one too, and came to the same conclusion. The way to go would be a microcontroller based design with relays etc. controlled from a PC. Then you could have a database of settings and virtual controls rather than a load of wafer switches and all the rest. It would be a challenging project, but you could have it printing graphs of the characteristics.

In practical terms, if you just want a useable instrument, you're far better off buying a factory made one secondhand rather than trying to make one, and that applies to pretty well any test gear. Last time I looked, Taylors and 2 panel AVOS were about £100. MKIVs had been over £400 but the last one I saw looked tidy and went for £250.

With a ready made item, you get a debugged design, often very high quality controls where they are needed, and a nice presentation. If you want to sell it for any reason, an AVO valve tester is an AVO valve tester, but your project is just your project.

If you had some special requirement, like rapidly matching up large numbers of audio valves, and only a dozen types, a purpose made tester might be a good propostion.

There's an article in the Jan 1998 Wireless World on a valve tester specifically for testing octal power valves. Let me know if you're interested.

Pete.

[joe]

I've been playing around with a home brew design for a while now, partly because I'd like to have one, but more in order to keep the grey matter ticking over now I'm retired (although I seem to have far less time than when I was working).

Valve holders were not too much of a problem given swapmeets and rallies etc.
Switches can be got from similar sources and I've got into the habit of buying any I see that look as though they might be of some use. If I get really stuck then there are always the "Maka-switch" kits from RS. In a similar way I've obtained two large 50uA FSD meters for a total of £1.50

Power supplies could be made from winding mains trannies with loads of tappings but in this day and age its probably far easier to use switch-mode supplies with either tappings on a much tinier transformer or with a network of resistors in the feedback to the control element.

Published data can also be obtained from various old manuals that turn up at swapmeeets etc, although I've acquired quite a few over the years.

I've unashamedly studied the circuits of the AVO MkIV & 163 as a sort of guide as well as any other published designs, but I've not seen the Wireless World one that XTC has mentioned. Pete, I've sent you a PM.

The idea of a computerised tester is an excellent one, but whilst I wouldn't mind dreaming up some sort of hardware interface to a PC I wouldn't have a clue about the software that would be needed.

Joe

[igranic]

For those of you who are true valve tester anoraks, you may like to check out the following link to a German website, a sort of cyber valve valhalla. It includes a comprehensive collection of pictures and some diagrams, including details of home-built "pocket" valve testers.

For the benefit of those who do not understand German, you can copy and paste the link into the the Google website translator, which can produce some interesting phraseolgy!

http://www.jogis-roehrenbude.de/Roe-Pruefer.htm

[Ed_Dinning]

Hi Gents, some further thoughts on this topic:
Use a plug board for electrode selection as this is by farv the simplest solution.
Use and LM317 programmable regulator with a simple switch and resistor chain to set the heater volts. This will go up to 30 volts at 1.5A if heat sunk.
Insulation test~ a neon and a 250V supply (with series resistor) and the use of a "shorting switch" as used on coil packs, wher one way is selected and all others are earthed.
I've built several of these in the past, even using RS thumbwheel switches for electrode selection.
Results processing could be via a Nat Instruments "virtual measuring system", these look good, the software can be obtained for free and they can produce some great looking displays.
I bresently have a nice looking home brew teste in a wooden case, complete with meters, switches and bases (no wiring or PSU's) That would be a good candidate for a build as above.
Let's get thinking/designing guys!

Ed

[Ed_Dinning]

Hi again Gents, just visited the excellent German valve tester site (must visit if possible when on hol); it shows an elegent way of selecting pins/electrodes by means of a matrix board. This could be easily constructed with jack sockets and jack plugs arranged as shorting plugs. Any better ideas?

Ed

[Alan_Douglas]

I've used two rows of banana jacks, one row for the socket pins and one for heater, cathode, grids etc. Short jumper cables with banana plugs at each end (I think they're known as wander plugs on your side of the pond) connect the two rows. Weston used the same system in their most elaborate lab model analyzer. Weston provided xtra jacks for the heater voltage metering, to get accurate readings despite the voltage drop when powering high-current valves.

[scotty29]

I built one 2 years ago, to a design from a 1954 Practical Wireless.The most time consuming task was winding the multy tap transformer, which was made easier by constructing a simple coil winding jig, with turns counter,(other wise your mind will start to wander and you have forgot the number of turns)
Anyone interested in the prodject, send me a PM and I will let you have the details, also pics.
Harold

[Mike Phelan]

Something I have been thinking about for a while, not really because I have a need for one, but as an interesting project, while I am saving up ebay proceeds for a CT160!
One possibility is to build a replica CT160, but there are a few stumbling blocks:
The transformer winding should not be too bad a job, but the thumbwheel switches will be a BIG problem, and proper rotary switches are virtually unobtainable, AFAIK. You could make these, but a lot of work.
The idea of using voltage regulators instead of tapped transformers is attractive, and a matrix board as on the German carrot-breeders' site with card overlays is very clever; though I would construct it slightly differently, maybe using a cannibalised prototyping board.
Steve Bench's design uses FET regulators for Va and Vg2 - simple.
The CT160 method of measuring gM with the meter in a bridge is ingenious - better still would be an oscillator that was switchable up to, say, 150mHz - not too difficul as no calibration would be needed. This would root out ECC85s that were no good for 118.7 mHz in FM front-ends.
You would still need rotary switches though - more if you only used one meter.

[Ed_Dinning]

Hi Gents, RS and Farnell have a range of thunbwheel switches, with various coding forms for between £5 and £8 per sector, a little expensive but they would make a smart front panel.
Ed

[igranic]

Here is another interesting web page for those determined to build their own valve tester.

http://members.aol.com/sbench101/#TubeTest

[Alan_Douglas]

Here's a simple high-voltage regulator:

http://www.antiquewireless.org/otb/resto0504.htm

[igranic]

Here is another valve tester web page. The site is also worth a look if you are inteseted in German valve radios.

http://www.oldradioworld.de/tt2.htm

[JHGibson]

If you have a voltmeter and a working radio that the valve to be tested fits, then you have a rudamentary valve tester.

Just earth the control grid and measure the voltage across the cathode bias resistor, then use Ohm's law to calculate the cathode current. Now connect a 1.5 volt battery between the control grid and earth with the negative side going to the grid. Find the new cathode current. You now know what the current change was for 1.5 volts, divide that by 1.5 and you have the mutual conductance in ma/volt. Multiply that by 1000 to get the answer in micromhos.
John.

[scotty29]

As promised. One I built from a 1954 design (from Practical Wireless )

[joe]

Hi Scotty

That looks like a really nice job of work there. If mine turns out half as nice I'll be well pleased.

If you have a copy of the design ( PW 1954) I'd be interested. I'll send you a PM.

Joe