The "Sussex" Homebrew Valve Tester.

[Mogliaa]

Hi everyone,

I'm finally in the process of building my tester. Has anyone thought about adding an extra AC meter to measure the valve mu? i.e. Steve Bench has this included in his tester and would be great to be able to measure mu in addition to Gm and therefore calculate Rp at the operating point.

Cheers,
Ale

[ColinTheAmpMan1]

Hi All,
I am new to this forum, having stumbled onto it a few days ago. I have been trying to find a mutual conductance valve tester for ages, but even the emission types are a bit pricey. This looks to be something to get me interested. I see that no messages have been posted since the end of last year and was wondering whether the custom transformers and PC boards are still available. I have seen that the digital meters are available on eBay, so that seems ok. What a great project; Mike is to be commended.
Colin.

[Ed_Dinning]

Hi Colin, yes I still wind the transformers on demand, PM me for further details.7

Perhaps the other contributors will let you know about PCB's, I know a batch were made at the start of the project but I'm not sure if any remain.

Ed

[G8UWM-MildMartin]

I have just one undrilled PCB left.
(Plus 8 that I have built up but still not got round to testing more than the oscillator on.)

[ColinTheAmpMan1]

All,
I read somewhere that the DMMs from the Far East exhibit non-linearity. If true, this might indicate inferior design and manufacture, or it might just apply to multimeters. I was wondering if anyone had any experience of such non-linearity or any comments to make?
Colin.

[neun elf]

Hello everybody,

My Sussex is finished,working and testing valves but I am not too sure how to interpret the results. I assume that a higher gm/V reading than the one given in the Avo manual is a good thing, but at what point does a lower reading equal a "dud" valve?
Similarly, what is indicated by a higher (or lower) anode current, compared to the one given in the manual.
Digital displays are great, but there was something to be said for the colour-coded scales on the AVO testers!
Sorry if the questions are a bit daft, but any help would be greatly appreciated. Thanks.

[Dekatron]

The RED/WHITE/GREEN or REPLACE/FAILING/GOOD scale AVO used is mentioned in the AVO CT160 manual and described as the centre point on the GREEN/GOOD scale being 100%, with an upper limit at 130% and a lower limit at 70%. The WHITE/FAILING part of the scale has a lower limit of 50% and an upper limit of 70%. The RED/REPLACE part of the scale is below 50%.

Let us say that you test a valve that should read 10mA/V according to the AVO VDM, that then corresponds to GOOD between 7mA/V and 13mA/V, FAILING between 5mA/V and 7mA/V and REPLACE below 5mA/V.

You'll just have to compare your reading with those figures for that valve. Just multiply the AVO VDM value with 1.0 for 100%, 1.3 for 130%, 0.7 for 70% and 0.5 for 50% and you'll get the limits.

This is what a moving coil meter can tell you with just a glance on the scale but which a digital meter readout can never tell you - if it does not know about the data in the VDM somehow and can adjust its display reading for that data.

It would be nice to have an LCD display with a meter face shown on that display so you could just look at it to get this kind of "understanding" of how good your valve is compared to the VDM. Maybe someone with the skills and knowledge can make something like this from a colour display and a small microprocessor, whishful thinking from my side here - maybe it even exists such a meter somewhere?

This could even be an "upgrade" or replcement for broken meters in AVO VCM's too!

[neun elf]

Thanks Martin,

You have confirmed my understanding of the gm/V for any given valve, but sometimes I get a gm/V figure that is near the value I expect, but the anode current is higher(or lower) than the figure given in the AVO valve manual, and I am not sure what the significance of this is, or how to interpret it.

[G8UWM-MildMartin]

Possibly it's something of no consequence if using auto-bias (cathode resistor) but would require a slight adjustment if using a grid bias voltage.
You can always adjust the bias voltage under test to give the rated anode current and see if it's reasonable.
(If you still can't get the specified anode current at a much less negative grid bias voltage, that suggests the valve is low emission i.e. worn out.)
In fact you can alter it, take the readings and plot the results on graph paper (or modern computer equivalent spreadsheet etc.) which is, after all, why the AVO instruments were named Valve Characteristic Meter...

[Andymic]

Originally Posted by Dekatron

It would be nice to have an LCD display with a meter face shown on that display so you could just look at it to get this kind of "understanding" of how good your valve is compared to the VDM.

Interesting Idea.
Should be quite easy to do.
You could stack a load of rectangular LED's (say 5 red, 2 Yellow and 5 Green) and use a transistor to switch each one on at a predetermined level.
Used to do that with audio amplifiers so you'd get a moving band display to show the output level.

Andy

[Pamphonica]

Remember that you would still need another control to set the "optimum" figure for each type of valve you are testing. A simple red/white/green display based on the Sussex gm meter fsd is meaningless.

The older Avo testers use a control to set the target gm (eg the "1mA/V" point you can see in Martin's scale picture above). Then the red/green scale is correct.

The newer VCM163 does have a mechanical overlay over the gm meter scale to show pass/fail, but unless I am doing a quick test on a lot of the same valves I just note the target value and read off the actual gm.

[Dekatron]

Lascar Electronics have a series of panel meters they call "Panel Pilot" which have programmable LCD TFT displays. I do not know how they can be programmed as I just found them but they sure do look nice, and they are expensive too of course. Check their homepage at: http://www.lascarelectronics.com/ and also http://www.panelpilot.com/ .

They might not have the resolution of an old original meter, but with software you can make whatever you want with them (within their limits of course). I especially like the feature where you have an analogue display as well as digital figures below that.

[Darren-UK]

A post regarding swordholder, who started this thread, has been moved to here.

[Bazz4CQJ]

Originally Posted by neun elf

Digital displays are great, but there was something to be said for the colour-coded scales on the AVO testers!

The problem of interpretation is still present with the AVO's, though the sub-scale in the gm meter can be used to some degree. My AVO sometimes shows brand new valves with Ia and gm values well in excess of the book values and it could be assumed that those values might fall towards the spec quickly as the valve burns in. Other brand new valves give values appreciably below the book values. I have tested valves which are working perfectly well in various pieces of equipment and been really surprised how low their results were on the tester.

The one thing you can do which helps provide confidence that a tester is giving true results is to use your tester on 'standard' valve which you have tested "by hand", i.e. set it up in a really basic text-book circuit and measured delta Ia for delta Vg under the specified conditions. I think it was Martin Scobie who suggested using ECL80's for this purpose as they are very cheap to acquire and they provide 2 valves in one envelope, so provide two standardised valves. Of course, you then expect your tester to give you much the same results on the standard valve. But the bottom line appears to be that valves often do give test results which are appreciably off the book values, but may still be quite serviceable.

[Ed_Dinning]

Hi Gents, I've been asked to wind some more transformers for the Sussex; these will be ready at the end of March, so PM me if anyone wants one.

Best regards, Ed

[ColinTheAmpMan1]

Hi All,
There was some discussion about roller-type switches some time ago and some folks used them in their own Sussex VT. I have purchased some from an eBay seller in the USA which seem to be good. They were in banks of six, so I had to buy three banks. They are made by Digitran and seem to be 45000 series Slimswitch. Although the rated load is 0.125 amps at 28 VAC or DC, the other specs give a non-switching current of 3 amps and a dielectric strength of 500 VRMS. Here are the spec sheets:

http://www.digitran-switches.com/spe...thumbwheel.pdf

I have, of course two sets of nines, only one of which I need to make a Sussex VT1.
Regards, Colin.

[Colin]

Hi Colin, thanks for posting. I'm interested in your spare set. I'll PM you.

Regards
Colin

[mikeydee]

Hi am formulating my sussex VT1. I have the parts but have a couple of questions about wiring and parts layout used. Would be great if anyone can give suggestions

1. Any specific cable Sussex's have used - was it single core / 702 / 1602 / Screened in places etc.
2. Also I am following Neun Elf's case design of the RS service case and the distance between the transformer and the case is going to be tight. I am not too worried about insulation as I can tackle that but I was wondering if there are any magnetic effects that should influence what the transformer is located next to.
3. Are the Ferrite beads recommended to be attached to every valve base pin or only certain pins.

[ColinTheAmpMan1]

Hi Mikeydee,

Here are my observations, though other, wiser, heads might want to correct me/add to my words:

1. Any parts of the circuit that are associated always with the heaters could use solid-core wire of a gauge that will carry 3 amps AC. This wire is probably best solid-core, as it would probably be advantageous to twist the heater wires tightly to reduce hum (It might interfere with the gm measurement). Bear in mind that if these wires are moved regularly for maintenance, then the use of stranded wire will eliminate the inevitable breaks that would happen with solid-core.

2. Any valve-base pins can be switched to any function and so should be considered to be heater, anode, cathode, screen or control-grid connections and treated appropriately. That is, they should all be wired with thick enough wire to take heater current and have ferrite beads to suppress oscillation. The top-cap control-grid connection should have a ferrite bead too, I would think (and the anode top-cap?).

3. Small valves (those on B9A and B7G, for example) rarely, if ever, need huge heater currents, while those on octal and similar bases are frequently output valves or power rectifiers and these do take a larger heater current. This would suggest that these valve-bases should be closer to the heater supply, so that lead-length is minimised. The wiring to these bases could advantageously be heavier gauge wire, with perhaps thinner wire connecting to the other bases' pins.

4. I wouldn't have thought that there was likely to be too much problem with magnetic effects from the mains transformer, just so long as you adopt the same logic as if you were constructing an audio amplifier; keep the transformer and power-supply components as far from valves that might be small-signal amplifying devices as you can. It might also be useful to keep the transformer as far away from any potential beam-tetrode valves as possible, too. I suppose this boils down to the B7G, B9A and octal bases, but probably all of the bases that you use, really.

5. If you look carefully at Les' useful document, you will see that his circuit diagram specifies screened wire in some places. Any wire that is always associated with control-grid circuitry should probably be treated this way. It would be overkill to assume that since any pin could be a control grid, that all valve-base wiring should be screened. Going that way, all of the wiring would be heavy gauge (heaters) and screened (control grids). Not a practical situation, me thinks.

I hope this all helps; good luck with the build.

Regards,
Colin

[ColinTheAmpMan1]

Further to what I said before, I had a look into my valve data book and came up with the following snippets.

6.3 volt heaters:
EF91 pentode on a B7G base takes 0.3 amps heater current
ECC8_ dual triode on a B9A base takes 0.3 amps heater current
EL84 output pentode on a B9A base takes 0.76 amps heater current
EL34 output pentode on an octal base takes 1.5 amps heater current
6V6 beam tetrode on an octal base takes 0.45 amps heater current
6L6 beam tetrode on an octal base takes 0.9 amps heater current
KT88 kinkless tetrode on an octal base takes 1.6 amps heater current
807 beam tetrode on an UX5 base takes 0.9 amps heater current
EZ80 rectifier on a B9A base takes 0.6 amps heater current
EZ81 rectifier on a B9A base takes 1.0 amps heater current

5.0 volt heaters:
5_4 rectifiers on octal base take 2.0 amps heater current

This might give an idea of which valve bases are likely to need the highest heater current provision and therefore be closest to the pin-selector switches. I would put them in the order octal, B9A, UX5, B7G.
Colin.