The "Sussex" Homebrew Valve Tester.

[smbaker]

On the positive side, my tester seems to be working really well, and I am thankful to the authors for the design and documentation. I went with a bit of a hybrid design for my tester/experimenter -- I have my own anode supplies from other projects that I integrated into the tester, but I built the Sussex grid supply and oscillator from the schematic.

I used 5V isolated DC-DC converters to power all of my meters. I ended up with a total of six meters. Hardest part of the project was making all those rectangular holes in the case.

[mole42uk]

Are you absolutely certain that the DC/DC converters are completely isolated?
No common ground anywhere? These cheap digital meters are useful but incredibly touchy about their power supplies!
Since you know the AC voltage is correct according to your multimeter, can you grab another supply for the AC meter (even a few 1.5v cells would do) and try that? See if it makes the AC meter behave.

[smbaker]

Originally Posted by mole42uk

Since you know the AC voltage is correct according to your multimeter, can you grab another supply for the AC meter (even a few 1.5v cells would do) and try that? See if it makes the AC meter behave.

Yeah, I already tried powering the meter separately from one of my benchtop power supplies, so that it was completely isolated from the tester. Same behavior.

[Dekatron]

If you have an oscillator with sinusoidal output signal then you can connect it via a diode to your AC-meter and use your DMM in parallel to see if they show the same voltage (just use a DMM that you know shows the correct voltage on half wave rectified and non-rectified sinusoidal voltages). You can use a potentiometer as a voltage divider if you don't have a variable output voltage.

This setup will ensure that you measure the half wave rectified voltage, shorting the diode will measure the non-rectified voltage. Don't use a smoothing capacitor as you will measure a DC-voltage then, with a small ripple voltage, and not the non/half-wave voltage.

I have had problems with a lot of the cheap Chinese meters, both LED & LCD showing the wrong values as well as being non-linear. I always check the ones I buy and compare the measurements with my Fluke 87. Some of the cheap meters were just as poor as my old Beckman RMS 3030 meter which was some 10% off on AC voltages, and even worse on half-wave rectified voltages.

/Martin

[smbaker]

Originally Posted by Dekatron

If you have an oscillator with sinusoidal output signal then you can connect it via a diode to your AC-meter and use your DMM in parallel to see if they show the same voltage (just use a DMM that you know shows the correct voltage on half wave rectified and non-rectified sinusoidal voltages).

With a 250mV Vpk 60Hz half-wave voltage applied (as shown on my scope), my Fluke reads 90 mV and the DPM reads 67 mV. It's pretty much that way throughout the voltage range -- the DPM reads about 75% of what the fluke reads.

How do you interpret this?

Also, for what it's worth, if I reverse the signal connections to the DPM, it gets completely wonky and both the DPM and the fluke start jumping around wildly.

[Dekatron]

I don't know what type of fluke you have but I guess it is of the True RMS reading type, the measured value suggests that as it is within the tolerance.

Full wave 250mV peak voltage should read 176.8mV RMS.

Half wave rectified should read half of that, some 88.4mV which is what your Fluke reads, plus/minus its tolerance.

The DPM seems more like mean reading meter than a True RMS reading meter to me and the strange behaviour when you switch polarities makes me wonder if you have gotten an accidental ground loop. Otherwise the DPM must be of a very poor design or broken.

[smbaker]

The new meter arrived today and is reading correctly. I'm now using the same backlit blue LCD meter that I see in many of the other Sussex testers. Kinda blew my nice RED LED color scheme, but I think I can live with that.

[mole42uk]

Excellent that it's working!
Shame about the colour scheme......

[Tony Jaques]

Hi there,
This information is probably no use to anyone who has already built the "Sussex", but it could be useful to someone who has not yet got to the chassis bashing stage.
I felt that because of the similarity between the AVO testers and the Sussex in respect of setting up valve connections it would be worth trying to use a set of thumbwheel switches. They would also look better than a row of rotary switches, be easier to wire up and as it happens they also cost a lot less.
A search for "thumbwheel switches" on Google or Ebay will lead you to a set of ten decade output switches with end cheeks for less than a fiver. When I got them - from China of course, (but by air mail!) I found that the switches are
marked KSA-2. Unfortunately an internet search failed to find any specification, so I decided to try a destructive test on one.
I first passed a current of 1.6A ac between connection 3 (a likely heater connection) and common. I must admit that I was slightly surprised that the printed tracks did not vaporise. I decided that a fair test (in view of the intended purpose) was five minutes. After this time the PCB was warm to the touch. The voltage drop was 0.3V. This equates to a contact resistance of 0.19 Ohms.
This test was repeated several times.
Next I applied 450 Volts between two adjacent connections (5 and common) for a few minutes. The switch did not break down.
Note that I made no attempt to change the switch setting with power applied because I felt that it would be pushing it - especially on the current test - and because the switches would normally be set up before any power was
applied in the tester.
So there you have it. I have probably stressed the switch well beyond what it is designed to stand and got away with it. I intend to build it into my valve tester in such a way that if the worst happens I can replace it with a panel of switches. I hope to get away with it!
What you chose to do is of course entirely up to you and entirely your responsibility.
Tony.

[smbaker]

I finally finished mine today -- I'm actually calling it my 'high voltage experimenter', but it doubles as a Tube Tester. The high voltage supplies are variable DC switching supplies with 10-turn pots, but the control grid board and the meters are from the Sussex design.

[retailer]

Hello Tony

I also did the same as you and purchased thumb wheel switches, in testing them I found the same as you that they 'should' do their job although I ran 2.5 amps through them. I initially purchased switches that were in banks of 4 with standard .1" pitch pins for the connections, they fit nicely on standard vero or perf board. A short time later I found thumb wheel switches in banks of 3 with tag strip type lugs for the connections on a surplus electronics site, they were cheap so I couldn't resist, they also passed the voltage and current test. If any one is contemplating thumb wheel types you need to be careful not to get BCD switches there are plenty of these about and at first glance look like regular decade types.

I did look at interlocking push button type switches, the type often used as band change switches on radios. Setup in columns, each column would control 1 valve pin and each switch on the column would be setup for anode, screen, grid, cathode etc. It would make for a very compact setup on the front panel.

I haven't built the Sussex yet as I had built the "RAT" tester designed by Steve Bench a few years before the Sussex surfaced, as my first build was a bit experimental using a lot of reclaimed components I had intended to build a second using a proper case but then along came the Sussex. I did a bread board type lashup of the Sussex for testing and found it not quite as convenient to use as the RAT tester as there was some interaction between the anode and grid controls, reminded me of the setup described by Avo to make a standard reference 12AT7 . I'm probably leaning towards another tester using the Steve Bench design as there is no interaction between the controls.

The Sussex is a fine design and the beauty of it is that there is a lot info available, many forum members are ready to help and advise, and very importantly the Avo data manual can be used for settings. Not only is this not possible with the RAT tester you are pretty much on your own if you decide to go ahead with it.

[Tony Jaques]

Hi there,
I have just completed my "Sussex" valve tester and I am well pleased with it. Like many others I owe a debt to Mike
Rowe. It is a shame he is (probably!) in no position to know of my gratitude.
I was going to attach external and internal images, but have not been able to. It is going to take me some time to
work how to make the quality so bad that they will fit the accepted parameters!
Tony...

[Tony Jaques]

Hi Triode,
I was pleased to see the response. As my first posting it was pleasing to see that someone read it. I like the idea of a push button matrix, and if my thumbwheel switches fail I will remember that one.
I have seen not interaction between Grid (1 or 2?) and anode supplies, but selecting the electrode leakage tests does change the control grid voltage because of the way that the 45 volt supply is applied (I think). This does not affect the results in Test mode though.
Tony

[Tony Jaques]

Hi there,
Having recently completed my "Sussex" I would like to suggest a few simple modifications. These are in no way
intended as any criticism of the design as presented; just changes that I found beneficial.

1. Having been "bitten" a couple of times in the early stages of construction and testing I added a 120k Ohm 2W
bleed resistor from the unused contacts of RL1 to ground to discharge the reservoir capacitor after switch off. The
120K was a compromise between discharge time and dissipation.
Incidentally, one of the bites was two days after last use. My reservoir could really hang on to its charge!

2. Because my HT supply happened to come up as 297 volts I changed the 15 Volt Zener diode in the supply stabiliser for an 18 Volt one instead. Not really necessary - but easily done.

3. In the leakage detector I changed R31 to 10K Ohms. In the event of a direct inter-electrode short the base
current with a 1K Ohm resistor is about 45 mA, which seems a bit excessive. More importantly the dissipation in R31
(0.6W) would be 2 watts! With a 10K Ohm resistor the dissipation become 0.2 W and a 1M Ohm leak can still be seen. Actually so can a 10 M Ohm leak (just)!

4. By mistake I ordered two 200mV meters instead of one and one 200V meter. So, because many valves require only a few volts of grid bias to be set, I decided to make a "feature" of my blunder and provide a 20 volts range as well as a 200 Volt one with simple resistive dividers.
This has already been quite useful so I would recommend that the grid volts meter be made a 20 Volt one (not
200mV!) with a switch to select an attenuated input for the 200 Volt range.
A penalty is that some awareness in use is needed because the decimal point stays put!

5. Because of some dodgy Gm readings with EF91s suspected to be oscillating I added an FX1115 ferrite bead to every valve pin. That fixed a problem that could easily happen with other valves.
Actually I have seen mention of this elsewhere, but I would add the comment that a bead needs to go on every pin
because it can not be known just which pins might be anode and grid pins.

Lastly, just a comment: I cannot recommend too highly that the LM317 version of the DC Heater supply is used. It
is far more stable than the original, and easier to make too.

I hope that that is some use to someone.
Tony.

[Tony Jaques]

Hi there,
This is not of any practical value, but might be of academic interest to someone, and might just reassure
others.
A few days ago I completed my "Sussex" valve tester, and of course the first thing that I had to do was round up
every valve that I had (not that many) and do some testing.
Everything was fine and I was very impressed with the tester - until I came to a DF91. I got nearly the expected
anode current (3.3mA instead of 3.5mA) but a Gm of zero! Another valve gave the same results.
Then I came across the comment about directly heated valves on page 40 of Keith Wevill's excellent manual. Oh
yes, of course. But just a minute; I did get a current flow. How? Where's it going to? And more important, could
it damage something?
Eventually I worked out the sneak path. For anyone interested it is from the heater H+ - RL2/11 - RL2/12 -
LED2 - R28 to the -45 volt line and then through the -ve power supply to ground. So the Heater/cathode light comes on as it should - but not for quite the right reasons.
The reason that approximately the right Anode current is shown is that the very high anode resistance of a pentode
(DF91) makes the valve behave as a constant current source, so with any reasonable "cathode" path roughly the right current will flow. But unless the path is a lowish impedance the gain, and hence Gm will be very low.
With H- earthed it will be found that in the Heater position of the function switch the Heater does not show as
functional (LED4) but the Heater/cathode LED does illuminate. Because I have used the much better LM317 1.4
and 2V heater supply there is now a path for LED2 from H+, through the resistors of the heater supply (135 Ohms), back to H- and therefore ground. LED2 (red) lowers the voltage on LED4 (Green) cathode to about 1.6 Volts so LED4 goes out! But at least Gm now reads OK.

Sneaky!

I suggest that the connection between H- and Cathode (earth) is done after the Heater test is done but before
switching to the inter-electrode short tests.

So can damage be done if the earth connection is forgotten? I don't think so. The current flow is low, but
in principle at least the voltage across the valve is increased by the presence of the -45 Volt supply in its
path.

Not a lot of use perhaps, but I hope it was interesting to some.


Tony.

[Tony Jaques]

Hi there,
I have attached a couple of images of my newly completed tester in case anyone else might find them interesting or even useful. They are I think pretty self evident, but a couple of comments may be helpful.
The most important thing is that I deliberately made the case a bit less in width than the length of A4 paper so that I could use my computer printer to tart up the panels, using sticky backed photo paper.
The copper winding visible on the larger transformer is my addition of a 5 volt winding tapped at 4 volts. I was lucky to be given a transformer (thanks Jed) that had enough of a gap to let me do this.
Of the two smaller transformers the one on the left is of 15-0-15 volts 3VA used to supply the -45 volts (actually
-50v) The one on the right was from a wall wart. It was stripped down and three new 9 volt secondaries wound on it for the displays which need all of 6mA each. Then an aluminium clamp was bashed up to make it chassis mounting. Cheap - but it works.

Tony.

[ian_rodger]

Tony,

Looks good. Any chance of a more detailed photo of the thumbwheel switch connections - I'm tempted to go down a similar route. Looks like you have added ferrite beads at the valve bases only?

Thanks,

Ian

[Tony Jaques]

Hello Ian,
I have "attached" two images that should show you more detail of the switching. Note the edge connector on the switch the centre "C" is the pole.
So wiring becomes very simple: The 9 "C"s connect to the valve base pins, and a wire is run through each of the similarly numbered holes, (except holes "0" because these are for "No connection"). These commoned connections are fed from the power supply arrangements of course.
Yes, I did only put ferrite beads on the valve pins. I have not yet looked
for signs of oscillation with the 'scope, because I don't think that there is any. But I do intend too.
I hope that that was some use.
Tony...

[mole42uk]

Using the thumb switches certainly makes for a smaller box!

Please do let us know how they perform in service - there was a lot of discussion about 'modern' thumb switches and their current rating for valve heaters. I'd be very interested to see if yours work well.

[Tony Jaques]

Hi mole42uk,
I waited a while to see how things would go. You should understand that my tester will not get a lot of use. It's one of my favourite classes of tool: not often needed but a treasure when it is.
The switches have given not trouble over about some 40 valve tests, but there is one
thing to bear in mind; the volt drop that I reported earlier. I recently was asked to check a couple of GZ32 bottles. These things take 2 Amps heater current at 5 Volts. The valve finished up with 4.5 volts. Now for a rectifier it is not that important, Basically if it hits the 50mA current limit with a few tens of volts applied then it is probably ok. But it might matter if testing some of the old higher function valves with a 4 volt heater.
For me that is so unlikely to happen that I am happy to stick to my rotary switches. If desperate I could always use an external supply.
I hope that helps.
Tony...