The Sussex VT 2021

[Richardgr]

Many have built Mike Rowe's 'Sussex' valve tester since it was first described here in 2009. At the time it was a revelation, since the only other alternatives were to track down vintage testers, and the prices of those had been steadily rising.

The philosophy of that tester was to revisit the classic AVO Mk IV using modern components. The valve configuration data was set using the same coding as the AVO VT manuals, some parameters specified (heater, anode, screen, grid), and the standard methodology followed to determine the health of the tube:

• Heater continuity
• Inter electrode shorts
• Hot heater/cathode insulation
• Anode current
• Mutual conductance (gm)
• Gas (optionally)

I have completed the last revision of the PCB, and have most of the peripheral components required to finish the build, but I have not done so. To emulate some of the fine examples of completed testers showcased on the original thread, a lot of time has to be expended on the case and wiring; the PCB part is trivial in comparison.

The reason I never completed the build was because I came across the uTracer. For €200 you get a complete kit, including components but excluding bases, a box and a power supply. The power supply is just an old 19-20V laptop brick, so generally not an issue to source. The PCB interfaces to a laptop, so there are no need of controls on the box - it really can be minimalist. My example is shown, with an example of the data from the GUI.

One benefit of the uTracer over the Sussex is that the Anode and Screen circuits are treated identically. Hence the screen current of a pentode is plotted alongside the anode, or in the case of a dual triode, both halves can be plotted together.

One benefit of the Sussex over the uTracer is that it retains the belt and braces thinking of the AVO, with more rigourous testing of the integrity of the valve before launching into mutual conductance readings.

I still have the goal to complete my Sussex. I have been reading through Les Carpenter's manual to get some inspiration. The problem with these bulletin-board designs is that they do not have properly organised configuration management. Really there should be a new thread for each revision of the PCB, and a formal method to gather enhancements and drive revisions. What happens is a single thread gets longer and longer, the design is updated, and it is bewildering to embark on a new build. It is the same on diyAudio with iconic designs like the Baby Huey.

I am hoping that I am throwing down the gauntlet here to see if there is still a demand for a Sussex, and if so how it should be specified.

To get the ball rolling, here are my suggestions:

• Common anode and screen voltage and current supplies, with screen current monitoring. (I am looking at doing this with a push button on the existing design)
• Support a minimalist implementation. I.e. isolation transformer for the 240V supply, bench supply for heaters, some generic transformers for meters and grid supply.
• By combining screen and anode circuits, dual triodes should be testable without switching, with the goal that a simple banana plug layout can be used to map electrodes on different tubes to the tester. Not having to contend with the wiring of 9 x 12 way switches would make the implementation a lot more trivial.

Finally, a suitable moniker. I suggest the 'Sussex VT MR2' as a homage.

[Electronpusher0]

I have built my own version of the Sussex, I call it the Felpham (it's where I live, a village in Sussex, where William Blake lived)

I attach the schematic and some pictures.

If there is any interest I will post some notes on the changes I made.

Peter

[Richardgr]

Looks very good, leveraging off the original design goals of Mike.

I would be interested to hear what you changed!

Could you consider a simple mechanism to select tube electrodes with patching cables, or do you consider the case and AVO-style mapping codes to be integral with this solution?

[Electronpusher0]

In my implementation of the Sussex the top panel with switches and sockets is connected to the tester proper by a flying lead and B9 plug and socket identical to that used on the 2 panel AVO tester. This means that I can unplug the top panel and plug in the left hand panel from my AVO 2 panel tester.
I can also use another panel I made with a matrix using 4mm plugs to select the electrodes instead of switches.
See here:
https://www.vintage-radio.net/forum/...45&postcount=3

I will write up my design notes on my changes and post them, it will not be until after the weekend though.

Peter

[Richardgr]

I am presuming that you insert the plugs into a matrix so that they are aligned with holes in a card, or follow a predetermined pattern? So similar to the AVO philosophy of having 10 digits keyed into thumbwheels so that a certain tube is patched in correctly. Fool-proof and repeatable if the instructions are followed.

If I understand your rationale correctly, I think your solution highlights a bit of a conundrum with valve testers: should they be designed to be used by people who understand the technology, or who just need to test a device?

The solution I have on my uTracer is to have two seperate 2mm banana plug sockets for each electrode option from the tester. [This is not my idea, I hasten to add, Ronald Dekker the creator of the device did it this way]. This means I can test a dual triode in one operation. The screen supply and measurements circuit is identical to the anode circuit. So to setup up an ECC81, for example, I use 2 plugs from the same section for Cathode and Grid, then Screen to one anode and Anode to the other. I can even decide if I want to use 12,6V across the heater, or parallel the 6.3V to each side.

The reality is that I am very rarely testing tubes in bulk. I might want to generate a set of curves for triode wired PCL84 pentode, check a couple of EL84s from an old tape recorder, prepare some characteristics graphs for tubes I sell (I think that is a real value adder), and so on. There are only 6 pairs of electrode/heater connectors to be patched to one of 9 pin sockets. It is quick, and flexible.

One weakness of the uTracer is the heater supply is derived from a pulsed 19v supply, so that the aggregated voltage is equal to the setting. This is a big advantage of the Sussex; if built as designed the tube is heated as it was designed to be, and people report that an external heater supply with a uTracer is a good idea.

A Sussex contains a lot of the components of the perfect benchtop supply for evaluating valve circuits. Regulated and variable screen and anode supplies, regulated and variable grid bias and switchable heater supplies. The transformer in the Sussex is not designed to supply the current demands of a benchtop supply, but it could be a key function of another implementation?

Just running stuff up a flagpole and seeing if anyone salutes!

[Station X]

Has anyone built a totally automated PC controlled valve tester?

I'm thinking you plug in a valve and enter its type from the keyboard. The tester then looks up the valve's pinout and characteristics. It sets the power and test circuit connections to the valve pins by operating relays arranged in a matrix. The voltages are set up and then the valve is tested with progress being shown on the screen. Then the results are presented in the form of a simple PAS/FAIL, test result figures and curves.

It would be impressive, but would never have the street cred of an AVO Tester.

[Electronpusher0]

The matrix I use is not card based, although it could be.
The vertical columns are the valve base pin numbers, the horizontal are the electrode number.
Its designed to use the AVO book, instead of setting a switch you insert a 4mm plug, this links between the row and column where inserted.
The rows and columns are strips of brass. Thin strips of plastic insulate the strips both between adjacent strips and between the row and column strips.

If there is interest I can post a dimensioned drawing, I will have to redraw my pencil sketch.

Regarding a computer controled valve tester take a look at this post
https://www.vintage-radio.net/forum/...6&postcount=47

Member "Electrical" has replaced the card matrix switches on a Mullard High Speed Valve Tester with relays with just that purpose in mind.

Peter

[Electronpusher0]

So as not to take this thrtead too far OT I will only post the wish list I used when creating my tester and will start a new thread to share the finer details of how I implemented it.

Peter

[Ed_Dinning]

Hi Folks, there was a design of an automated curve tracer in Electronic Engineering in the 40's. What a beast ! It used several uniselectors to do the curve tracing and some others for test selection. Output was displayed on a CRT

The transformer for the Sussex was designed to a size and price; hence it is not suitable for "burn in" testing of valves and only covers heater voltages from 2.5V to 12.6 . External supplies must be used for voltages above this level and also for battery valves at 1.4, 2, 4, 6 volts
I did a successful prototype transformer that was all singing/ dancing, but this was both large and expensive.

Ed