The Valve Interrogator

[Meat Pye]

Ever wanted to test your valve heater's continuity without fumbling with those test meter leads? It's a fiddly job even when you know that the heaters are on pin 4+5 on a B9G and those on a B7G are on 3+4. Want to test for cathode/heater shorts or leakage? Ever wondered if any of the other pins are connected to each other? Well now you can! With the valve interrogator you just pop the valve into the appropriate socket and it tells you all you need to know before doing an emissions test in a valve tester. All this can be yours for the price of a few evenings slaving over a hot soldering iron.

Well that would be my marketing spiel if I was selling this little unit but since the idea is gratis - would this be of any use to anyone with a large collection of valves? Let's start with how it all works. Referring to the schematic, an oscillator (555) generates the square waves (the frequency of which can be varied by the 1M pot) which clocks a 4017 decade divider. ONE of the 4017s outputs is always high and they in turn drive (via a diode and 10k resistor) an npn transistor which lights up an led. So far, so child's toy/light display. The junction of each diode/10k resistor is connected to the pins of a valve socket with the 10th o/p wired to a 4mm socket. When any connection is detected between any two (or more) pins on the socket then two (or more) leds will light up.

Let's take a valve for a test drive! Plugging an EF80 into the B9G socket the following led display was observed:-

1+3
2
3+1
4+5
5+4
6
7
8
9
10

When o/p1 goes high it switches on transistor 1 (Q1) and (via an internal EF80 connection) Q3, lighting up leds 1+3. o/p2 goes high and lights up led2 followed by o/p3 lighting up led1+3. When o/p4 goes high it lights up led4 and (via the heater) led5. Both these leds light up with equal brightness something that would not happen if the leds were driven directly from the 4017. The EF80 has no top cap and so the 10th o/p is not connected. Plugging an ECC81 in gave the following results:-

1
2
3+1
4+5+9
5+4+9
6
7
8
9+4+5
10

The ECC81 is a double triode with 2 separate heaters with a common return which makes 3 leds light up together.

Why use transistors when the leds could be connected directly to the 4017's o/ps? The 4017 can only supply 10 to 15mA on each o/p and this would not give a bright display with 2 or 3 leds lit simultaneously. Ultrabright leds would solve this but would not show up any high resistance leakage. Not having a leaky valve to hand I hooked up a 390k resistor between 2 pins and saw 1 led light up brightly while the other was dim but visible. Repeating with a 1M resistor the dim led was too dim to be seen. The use of transistors (and 5V regulated supply) also ensures that the voltage and current across the heater element is miniscule - essential for delicate battery valves, but not for mains ones. The diodes (1N4148) on the 4017 o/ps are there to prevent a high o/p driving into the low o/p should there be any connection between them via the heater or internal valve connections. Pin 15 (reset) is connected (via a switch) to ground for divide by 10 and to pin6 for divide by 7. This gives quicker operation for 7 pin (or less) valves.

Here's a couple of photos including an internal shot of the spaghetti junction wiring!

[Meat Pye]

I decided to upgrade the Interrogator as I wasn't entirely satisfied with some elements of my design. The first upgrade was to replace the green leds with hyperbright ones; 6 blue and 3 yellow for pins 3,4, and 5. These yellow leds would represent the heaters for 9 pin (4+5) and 7 pin (3+4).

The next to go was the 7805 regulator since the 555 and 4017 chips can run from 5V to 15V and, accordingly, the dropper resistors for the leds were increased to 1k which gives an led current of between 1.5 and 8.5mA depending on the voltage.

It was then decided to test my handful of EF80s on the Mk2 Interrogator, and the results were quite interesting. They all showed the same result except for 2 that showed a tenuous connection (1 led bright an another dim) between pin2 (g1) and pins 1+3 (k). A third one showed a tenuous connection between the heater (4+5) and the cathode (1+3). This would imply a cathode/heater short but strangely, when measured on a DMM the resistance was too high to register i.e. >2M ohm. Using resistors connected across 2 pins I estimated that the resistance of the 'short' was of the order of 10-20M. I could still see a connection with a resistance of more than 25M so it's clear that my circuit is detecting some kind of leakage fault.

As an afterthought, does anyone have any valves that are known leaky or shorted to heater so that I could test them? My thermionic collection is rather too small to give a good sample.

[Nymrod121]

What a clever piece of testgear that is - and the build quality is excellent

You mentioned a schematic in the OP; will this be released at some point?

In keeping with the thread title, perhaps a momentary flash of light could be used to encourage recalcitrant subjects ... an EB91 on full filament volts should suffice

Regards
Guy

[Meat Pye]

Thanks. You'll find the schematic in the the OP - it's the 3rd pic.

I designed the circuit for those of us who have a few valves and can't justify the expense of a valve tester. If you ever buy a job lot of valves then the ability to do a quick check of their heater continuity would save a lot of time instead of doing an ohmmeter test. Unknown valves might possibly be identified by knowing which pins are connected to each other.

The basic idea could also be expanded, with more connections, to test multicore wiring etc.