About Valve testing

[Mike Watterson]

The main aim is so that people can understand what an AVO 163 or Mullard HST or any of the modern curve tracers need to do, as as our fav Penguin Jeffery remarked you only need some PSUs.

BUT it can't be emphasised enough that except maybe at VHF/UHF, the handiest valve tester is the equipment it's for. So I may do some posts on the different valve functions in a radio, amp, record-player, transmitter and tape-recorder, indicating what can be measured to decide if the valve needs replaced.

A valve tester or curve tracer is useful if designing from scratch, especially if the valve is plentiful but lacking data, or to help guess what a random valve with missing markings might be.

I used them a lot by visiting a TV shop in 1960s because I was ignorant and didn't know how to tell measuring the valve in the radio. Historically used more for production, design or checking new stock than repairing sets, because people knew how to repair and up to the mid 1960s had stocks of new valves.

People also write now about using them to match valves for amps, either push-pull or stereo etc. That wasn't done historically. It's a bad design that needs parts selected on test. Some designs relied on negative feedback. Others, especially radio transmitters with paralleled valves would have bias adjustable for each one. On some sets you took off one top cap alternately and set the quiescent cathode current. Other models you measured cathode resistors or opened test links to the output transformer.

An AVO VCM163 is nice to have, but it takes a lot space. May need a lot of specialist restoration and really not used much in repair or restoration. I've not taken my uTracer out of its case since I last got a mysterious rare Russian Rod Pentode a few years ago!

The ad-hoc PSU with leakage and PIV test supplies is a handy general piece of kit, so I'm building one anyway. It certainly will be better for valve testing than some old cheap testers or the specialist Mullard HST (which is an ingenious piece of kit for a valve factory!).

[Mike Watterson]

Anyone else remember TVs with valve UHF tuners? Maybe PC86 and PC88

http://www.r-type.org/exhib/aaa0865.htm
http://www.r-type.org/exhib/aaa1566.htm

If UHF stopped, you simply replaced those! Maybe less than 4 years life?

[ppppenguin]

Not just short life but insensitive and noisy. No wonder BBC2 had a bad reputation when it started in 1964.

They were some of the very few P series valves with a heater voltage of less than 6.3V.

I've loads of NOS around. I wonder if the audiophools might like them

[BusyBee]

Thnk you Mike.

It's useful to get all this information collected in one place.

Tracy

[Mike Watterson]

So the minimum to test valves is

1) The equipment the valve is used in
2) A DVM/DMM. The cheapest ones might have a 750V DC scale, but often flash over above 300V. They also often have only 1 M Ohm input on all voltage ranges. The 50V range on an old 20 K Ohm per volt analogue meter is as good and the higher ranges are better. Get two that do at worst 200uA DC current (so you can see 1 uA) and 20 M Ohm on voltage ranges.
3) A leakage tester. Not a Megger or Electrician's tester, but something limited to 200 uA on a short and at least 300V into 20 M Ohms. A flash from a single use camera works. Remove the xenon tube and the trigger button. Replace the 300V electrolytic with 1uF 400V and put at least 2 x 1M in series on the output. A series neon with a 200V 0.1 polyfoil cap across it will give visual indication.

Optional is a signal generator. It's handy if it has AM, but that can be added. It also needs an attenuator, which can be added, to give down to the level of a very weak station. The harmonics of a 30MHz max generator may work up to 90 MHz or 120 MHz for VHF. You don't need FM on a generator to test an FM tuner!

A 20 MHz scope is needed for TV and maybe taperecorders, but not for amps, recordplayers, radios or transmitters. A scope and maybe a spectrum analyser are needed for designing stuff.

Also optional, particularly for valve TVs, SMPSUs and series heater chain mains radios is a tungsten or halogen lamp in series. I've only ever needed it for one radio.

Obviously you need a separate 40W or more soldering iron, with 3mm flat or 3mm chisel tip and 60:40 solder. You might want separate lead free solder and iron for "modern" stuff. A spring loaded solder sucker for tags and fluxed copper braid for PCBs.

Don't plug in the equipment "just to see". The first thing is to check the valve on the output stage. There are parts to replace and things to check before powering on to check the output valve.

Next: Testing the output valve using the equipment instead of a valve tester.

[Mike Watterson]

Testing the output valve using the equipment instead of a valve tester

For this we need to decide what kind of power supply is used and what kind of valve and output stage and what kind of bias.

1) Supply.
Series or parallel heaters / filaments, or even a pair of paralleled series chains.
Isolated transformer for heaters and HT or partial transformer/autotransformer/ballasts only (Live chassis). WARNING some USA sets have metalwork connected via 0.1 uF paper cap to one side of mains. MUST be replaced by safety 10nF rated 500V to 1kV. Some sets have the ballast resistor as part of line cord to convert from 120V to 240V. Needs replaced.
Valve or Metal/Selenium or Germanium or Silicon rectifier(s). A rectifier valve can have its own heater supply, or be part of the series chain, or run direct off USA 117V nominal mains, or off main heater supply.
Is there a Barretter? A big glass tube that looks a little like a festoon light bulb. It's a kind of Thermistor to regulate the mains voltage. Is there a regular series thermistor? It limits inrush current as it may damage panel lamps or some valves. The inrush otherwise can be x2.5!
Is it battery only or mains and battery using 1.4V or 2V battery tubes?
Is it battery using selected mains 6.3V or 12.6V valves direct off car battery? 10.8V to 14V. Then the audio will be transistors.
Is it battery using a normal HT on valves by a vibrator pack from 6V or 12V. Some of these use an extra pair of contacts to "rectify" the transformer secondary. May be 80Hz to 200Hz. Some can also use mains.

This can all be deduced without a schematic. A taperecorder, transmitter or TV needs a schematic. Most radios and almost all record players can be figured out without one.

2) Audio output Valve.
a) Direct Triode with 2, 4 or 6V filament. Mostly before 1931, but a few sets used them up till 1941. There is often a tapped grid bias pack.
b) Direct Pentode with 1.4V or 2V filament. The 4V was rare. Very early sets used a tapped grid bias pack. Some used a Mallory mercury cell. This is what the button format and mercury battery was invented for. Company named after Mallory and became Duracell. Later sets used a resistor in the -HT connection to 0V. A very few sets had the grid bias in the HT pack. Some 1950s sets use filtered connection to the local oscillator grid for a -9V grid bias. A 7.5V to 12V battery (or battery mains) set using 1.4V valves will put the audio output valve(s) at the +7.V, 9V or 12V battery connection so that the grid leak resistor to 0V is neagtive. The first audio preamp is usually at the 0V end of battery (and mains supply for least hum) and then the Mixer/oscillator is next.
c) Pentode or Beam Tetrode with indirect heater. Grid Bias is usually derived by the voltage on the cathode series resistor. On push pull there may be a negative supply and always on transmitters. Occasionally this cathode resistor has a 1.35V tap for a DM70/DM71 magic eye filament.

3) Kind of output
a) Very old sets, with a triode or pentode, used a direct 2,000 or 4,000 load such as a horn, moving iron cone speaker or 2K + 2K headphones. This would burn out when the grid bias battery failed after about a a year, or if the tap came off.
b) Later a transformer with a 2 Ohm to 16 Ohm moving coil and permanent magnet was used. The output transformer would burn out when the grid bias battery failed after about a a year, or if the tap came off.
c) Later still a capacitor rather than a transformer was used to drive the output from the prior anode. This would go leaky in 5 to 10 years and the output transformer would burn out.
d) Mains radios are common from 1934. Early models ran the HT through a series winding on the speaker instead of a magnet. This might include a humbucking winding. Now the leaky cap can result in the output transformer or the field winding burning out.
e) Due to design of transformer a capacitor called a tone corrector (a snubber) is added to the output anode. it can go to HT or OV. When it fails the HT rectifier or main HT transformer burns out. The failing grid to proir anode cap may have already damaged the output valve and output transformer.
f) The output may use a centre tap transformer and two identical valves. Sometimes these are in the same envelope such as KLL32 and ELL8x
g) A few mains / battery radios use all battery valves and switch from a battery Pentode to regular mains radio Pentode on mains. These use a custom audio transformer, inevitably burnt out.
h) There may be negative feedback from anode, or loudspeaker or an extra winding to grid or volume control. Check the leakage of capacitors.


Proceeding
Some people talk about a "THAT" capacitor. There are many problematic capacitors. Leaky paper caps on the screen grids (g2) will kill gain. Leaky paper caps from mains to metal work may kill users. The "Tone Correction" cap is a problem.

The power supply needs to be got working first, but after replacing / reforming caps.

Unplug all the valves. Do not connect to supply. Connect LV PSU to any Electrolytic on PSU HT and adjust current limit to a few mA and turn up voltage to 30V. Leave at least all day. If the current doesn't drop a cap is leaky and not reforming. If the voltage vanishes the instant the PSU disconnected, then the cap has dried out.
Replace any metal/selenium rectifiers with 1N4007 and 47 Ohm to 100 Ohm 5W resistor. Valves in a series chain need replaced if filament/heater open circuit.
Check resistance of ballasts/transformer winding and leakage of transformer. Check primary of audio transformer.
Disconnect + end of cathode capacitor and reform it at 25V and current limit. Replace if needed.
If parallel or no valve rectifier, you can power up with all the valve out except rectifier once caps, leakage etc all checked. Check HT. Measure g1, and cathode of o/p valve socket that they are 0V. Measure g2 and anode that they are close to HT volts.
Add output valve and measure cathode voltage. Switch off
Calculate current from V/R where R is cathode resistor. if it's very low and all voltages are OK, then the valve is worn out. If it's very high, recheck the grid circuit.

Basically any paper capacitor (most old ones from 1000pF to 1.5uF) is going to need replaced unless it's a hermetic sealed type in a metal can (not the TCC as the end bungs leak). Cathode and Discriminator electrolytics tend to dry out (low volume and no FM) and HT ones often "reform". Only USA Silver Mica are prone to failure. Polystyrene are OK unless you melt them soldering. Mustard and rainbow sweet caps are usually plastic & foil or metalised foil so usually OK, though I've had faulty poly caps in a cassette deck. The two colour plastic body low voltage Electrolytics are usually scrap.

If the set is an isolated chassis (decent transformer) you can now get hum touching the o/p valve grid via a 10nF cap and 100K resistor for safety.

If in doubt, don't turn it on and ask!

[Mike Watterson]

Note that for the battery radios
LT on a PSU is less than number on a dry battery
1.5V -> 1.35V
7.5V -> 6.8V
9V -> 8.2V

Or adjust for 1.35V on the f+ of the lowest valve in a series chain. DAF91, DAF96, 1U5 etc.
For 2V valves use 2V

For HT you can use
22.5 -> 18
45V -> 36
67 -> 50
90 -> 80 (but 60, ie. 2 x LV PSUs should run!)
108 to 144V: use 90V.

Put a 20mA fuse in HT lead.

The batteries are higher voltage when fresh and endpoint is about 1V per cell with HT and about 1.1V per cell on LT.

Many signal mains valve (6.3V) will run at 12 to 30V HT. May need +1V gb for ECH81 (=ECH83). Valves such asT ECC82, EBF89 and many others. Only data sheets for Car Radio applications (10.8V to 14V supply) cover the operation. Max frequency for oscillators is reduced.

[Mike Watterson]

Note that real Tetrodes are also known as SG valves (Screen Grid) and were only used for RF or mixers or oscillators. The Beam Tetrode is unrelated and is really like a Pentode.

[Mike Watterson]

Tolerances when testing a valve in the equipment

Especially any Pentode.

Because of how they were made the resistors more than 100 k Ohms tend to go high and less than 1000 Ohms go low. Often the output valve screen grid (g2) connects to an HT or the output transformer, so not an issue. Many resistors were 20% and grid leaks (g1 to 0V) can be 40K to 5 M Ohm on the output valve and it won't make much difference. The cathode resistor should though be within 10% and nearest 5% part is fine for replacement, such as 220 Ohms for a part marked 200 on the schematic. Screen grid to HT resistors going high will reduce gain, but fine if within 30% of original value.

AGC capacitors are not critical as long as working.

Electrolytics can be +90% and -40% and it won't matter as long as they are neither leaky (oxide broken down or high ESR (Effective Series Resistance), usually due to drying out. Same voltage or higher for replacements and even twice the value is fine except on two places. The rectifier valve has a maximum direct connected capacitance on data sheet (peak current) and the replacement FM discriminator or similar should be nearest next highest value, thus 4.7uF for a 4uF part. It is usually -ve to the diodes (semiconductor or 2 in an EABC80) and +ve to chassis.

Nearest +20% is fine for replacing any paper caps. I usually replace the "tone correction" with a 1KV ceramic of nearest value. Others are replaced by suitable 250V or 400V metpoly/polyester caps. I may use PCB types with tied on and soldered extension wires and sleeves rather than the round yellow ones. Many paper caps in battery sets can be replaced by 50V or 100V ceramic parts. I might snip out most of the "live" end of a paper cap and leave it in place.

If the g2 voltage is near HT with no valve and drops alarmingly and cathode voltage is madly high, did you really replace cap feeding g1? If the g2 drops a lot but cathode voltage is OK, suspect an HT line resistor gone high, or the g2 to HT resistor if there is one.

Don't rely on an old tester bridge from 1960s or earlier, but a DMM or a modern component tester (discharge caps twice before testing). Unplug mains before testing or changing parts.

If you have no leakage tester reading socket voltages with HT live and no valves fitted will reveal leaky caps. A DMM even on 20 M Ohm range can't see capacitor leakage. You need uA meter and 100s of volts, or the safe flashgun based tester.

[ppppenguin]

A warning that modern capacitor testers will usually give a higher than expected capacitance for leaky parts. This is because they measure rate of charge which will be lower than expected if the charge is trying to leak away.