PX4 "no glow" mystery

[newlite4]

I have two pairs of Osram PX4. Two of these have the narrow envelope with a filament reading of around 1 ohm, the other two have the wide envelope and have a filament reading of around 1.6 ohm. With the expected filament volts of 4VAC at the socket, there is no discernable glow from any of these devices although the envelopes become hot to the touch and a low level of amplification is noted through the speaker. What's going on?
Neil

[ronbryan]

Is there any sign of a vacuum? I had this symptom (i.e heat, no glow and no emission) with an expensive EA53 diode detector in an HP Voltmeter and it was attributed to gas.

Ron

[Hartley118]

A white getter is a sure sign of a valve that has 'gone to air'.

Martin

[snowman_al]

... ''there is no discernible glow from any of these devices ''...
Isn't that normal for PX4s? I have had to take pictures in a darkened room in the past to convince myself they were on. Like the one attached.
Best measure the current at 4 volts and see if it meets the 1A spec. Less indicates one of the filaments is open circuit...

I've added a PX25 for comparison it draws 2 Amps.

[David Simpson]

Neil, is not a 1920's PX4 a "dull emitter" Triode ? I've some notes on ones I tested for a BVWS chum about 3 years or so ago. Will look them out.

Regards, David

[newlite4]

Thanks everyone. Al, you must be right, since this is a directly heated cathode then it is not necessary to heat up a separate cathode sleeve to promote electron emission, thanks for the pics. David, I wondered if they were dull emitters but thought that dull emitters were only meant for battery sets, it would be the same anomaly as stated above where the cathodes are directly heated. Thankfully the getter deposits are intact and silvery, if not I would be in the process of making a set of high-end audiophile bookends with them .
Overall, hoping that these PX4s are OK, sets do come in where the PX4 and PP3 valves have been robbed and it is great to have the opportunity to save them.
Neil

[Sideband]

The only directly heated output valve I have is a PM24M and the filament glow is only visible in subdued lighting as Snowman Al demonstrates in his pictures. The PX4 will be the same.

[Techman]

The glow won't be visible in bright light with these valves.

The reading of 1.6 ohms as against 1.0 ohms indicates that you've got a problem with the higher resistance reading ones, although they may still work well enough to be usable. However, if you took any of the readings while the valves were still warm then they will be meaningless - let the valves sit and cool right down before taking any filament resistance readings.

[Techman]

Just to add to the above - if all your valves test good then I'd suspect that the readings of 1.6 ohms were taken while the valves were still warm, so let them cool right down to cold and take them again. the second point is did you 'zero' your meter? Put your probes together and note the reading, was it 'zero'? If it wasn't (which it often isn't) then you will need to subtract this reading from any 'taken' filament resistance readings to get a sensible figure to work with.

[newlite4]

Thanks Techman, I am wondering if the larger envelopes that measure 1.6 ohm have a slightly different method of construction to the slimmer 1 ohm types. Measurements were taken when cold using a zeroed Fluke DMM.
Neil

[Techman]

Originally Posted by newlite4

I am wondering if the larger envelopes that measure 1.6 ohm have a slightly different method of construction to the slimmer 1 ohm types.

I wouldn't have thought so. I've usually found that they all measure round about an ohm or so cold, give or take. A higher resistance usually indicates either an actual filament string problem or perhaps much more likely a high resistance solder connection within one or both of the filament pins on the valve base. 1.6 ohms isn't awfully high, but higher than it really should be, I would say.

[Radio Wrangler]

These valves have multiple lengths of filament, all connected in parallel inside the bottle.

If one or more lengths fail, either through breaking or failure of a weld, then you see higher resistance from there being fewer strands in parallel.

These things act like several valves in parallel, and they tend to fail one sub-valve at a time if filaments break.

If raising a mortgage to buy one of these valves, it's a bit of a game ensuring the one you've bought is firing on all cylinders, and remains that way after shipping. Add in the possibility bad welds that make connection only when hot or only when cold and recriminations can fly.

Valves aren't as easy as people think. Transistors aren't quite as hard as people think.

David

[Techman]

Yes, if you've received these valves through the post, then it may well be that a filament string is broken in each one. It's easy enough to see in subdued light if all four pairs are glowing. Also it's easy enough to apply 4 volts and check that the current being passed is as the specification states. I can tell you from experience that these valves don't travel well through the post. The last one I received had busted filament strings due to poor packing. It's a shame to think that a valve like this survives for many decades in perfect condition, only to be ruined by having it rattling around in its box through the post. You can get away with it with most valve types, but with these having long, unsupported filament strings, they don't travel well when posted - another more standard type of valve in the same package survived perfectly, although that one wasn't rattling around loose in its box like the PX4...it's just a bit of bad luck at the end of the day and how people pack things and how the post treats them on route. The worst thing is when the broken filament is dangling about and shorts onto another electrode, that's fun - not...as I found out! All that aside, high resistance in the valve base pins is a very common problem, and one that I've had to deal with in the past with these valve types due to years of operating at relatively high heater current - you don't get the same problem with valves that operate with lower current heaters, so low voltage and high current heaters are bad news...high(er) voltage and low(er) current heaters are much better in this respect.

It's good to see these valves being used as they were intended in vintage radios and grams, and not for audiophoolery! Although that's not to say that the don't look quite 'pretty' when displayed in an open 'on-view' amplifier - I suspect that the audiophool finds it rather disappointing that for something with such a large glass bulb, that the glow is so dim as to only be seen in a darkened room - I find that part quite funny!

[ms660]

In the one in the link below all the filament lengths aren't connected in parallel, there are two circuits in parallel, each circuit consisting of two inverted V's connected in series, you can see the series connecting strap for the two outer inverted V's and the filament supply wires from the pinch:

http://www.r-type.org/exhib/aba0002.htm

Lawrence.

[GMB]

When I was a child I ran into a batch of 6BW6 valves that behaved very oddly.
There was no visible glow from the heater, but the envelope got very hot.
The getter was silver - but didn't quite look as good as other valves.
There was some conduction but not much.

I am now sure that these valves all had defective pin seals and so a small quantity of gas had got in. It doesn't take much gas to cool the heater so much that it is not seen to glow - but the heat is still being generated.

It occurs to me that it should be possible to access this effect my measuring the cold and hot resistance of the filament - from which I think it should be possible to calculate the actual temperature of the filament when hot. If the filament power is correct but the temperature is not then this is a sure sign of gas.

[newlite4]

These PX4s came from a house clearance that I did fifteen years ago. When I went into the attic I encountered many thousands of valves and I mused that there may be some PX4s here. I picked out six PX4s, two of them were BNIB and in mint unused condition so I arranged with the executor to sell these at auction. All proceeds (less auction commission) I donated to the local hospice, they fetched £320! The remaining four having been in used condition I felt better about keeping for repairs to plundered sets. Quite a good find really as I would never normally buy valves like these.
Lawrence, I can see looking at the valve museum page that there are several opportunities for filament connection failure due to thermal cycling within the envelope, this does heighten the need for filament resistance checks if anyone is in the process of buying these.
I am a little cautious about testing the filaments on an external DC supply in case the supply throws a wobbly and cooks the filaments, unlikely I know but I would not want to take the chance. If they were EF80s I would not be too bothered if the control pot on the PSU decided to go high.
My main concern was the lack of a glow but all of you have put my mind at ease now and I can at least use these valves as viable spares.
Neil

[ms660]

Out of interest the PX4 was made with two different filament currents and max anode voltage specs, the early ones were 0.6 A @ 4V, the later ones were 1 A @ 4V, the later ones having a higher max anode voltage rating and a greater anode power dissipation rating.

Lawrence.

the PX4 was made with two different filament currents and max anode voltage specs,

That explains the OPs different resistance readings, so it looks like all of them are good.

[newlite4]

This latest information from Lawrence is beginning to make sense of the discrepancy between heater resistances. I notice that the repro PX4s from EML run at 1.6 amp heater current which ties up with the later production spec. Note that the larger envelopes have the lower heater current. The slimmer envelopes with the 1 amp heaters would appear to be the later version.
Neil

[PJL]

You can normally tell if a filament is broken as one or more of the filament tension springs at the top will be released. The Osram datasheet for the balloon version shows 1A so I guess the 4V 0.6A PX4 was very short lived so unlikely to be yours.