Valve warm up times

[trickie_dickie]

Is there a post anywhere that gives typical warm up times of heaters in various valves used in radios & amps from times gone by?

[Synchrodyne]

For the American series-string TV valves with controlled heater warm-up time, the latter parameter was usually quoted as 11 seconds typical.

The 600 mA series-string types were first introduced in 1954, and the 300 and 450 mA types following in 1956. As many of the 300 mA types were reworked versions of existing valves (e.g. 6AU6A was the controlled heater warm-up version of the 6AU6), one may suppose that the originals had somewhat variable warm-up times.

In 1961 there was also released a set of controlled heater warm-up time derivatives of the American 100 mA series-string radio valves that had been introduced in 1959. These had a 20-second nominal warm-up time.


Cheers,

[Synchrodyne]

Mullard introduced controlled warm-up time series-string TV valves in 1963, as shown in the attached advertisement. Whilst the means by which this was achieved were discussed, there was unfortunately no mention of actual warm-up times.


Cheers,

[broadgage]

The actual warm up time would also be dependant on the type of heater circuit.
A constant voltage supply such as a heater transformer or a battery will give a quicker warm up than a resistance dropper from a high voltage mains supply.

A cold valve heater connected to a constant voltage supply will draw extra current initially, and this extra energy will rapidly heat the heater.

A mains dropper in series with the heaters approximates to a constant current supply and results in much slower warming. The cold heater energised from a constant current supply will initially drop a lower than nominal voltage, and this lower dissipation will slow down the warming up.

Directly heated valves will normally warm up quicker, since ONLY a filament has to reach operating temperature rather than waiting for a filament to heat a cathode.
Some directly heated valve circuits were rather slow to heat up though for another reason. If the filaments of several directly heated valves are connected in series, then those at one end of the chain are subjected to the HT current of the earlier valves AND to the design filament current. To obviate over-running the filaments thereby they were shunted with a resistor to bring the filament current down to the nominal value.
When such as set is first turned on, there will be no HT current and warming up thus delayed by the initially reduced filament current.
Valve battery/mains portables existed that used 5 directly heated filaments in series, from a 5 cell battery of 7.5 volts nominal. These were slow to warm up on battery power, due the reduced filament current, and on mains supply slower still due to the virtually constant current filament supply via a dropper resistance.

The actual mains voltage would influence warm up times. A set designed for 240 volts will warm up quicker on 250 volts than on 220 volts, both voltages are within the expected range of a nominal 240 volt supply.

Vehicle mounted equipment would also warm up quicker with the engine running (14 volts actual) than with the engine stopped (12 volts actual)

Anecdotal reports suggested that cold weather extended valve warm up times, I doubt that variations in UK room temperature from a chilly 5 degrees to a warm 25 degrees actually had noticeable influence. More likely is that low temperatures reduced the effectiveness of batteries, or for mains sets, tended to result in low line voltage due to extra load.
I suppose that extreme cold in polar regions or aircraft might have had a noticeable effect.

Some equipment over-ran valve heaters to reduce warm up times, American juke boxes did this. After a few seconds the heater voltage was reduced to normal.

[paulsherwin]

Heater warm up times vary a lot even within the same valve type, depending on how the valve is constructed. The best example I can think of offhand is the UL41, which seems to have been redesigned sometime in the 60s to reduce heater/cathode breakdown and consequent leaks. Late production UL41s take a *lot* longer to warm up in a typical AC/DC radio.

[GrimJosef]

Yes. M-OV KT66s (GEC, Osram, Marconi) warm up more quickly, at least in terms of cathode emission, than the modern ones from the Chinese Shuguang plant. I wondered whether the main difference was in the cathode coating formulation (no longer any thorium ?).

Cheers,

GJ

[G6Tanuki]

As has been noted, it depends on the circuit: I remember some 1950s TVs that had series-heaters and included a couple of EB91/6AL5 double-diodes in the string: the filaments of these would light up like a lamp-bulb for a few seconds immediately on switch-on! I guess they'd have been up to full-emission in a second or so...

One juke-box design used 'accelerated warm-up' of its push-pull 6L6 output stage by feeding the cold heaters from a separate 9V winding on the mains transformer; once they started drawing cathode-current the heaters were switched back to 6.3V for normal operation. The idea being to get the output-stage ready to play in less time than the juke-box-selector took to find and drop the needle onto the relevant 45, because paying punters would have been annoyed not to hear all the music they had paid-for.

Some TV sets with series-heaters did delay things quite a bit by way of Brimistors or similar in the heater-chain; it was not unusual for tellies to take a minute from switch-on before displaying a picture; one GEC model I remembered was particularly noticeable because the LOPT would sometimes 'squeal' at a low-ish frequency during the warmup process before eventually achieving lock once the signal-frequency stages started to work and provide sync.

[Peter.N.]

It can also depend on the condition of the of the valve, power valves particularly can take much longer to give acceptable audio quality if the emission is low.

Peter

There is an added delay in a valve telly, once the line output stage is running that has to warm up the EHT rectifier, add ten seconds or so.

[PJL]

Warm up time is for series heater chains and will refer to the heater reaching a stable resistance rather than the time it takes to heat the cathode and conduction to stabalise.

[Goldieoldie]

Valves used in portables running a 1.4 v heater are almost instantaneous.How is this achieved with such a low voltage / current ?

[G8HQP Dave]

Only a very thin filament to heat.

[Synchrodyne]

Originally Posted by PJL

Warm up time is for series heater chains and will refer to the heater reaching a stable resistance rather than the time it takes to heat the cathode and conduction to stabalise.

The book “RCA Electron Tube Design, 1962” provided “chapter and verse” on the subject. I have extracted the pertinent text, attached below.


Cheers,

[trickie_dickie]

This is great. I think G6Tanuki has given me an answer. I'm thinking of using an EB91 to rectify a Bias supply up to -60 volts to bias the output valves in my 5-20 amp but it needs to be set before the o/p valves warm up.

I think it may be worth testing the valves starting with the GZ34, measure the time it takes to charge up the res cap 8mF, with no other valves fitted.
Then measure the time the EB91 takes to charge up the Bias cap 10mF.
Then I could measure the time it takes for the o/p valves to reach nominal current flow.

Many thanks everyone.

[PJL]

Or is this a case of where use of a silicon rectifier is justified?

[GrimJosef]

Originally Posted by trickie_dickie

... I'm thinking of using an EB91 to rectify a Bias supply up to -60 volts to bias the output valves in my 5-20 amp but it needs to be set before the o/p valves warm up ...

It's worth thinking about other power control sequences too. I have seen more than one valve amp with problems caused when someone has briefly switched the mains supply to the amp off and then back on again. Things which were assumed to take a long time to warm up can still be hot after a brief switch-off and can start conducting surprisingly quickly. One way round this is to use a time-delayed 'soft-start' relay on the HT supply, or on the main rectifier's heater supply.

Cheers,

GJ

[bikerhifinut]

Originally Posted by trickie_dickie

This is great. I think G6Tanuki has given me an answer. I'm thinking of using an EB91 to rectify a Bias supply up to -60 volts to bias the output valves in my 5-20 amp but it needs to be set before the o/p valves warm up.

Then it's no longer a 5-20, you've substantially altered the output stage and there may well be more to consider than that.

[Radio Wrangler]

Are you putting fixed bias on the EL34s in place of the cathode resistor's self bias?

If so, be aware that the cathode resistors create degeneration at the frequencies below where the cathode capacitors shunt them. This gives a greater degree of stability of bias conditions than preset bias can achieve?

It's why some guitar amps with bias pots also sport individual current meters for each valve's current.

Powering up with no cathode resistors and no bias would be a very bad thing indeed. Silicon would be a very good idea for the bias rectifier. If your EB91 failed, it would take the EL34s and maybe some transformers with it. I also seem to remember EB91s having a section of heater wire that flashed up very bright during power-up.

David

[G8HQP Dave]

I would not want to trust a set of output valves to one rectifier heating up quicker than another.

[kalee20]

Absolutely! I'd stick cathode resistors in (470Ω) shunted by normally-open relay contacts, the relay being energised by the bias supply. If the output valves warm up before the bias supply, the output valves get cathode bias till the fixed bias is established.

Might need hunting for a sensitive relay, if using EB91's. But should be do-able!