This has been commented on , on and off over the years.
Cathodes get stripped if they have HT on them before the heater comes up to the right temperature etc. This reduces tube life and`performance suffers.
Some tube circuits are put into standby by having a partial heater voltage on them ( 60 % of nominal voltage ? ) and no HT voltage on the anodes.
Is this a safe condition for the heaters ? Power up will bring up the heater voltage to nominal value and I guess HT comes up instantly ( due to a relay or other switching schemes ).
What would be considered the best way to put the circuit in standy without damaging the heater or the cathode coating and ensuring normal tube life and performance.
Cathodes get stripped if they have HT on them before the heater comes up to the right temperature etc. This reduces tube life and`performance suffers.
Some tube circuits are put into standby by having a partial heater voltage on them ( 60 % of nominal voltage ? ) and no HT voltage on the anodes.
Is this a safe condition for the heaters ? Power up will bring up the heater voltage to nominal value and I guess HT comes up instantly ( due to a relay or other switching schemes ).
What would be considered the best way to put the circuit in standy without damaging the heater or the cathode coating and ensuring normal tube life and performance.
Remember the "insta-on" tube TV's of the 60's and 70's? When off, they put a diode in series so effectively the tubes were seeing about 50% of the AC peak voltage through their heaters.
Didin't seem to hurt them - the last "insta-on" TV used by my family lived 34 years and only one tube changed (the video out, IIRC) in that time
Cheers!
Didin't seem to hurt them - the last "insta-on" TV used by my family lived 34 years and only one tube changed (the video out, IIRC) in that time
Cheers!
Really? I've read info that describes the idea of cathode stripping as a bit of an old wives tale. I can accept out of bias conditions as a possible failure mode due to bias circuits not being up to speed when HT is applied and the tube going into a run-away failure, but cathode stripping - has anyone actually seen this in real life?
"By the 1950s, lifetime was a predominant concern of tube designers and industrial users, and if switch-on surges would significantly reduce lifetime then this would be reflected in the literature of the time, which it is not. There could conceivably be some benefit with directly-heated filamentary tubes, whose filaments are more mechanically fragile. The high-end audio world is full of mystical beliefs, and very probably this is one of them." http://www.john-a-harper.com/FilamentHeating/index.html
"........or power off? ....."
Power on failure is related to failure due to inrush current. The switch on current is much larger than the operating current as the heater resistance is very low due to the low starting temperature. Additionally the heating is worst at the weakest points on the heater coil and that eventually breaks down. Just like filament lamps. Switch off doesn't have such problems.
However cyclic heating and cooling of the heater would also cause metal fatigue and cause eventual breakdown. Both these must be why Tommy Flowers wanted to keep the tube circuits permanently on.
Maybe strong physical movements when the heater is still hot ( after switch off ) could also cause failure ?
But then the focus of this thread is on what is the best method to ensure long tube life. For standby status , reduced heater voltage and no HT is good enough ....so far ! On " power up " I guess the heater voltage can be switched up to full instantly and let the HT rise a bit slower .
Power on failure is related to failure due to inrush current. The switch on current is much larger than the operating current as the heater resistance is very low due to the low starting temperature. Additionally the heating is worst at the weakest points on the heater coil and that eventually breaks down. Just like filament lamps. Switch off doesn't have such problems.
However cyclic heating and cooling of the heater would also cause metal fatigue and cause eventual breakdown. Both these must be why Tommy Flowers wanted to keep the tube circuits permanently on.
Maybe strong physical movements when the heater is still hot ( after switch off ) could also cause failure ?
But then the focus of this thread is on what is the best method to ensure long tube life. For standby status , reduced heater voltage and no HT is good enough ....so far ! On " power up " I guess the heater voltage can be switched up to full instantly and let the HT rise a bit slower .
On the other other hand [sic], if you leave heaters on for long periods with no HT you run the risk of "cathode poisoning" where non-radiating layers develop in the cathode coating. They had to develop special valves for computers (where some valves could be idle for long periods) to overcome this problem. There are no free lunches : what a horrible world!
........"cathode poisoning"............
Right, that jogs my deteriorating memory! Thanks.
So applying a low HT supply will keep the tube at a strange operating point. It will keep dissipation down but is this odd operating point OK ? Does it overcome "cathode poisoning" ?
Or do we say that heater on ( partial voltage ) and no HT is out of the question if it's going to be that way for several hours ?
Right, that jogs my deteriorating memory! Thanks.
So applying a low HT supply will keep the tube at a strange operating point. It will keep dissipation down but is this odd operating point OK ? Does it overcome "cathode poisoning" ?
Or do we say that heater on ( partial voltage ) and no HT is out of the question if it's going to be that way for several hours ?
I'm afraid you've reached the limits of my knowledge on this subject. I don't know how long the valves have to be idle for the poisoning to start happening. It's probably more than a few hours.
Personally I use valve rectifiers in the HT supply and switch amplifiers off when I'm not using them. The valves can be used after a solid-state supply as slow switch-on devices.
Personally I use valve rectifiers in the HT supply and switch amplifiers off when I'm not using them. The valves can be used after a solid-state supply as slow switch-on devices.
Slow 'switch on' in solid state supplies is not too hard.
I wouldn't use a tube rectifier for that reason. But it does reduce switching noise spectrum compared to solid state switching. I guess keeping valve systems on 'standby' is NOT a good thing to do.
That's it. I've have to leave early morning and I must now go to bed now. Will read posts after a couple of days !
Cheers.
I wouldn't use a tube rectifier for that reason. But it does reduce switching noise spectrum compared to solid state switching. I guess keeping valve systems on 'standby' is NOT a good thing to do.
That's it. I've have to leave early morning and I must now go to bed now. Will read posts after a couple of days !
Cheers.
I think the idea comes from ham radio amps, not audio HiFi amps. The radio amps sometimes run with B+ well into the kilovolts range.
It is a problem with only certain noted types like the 6080, 6336/A/B/C, and 6AS7G as well as in very high voltage broadcast and ham radio applications.. It is not normally an issue with tubes typically found in most of our tube hifi gear.
Cathode interface (poisoning) is a problem that can occur when plate current is extremely low or non-existent and the filaments are heated to normal temperatures. Tubes designed specifically for computer use are designed for extended periods of operation fully cut off, and usually don't suffer from this problem. Certain tubes used in Tektronix tube scopes (in the distributed vertical amplifier) were also low interface types.
CCS based heating for dhts or even idhts (in series only) eliminates high inrush currents and reduces thermal stress and hotspots. Ramping up the filament voltage in parallel filament strings is probably helpful.
In 30 some odd years of playing with tubes I have had two filaments burn out at power on, interestingly both were Mullard 12AX7A (both purchased new) and few other incidents that stood out.
Most of my tube failures have been unrelated to filament issues.
I wouldn't spend much time worrying about these issues, and more importantly would not needlessly complicate a design with circuitry to protect against something that does not appear to be an issue in most cases.
Cathode interface (poisoning) is a problem that can occur when plate current is extremely low or non-existent and the filaments are heated to normal temperatures. Tubes designed specifically for computer use are designed for extended periods of operation fully cut off, and usually don't suffer from this problem. Certain tubes used in Tektronix tube scopes (in the distributed vertical amplifier) were also low interface types.
CCS based heating for dhts or even idhts (in series only) eliminates high inrush currents and reduces thermal stress and hotspots. Ramping up the filament voltage in parallel filament strings is probably helpful.
In 30 some odd years of playing with tubes I have had two filaments burn out at power on, interestingly both were Mullard 12AX7A (both purchased new) and few other incidents that stood out.
Most of my tube failures have been unrelated to filament issues.
I wouldn't spend much time worrying about these issues, and more importantly would not needlessly complicate a design with circuitry to protect against something that does not appear to be an issue in most cases.
Cathode-stripping..
I also believe that cathode-stripping is a risk only with Very high voltage Ham transmitting types....
All 50's 60's and what 70's radios and TV's here in europe had valves generally had the HT applied before the heaters of the rest of the set had got to temperature, as most radios etc had directly-heated recs etc and the TV's tended to use a silicon diode rectifier--Very rarely did I see a genuine 'low-emission' type valve failure....
Over the years, Ive seen quite a number of O/C Heaters, and nearly without Exception, They were Mullard Valves and CRT's and even in sets fitted with a thermistor heater current limiter designed to apply a slow warm-up....
--Of different types, some large pentodes, some small signal triodes, But Always bloody Mullard!
(Why on earth are Mullard valves held in such high regard, when they cant even get the heater right!--Good marketing I guess....)
I have around 12 of EF91 types, of which I have 8 Mullard CV4014/M8083 'special-quality' types, four of the 'blue-glass' type and four clear, so are of a different batch range...
Of the 8 Mullard, 5 show that sudden bright heating when first powered, where the heater is welded to the lead-out wires-One is particularly bad, and glows nearly white for a half-second or so. Heater supply checked, and is 5.9-6.0V
All these valves are NOS...
The other four are I believe STC and are CV138,(EF91 equivalent) and are quite well used, dont show this effect at all.
--And, That is just the EF91 types Ive had issues with!
take a look at your examples of Mullard valves and see how many are apparently poorly assembled, where the structure is not straight or the Micas are cracked etc, and even the bulb itself mis-shapen......
Nah, I Dont like or trust Mullard....
I also believe that cathode-stripping is a risk only with Very high voltage Ham transmitting types....
All 50's 60's and what 70's radios and TV's here in europe had valves generally had the HT applied before the heaters of the rest of the set had got to temperature, as most radios etc had directly-heated recs etc and the TV's tended to use a silicon diode rectifier--Very rarely did I see a genuine 'low-emission' type valve failure....
Over the years, Ive seen quite a number of O/C Heaters, and nearly without Exception, They were Mullard Valves and CRT's and even in sets fitted with a thermistor heater current limiter designed to apply a slow warm-up....
--Of different types, some large pentodes, some small signal triodes, But Always bloody Mullard!
(Why on earth are Mullard valves held in such high regard, when they cant even get the heater right!--Good marketing I guess....)
I have around 12 of EF91 types, of which I have 8 Mullard CV4014/M8083 'special-quality' types, four of the 'blue-glass' type and four clear, so are of a different batch range...
Of the 8 Mullard, 5 show that sudden bright heating when first powered, where the heater is welded to the lead-out wires-One is particularly bad, and glows nearly white for a half-second or so. Heater supply checked, and is 5.9-6.0V
All these valves are NOS...
The other four are I believe STC and are CV138,(EF91 equivalent) and are quite well used, dont show this effect at all.
--And, That is just the EF91 types Ive had issues with!
take a look at your examples of Mullard valves and see how many are apparently poorly assembled, where the structure is not straight or the Micas are cracked etc, and even the bulb itself mis-shapen......
Nah, I Dont like or trust Mullard....
So the consensus is that we can leave the heater partially on with about half the nominal voltage and have no HT. This way we cut down over all power consumption when the amp is not in use but keep things slightly warm and have a quick return to the 'ideal' operating condition/temperature within a few minutes instead of an hour or more ! Long term disuse will of course require the system to be completely turned off. By that I mean anything over say 6 hours could be called long term disuse ?
That's not my belief at all. I think you should turn them on when needed and off when not. I haven't had a single demonstrable tube failure other than the noted Mullards in several decades due to heating issues. Leaving them partially powered wastes energy for no benefit that I can determine.
I don't think that reduction in warm up time is significant enough to justify the additional complexity.
It will take just as long to return to "ideal" operating conditions in either case, a few seconds less heating time until the tubes start to conduct does not materially change how long it will take for the tube internals to reach thermal equilibrium for the new operating conditions. To be otherwise you would have to have plate power applied and be dissipating appreciable power which by definition is not a good idea unless you can assure you are not depleting the space charge around the cathode under such conditions - without great care this is the exact scenario where cathode stripping can occur in generic receiving tubes.
An HOUR or more? It only takes a few minutes.
But if you can't wait even that long yes bring the heaters down, here is a kind of "sleep mode" that is not hard to do. Look at the link below and scroll down to the bottom and start at "Improved standby switching". You can do quite a lot by just addind a diode and a resistor to the common design.
http://www.freewebs.com/valvewizard1/standby.html
lol - an hour????
you have design issues... Hey, Class A is seriously carbon un-neutral already. Standby with heaters on half power is the same as leaving your car idling all night so that engine wear on start-up isn't an issue. Turn it off. You aren't listening anyway. Or look to a soft start system instead.
you have design issues... Hey, Class A is seriously carbon un-neutral already. Standby with heaters on half power is the same as leaving your car idling all night so that engine wear on start-up isn't an issue. Turn it off. You aren't listening anyway. Or look to a soft start system instead.
With regard to malfunctioning "Mullard" valves, it has to be remembered that "Mullard" was a wholly-owned subsidiary of Philips and whilst it had several factories in the UK, the valves branded as "Mullard" could be sourced from anywhere in the far-flung Philips empire. If you want to avoid failures in the future you will have to decipher the manufacturing codes etched on the valve glass which specify valve type, factory and manufacturing date.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.