Also worth at least thinking about 'cathode hot spot' as a possible hazard before implementing any deviation from normal heater supply design IMO, even in triodes/pentodes. Such as constant current, low voltage, etc ............anything which delays transition of cathode temperature through vulnerable range.
Thermionic emission is described by Richardson-Dushman equation
Hence, for a given temperature, any mechanism that increases the work function of the cathode will produce emission decay.
Ignoring natural causes for long term emission decay, as cathode evaporation, bad quality metal-oxide coating interphase, too thin oxide coating, etc.; only we can control ion bombardment in some particular cases.
Let’s consider again Ion Bombardment
Case 1: Vak = 0, Hot Cathode
The energy of electrons emitted by the cathode is fairly low, then the probability of ions creation is also fairly low, furthermore, the dense electron cloud makes ion-cathode collisions with enough energy are even less likely.
In this scenario, a short time without applied B+ would be innocuous.
Case 2: Vak = B+, Hot Cathode
Since the 40’s there are documented evidence of ion bombardment at low energies, e.g. 5eV, 10eV and 15eV
Note on Volt-Dependent Poisoning Effects in Oxide-Cathode Valves - Abstract - Proceedings of the Physical Society. Section B - IOPscience
At such low energies, instead of cathode sputtering, we have cathode poisoning because of ion adsorption mechanisms, in the long term they can form a layer that produce emission decay.
This is the most common case and it is beyond of our scope, nevertheless it is reversible at some extent, by controlled cathode sputtering, experimentally verified with more than a hundred CRTs.
Case 3: Vak = B+, Cathode warms from cold
This is the worst case scenario, because we could have cathode sputtering and for sure we can have copious cathode poisoning.
All our efforts must be made in order to avoid this case; as it is the only that we can prevent, it is silly do not do it. IMHO.
J0 = A0 T² exp ( - e φ / k T )
Hence, for a given temperature, any mechanism that increases the work function of the cathode will produce emission decay.
Ignoring natural causes for long term emission decay, as cathode evaporation, bad quality metal-oxide coating interphase, too thin oxide coating, etc.; only we can control ion bombardment in some particular cases.
Let’s consider again Ion Bombardment
Case 1: Vak = 0, Hot Cathode
The energy of electrons emitted by the cathode is fairly low, then the probability of ions creation is also fairly low, furthermore, the dense electron cloud makes ion-cathode collisions with enough energy are even less likely.
In this scenario, a short time without applied B+ would be innocuous.
Case 2: Vak = B+, Hot Cathode
Since the 40’s there are documented evidence of ion bombardment at low energies, e.g. 5eV, 10eV and 15eV
Note on Volt-Dependent Poisoning Effects in Oxide-Cathode Valves - Abstract - Proceedings of the Physical Society. Section B - IOPscience
At such low energies, instead of cathode sputtering, we have cathode poisoning because of ion adsorption mechanisms, in the long term they can form a layer that produce emission decay.
This is the most common case and it is beyond of our scope, nevertheless it is reversible at some extent, by controlled cathode sputtering, experimentally verified with more than a hundred CRTs.
Case 3: Vak = B+, Cathode warms from cold
This is the worst case scenario, because we could have cathode sputtering and for sure we can have copious cathode poisoning.
All our efforts must be made in order to avoid this case; as it is the only that we can prevent, it is silly do not do it. IMHO.
Last edited:
A good source of statistics is book called "getting there most from vacuum tubes." It is the only book I know of that covers vacuum tube failure rates and the causes and has stats to back it up. In was written in about 1960 which means it is relatively modern as tube books go.
The book does say the cathode depletion is "real" but say it takes tens of thousands of hours to happen and almost certainly some other problem will happen first. The danger they say is not this kind of normal use but from over heating, red plating, arcing, allowing the glass envelops to over heat and other abuse.
The number one thing you can do to extend tube life is control the heater voltage. Run the tubes within the manufacturer's specs but on the low end of the range. 5.9 volts puts you 5% under the nominal 6.3v which is still well within the specified plus/minus 10% range.
Heater biasing as a way to address "hum" is covered in the book also. The author suggest some way to "balance" the voltage in the two legs of the hater circuit by connecting the boat to the center tap of the heater transformer. Lacking a CT you can use a resister network and or a 200 ohm pot.
The book should be on this web site. (copyright has expired so the PDF is 100% legal.) Technical books online
But on the other hand why bother to care about tube failure rates? Tubes are cheap and easy to replace. If you want to prevent expensive failures in a tube amp then (1) place fuses on the high volt sides of all transformers. and (2) but an aprox. 400 ohm power resister in parallel with the speaker terminals. It will rob only a slight amount of power and provide a safe load even of the speaker fails or the speaker leads become disconnected. Using a minimum 10R cathode resister on power tubes will also save the OPT if a tube shorts it will limit the current until the fuse blows.
I'l sure many of us have seen blown transformers. This is generally an expensive problem. But who here as ever seen early (less then a thousand hours) reduced cathode emission that was not caused by abuse in audio gear that runs with B+ under 1KV?
They want $33 to read this. Can you summarize for those of us who can't read the paper?
What was the MAGNITUDE of the in-circuit effect?
How did it change the characteristics of the tube relative to the curves published in the tube data book?
Was the total effect greater or less than the normal manufacturing tolerances?
Last edited:
I know, I did not read the paper either.
I cited it just to show the energies involved on ion bombardment under normal operation.
You can also search for an old paper with almost the same numbers, i.e. 5eV, 10eV and 16eV
Metson, G.H., 1949, Proc. Phys. Soc. Lond.. 62B, 589.
What is the exact number of gas atoms inside the valve?
Does it include gas atoms absorbed into the anode?
Last edited:
It claims gas is evolved from the metal parts of a valve (actually the grid in the paper, which was used an an anode for testing) due to dissociation of a surface film. Some of the gas gets deposited harmlessly on neutral surfaces while some finds its way to the cathode and poisons it. There it becomes negatively ionised and so returns to the grid again. The process continues until all the gas is used up and the effect disappears, and the cathode appears to recover fully.
Hmm, that's the Tomer 1960 text again, which doesn't actually cover failure rates nor present stats, and most of the rate graphs don't even have a scale - it isn't serious literature IMO. Merlinb described it on this thread as "more of a 'beginner's guide' than an academic study", which just about sums it up IMO. I wouldn't place much value there as to indicative rates, except as a well meaning record of common-wisdom and lore - and such things aren't necessarily the way things really are.
Wouldn't this aggravate the process which allegedly might lead to cathode hot spot generation, in principle ?
Not if they are NOS and/or rare/irreplaceable they're not !ChrisA said:
Last edited:
For clarification, I mean Tomer 1960 doesn't explicity cover failure rates numerically or statistically.
It's just not that kind of text. But if there is any such text out there, now that would be interesting.......
I was ablate read this thanks to a linkI was sent. What this paper talks about is a really smart physics experiment where thy were able to identify the compounds involves by looking at the activation energy associated to the various oxides. The paper does not address tube performance.
I wonder if some one could propose a measurement that would capture the effect they are looking for.
I would not suggest deviation from the normal heater power supply design. It's normal to allow the heater voltage to go plus/minus 10% from the nominal 6.3VAC. I suggested a 5% deviation. What I suggested and did was 100% mainstream and within specs. "hum balance" pots are also common.
In some cases elevated (biased) heaters are actually required to stay within the manufacture's specs for heater to cathode voltage. (may apply to some designs with cathode well above ground)
That said, I have tried some experiments with 12AX7 tubes. I've looked at their curves at 5.7 volts and down to 5.1 volts. The last value, 5.1 volts, is about as low as they go and still work fine. Well at least the samples I had handy.
What got me started on this is at the design o the Hammond Novachord. Novachord Restoration Project
This was a tube based synth built in 1938 that contained about one ba-zillion tubes. I read that Hammond used 5.1 VAC on the heaters to reduce tube failure because abnormal voltage you'd be replacing a tube a month in a monster like this. So I wondered "would 5.1 volts work?" I think so but I'd not recommend it.
I dod have a project planned where I want to go slightly un-conventional with heaters. I will be a microphone preamplifier with built-in audio compression. I'll use DC heaters. I want at least 65db gain and also for the amp to be dead-quiet. I'm thinking of placing one LM7806 6-volt regulator chip at every tube socket and sending unregulated 16VDC around the chassis. The lm7812 have something like 100+ db of isolation.
Member
Joined 2009
Paid Member
I'm still using the first SET I ever built and it keeps going strong. I'm getting sick of it, I mean, when will it break - I was kind of hoping to have an excuse to pull a tube out and put in a new one - but it just sits there year after year.......��
Mind you, it uses a tube rectifier so no cold-on B+. I have thought about changing it for a SS rectifier because the power transformer gets a bit too hot for my liking and tube rectifiers are very inefficient beasts indeed.
Last edited:
For power tubes, you'll get the longest life if you design in accordance with manufacturer's recommendations.
For low level tubes, it's a bit more complex. In early computers based on vaccuum tubes, and certain professional test equipment, heaters were run low, at e.g around 5 volts. If the tube is drawing anode current according to datasheet application values, that will seriously shorten tube life. However, in such applications, designers were carefull to keep anode currents well below datasheet values as well. Thus the space charge still built up and the cathode emission coating was not stressed. And often the anode voltage was well within ratings as well.
It is the combination of quite low anode current and low heater voltage that brings longer tube life, not just low heater voltage.
Last edited:
- Status
- Not open for further replies.