More heat...
There are three types of cathodes - it doesn't matter whether it's directly heated or indirectly heated, it's the emissive surface that is important.
Pure tungsten (brilliant white): Immune to cathode stripping
Thoriated tungsten (bright yellow): Susceptible to cathode stripping.
Oxide-coated (red): Very susceptible to cathode stripping.
So (as a quick way of remembering), the dimmer the filament, the more susceptible the cathode is to cathode stripping.
845 is thoriated tungsten, 300B is oxide-coated.
There are three types of cathodes - it doesn't matter whether it's directly heated or indirectly heated, it's the emissive surface that is important.
Pure tungsten (brilliant white): Immune to cathode stripping
Thoriated tungsten (bright yellow): Susceptible to cathode stripping.
Oxide-coated (red): Very susceptible to cathode stripping.
So (as a quick way of remembering), the dimmer the filament, the more susceptible the cathode is to cathode stripping.
845 is thoriated tungsten, 300B is oxide-coated.
Cathode stripping is perhaps a slightly misleading description. Cathode bombardment would be more accurate - remember that we are considering damage to the cathode emissive surface.
The vacuum in a valve is not perfect, so there are copious stray gas molecules randomly floating between the anode and cathode. If an electron should be accelerated towards the anode from the cathode, there is always a chance that it will strike a gas molecule and have sufficient energy to remove an electron from that molecule, rendering it positively charged and attracted to a lower potential such as the cathode. The problem is that the ion is not merely a single electron, it's a molecule with a nucleus composed of (heavy) neutrons having considerable momentum when it strikes the cathode emissive surface.
If the only force on particles between the cathode and anode was the accelerating electrostatic force of the anode potential, then the repellant force of the electron cloud above the cathode surface would be sufficient to completely prevent damaging cathode bombardment. However, Brownian motion means that some ions have additional momentum, and that tips the balance in favour of cathode bombardment.
We need to prevent/reduce ion bombardment, preferably by not producing positive ions, or by maintaining a protective space charge around the cathode emissive surface. In general, that means that we shouldn't allow the anode to become positive before the cathode is able to develop a space charge.
The vacuum in a valve is not perfect, so there are copious stray gas molecules randomly floating between the anode and cathode. If an electron should be accelerated towards the anode from the cathode, there is always a chance that it will strike a gas molecule and have sufficient energy to remove an electron from that molecule, rendering it positively charged and attracted to a lower potential such as the cathode. The problem is that the ion is not merely a single electron, it's a molecule with a nucleus composed of (heavy) neutrons having considerable momentum when it strikes the cathode emissive surface.
If the only force on particles between the cathode and anode was the accelerating electrostatic force of the anode potential, then the repellant force of the electron cloud above the cathode surface would be sufficient to completely prevent damaging cathode bombardment. However, Brownian motion means that some ions have additional momentum, and that tips the balance in favour of cathode bombardment.
We need to prevent/reduce ion bombardment, preferably by not producing positive ions, or by maintaining a protective space charge around the cathode emissive surface. In general, that means that we shouldn't allow the anode to become positive before the cathode is able to develop a space charge.
What sorts of collision energies are required to ionize a typical gas molecule (nitrogen, oxygen, water, by and large)? What sort of electron kinetic energies result in a typical tube construction with typical voltages used in audio? Say, a 400V plate potential and a 4mm cathode-plate spacing? How do these compare?
Pedantic points, unimportant: you really mean "protons and neutrons." And you mean a Boltzman distribution of momenta, not Brownian motion, a related but different phenomenon.
Pedantic points, unimportant: you really mean "protons and neutrons." And you mean a Boltzman distribution of momenta, not Brownian motion, a related but different phenomenon.
Fluorescent tube cathode-stripping
Fluoro tubes also have directly-heated coated cathodes. Where I worked we made some really good quality inverters to allow 15 watt tubes to run from a 24 vdc supply. We paid close attention to pre-heating the cathodes before the high voltage was applied to the tube. In one particular test that ran for about 6 weeks, on a 3 seconds on, 7 seconds off basis we got over 350,000 starts(!) from the tube before we gave up and decided it was good enough. Tube was still running OK. We then tried it using using no preheat time and the tube would fail in about 2 days. During the test you could see the tube ends gradually going black from the cathode material being splattered on the inside surface of the glass. Gradually the emission would be concentrated in a smaller and smaller area of the filament and it would eventually melt and go open circuit at this point.
Not exactly the same as a high-vacuum valve but from that it would seem the best way to protect your expensive bottles is to get the cathodes up to operating temperature, then ease the HT on over several seconds.
Fluoro tubes also have directly-heated coated cathodes. Where I worked we made some really good quality inverters to allow 15 watt tubes to run from a 24 vdc supply. We paid close attention to pre-heating the cathodes before the high voltage was applied to the tube. In one particular test that ran for about 6 weeks, on a 3 seconds on, 7 seconds off basis we got over 350,000 starts(!) from the tube before we gave up and decided it was good enough. Tube was still running OK. We then tried it using using no preheat time and the tube would fail in about 2 days. During the test you could see the tube ends gradually going black from the cathode material being splattered on the inside surface of the glass. Gradually the emission would be concentrated in a smaller and smaller area of the filament and it would eventually melt and go open circuit at this point.
Not exactly the same as a high-vacuum valve but from that it would seem the best way to protect your expensive bottles is to get the cathodes up to operating temperature, then ease the HT on over several seconds.
Photomultiplier Experience
How much enrgy is required to strip an electron from an atom?
Who knows BUT this experience sort of relates:
I run a 14 dynode photomultiplier tube with 2200 volts anode to cathode as the sensor in the receiver of a Laser Airborne Depth Sounder System (Hydrographic Survey).
As the tubes age and some gas gets in we see increasing "afterpulsing". This is when a "photo" electron being accelerated up the tube, toward the anode, hits a residual gas atom (probably Helium), stripping off an electron and turning it into a positively charged ion. This ion is then accelerated back the other direction and smashes into the photocathode, busting loose a few electrons which then acccelerate up to the anode, causing a false "after pulse".
Space charge saturation (electron cloud) around the cathode won't do anything to impede a positively charged ion getting back to the cathode.
I've been trying to get the photomultiplier tube manufacturer to add a Titanium getter for years - Titanium has a HUGE affinity for Helium.
As it is we just change the tube every 3 to 4 years (maximum).
Cheers,
Ian
How much enrgy is required to strip an electron from an atom?
Who knows BUT this experience sort of relates:
I run a 14 dynode photomultiplier tube with 2200 volts anode to cathode as the sensor in the receiver of a Laser Airborne Depth Sounder System (Hydrographic Survey).
As the tubes age and some gas gets in we see increasing "afterpulsing". This is when a "photo" electron being accelerated up the tube, toward the anode, hits a residual gas atom (probably Helium), stripping off an electron and turning it into a positively charged ion. This ion is then accelerated back the other direction and smashes into the photocathode, busting loose a few electrons which then acccelerate up to the anode, causing a false "after pulse".
Space charge saturation (electron cloud) around the cathode won't do anything to impede a positively charged ion getting back to the cathode.
I've been trying to get the photomultiplier tube manufacturer to add a Titanium getter for years - Titanium has a HUGE affinity for Helium.
As it is we just change the tube every 3 to 4 years (maximum).
Cheers,
Ian
To confuse this issue a little.
I just read on following
Source: http://yarchive.net/electr/tube_time_delay.html
"One common argument used by tube rectifier aficionados is the so-called cathode stripping effect where high voltage applied to the plate of a tube before the cathode has warmed up can strip the cathode of emitting material.
Unfortunately, this effect only occurs at high voltage, typically above 10
kilovolts. It is not a factor in small receiving or transmitting tubes.
If it really were a problem, it would destroy the tube rectifiers which have plate voltage applied immediately at turn-on.
It's only affect tubes typically >10 KV, while not affecting small signal tubes or transmitting tubes.
Just wondering is there any official research or paper publication to precisely define this concern ?
Thank & regards
I just read on following
Source: http://yarchive.net/electr/tube_time_delay.html
"One common argument used by tube rectifier aficionados is the so-called cathode stripping effect where high voltage applied to the plate of a tube before the cathode has warmed up can strip the cathode of emitting material.
Unfortunately, this effect only occurs at high voltage, typically above 10
kilovolts. It is not a factor in small receiving or transmitting tubes.
If it really were a problem, it would destroy the tube rectifiers which have plate voltage applied immediately at turn-on.
It's only affect tubes typically >10 KV, while not affecting small signal tubes or transmitting tubes.
Just wondering is there any official research or paper publication to precisely define this concern ?
Thank & regards
Another concern, apart from the risk to cathodes, is that of turn-on surge. If you turn on the HT quickly, by using SS diode rectifiers, the tubes are not yet heated and cannot conduct. There will be no current drain on the PS due to the tubes in the amp, so all the current through the rectifiers ia available to charging the smoothing and reservoir caps. This gives a great surge through the caps and some people think that's bad.
The remedy is to use a delay and a soft start, i.e. let the tubes get hot first, then allow the HT to build slowly. This can be achieved by using thermionic indirectly heated rectifiers, which can take up to half a minute before they start conducting, depending on type, and turn on quite slowly. This solution avoids the turn-on surge and also deals with any risk of cathode stripping, either real or imagined. It also results in a relatively
"clean" HT, free from SS diode "hash".
To minimize the impedance which is incurred with thermionic rectifiers, with the resulting voltage drop and "sag", thermionic TV damper diodes can be used. These have a long warm-up time and are quite rugged. They work well in either a "full-wave" circuit (center-tapped power transformer secondary) or in a Gratez hybid bridge, in which they make up the positive-going half of the bridge and SS diodes make up the negative-going half. TV damper diodes tolerate high PIV, high currents and usually high heater-to-cathode voltages.
The remedy is to use a delay and a soft start, i.e. let the tubes get hot first, then allow the HT to build slowly. This can be achieved by using thermionic indirectly heated rectifiers, which can take up to half a minute before they start conducting, depending on type, and turn on quite slowly. This solution avoids the turn-on surge and also deals with any risk of cathode stripping, either real or imagined. It also results in a relatively
"clean" HT, free from SS diode "hash".
To minimize the impedance which is incurred with thermionic rectifiers, with the resulting voltage drop and "sag", thermionic TV damper diodes can be used. These have a long warm-up time and are quite rugged. They work well in either a "full-wave" circuit (center-tapped power transformer secondary) or in a Gratez hybid bridge, in which they make up the positive-going half of the bridge and SS diodes make up the negative-going half. TV damper diodes tolerate high PIV, high currents and usually high heater-to-cathode voltages.
my diy-amp has 4xpy500A for rectification.....
I make sure i always preheat the py500A
Nevertheless, within 2 years they have developed a 'mirror-blackish' on the top of the tubes...at the top-cap. (that is in this case the cathode connection)
i suspect this is.....'moved' getter and a residu of other materials
could this be cathode stripping also??
or is it just conatmination???
greetings
I make sure i always preheat the py500A
Nevertheless, within 2 years they have developed a 'mirror-blackish' on the top of the tubes...at the top-cap. (that is in this case the cathode connection)
i suspect this is.....'moved' getter and a residu of other materials
could this be cathode stripping also??
or is it just conatmination???
greetings
kathodyne: It could be removed cathode, but that could be down to your ripple current.
ray_moth: I thought TV damper (or efficiency as they're known in Europe) diodes would be slow warm-up too, but when I measured 12CL3, they took exactly the same time to start conducting as any other thermionic diode - about 11s.
gingertube: What makes you think helium? If your photomultipliers are as expensive as the ones we used in telecines (£1500 each in 1992), I can see why you'd want to protect them.
ray_moth: I thought TV damper (or efficiency as they're known in Europe) diodes would be slow warm-up too, but when I measured 12CL3, they took exactly the same time to start conducting as any other thermionic diode - about 11s.
gingertube: What makes you think helium? If your photomultipliers are as expensive as the ones we used in telecines (£1500 each in 1992), I can see why you'd want to protect them.
EC, I've chewed this one around and I still have some trouble with the mechanism you propose. Here's the problem: the concentration of gas in a tube is quite low. Now, when there's a significant electron flux, there could indeed be a significant cross section for collision and ion formation. Problem is that at start-up, the electron flux is miniscule. The probability of electron-ion collision will certainly be proportional to the electon flux and the gas molecule concentration- if both are low, that cross-section will be very, very close to zero.
Things are different when there's a very high electric field. But at normal operating voltages, the field between cathode and plate is not horribly big, in fact, no bigger than the field between an insulated wire carrying B+ and the grounded chassis a few mm away.
Things are different when there's a very high electric field. But at normal operating voltages, the field between cathode and plate is not horribly big, in fact, no bigger than the field between an insulated wire carrying B+ and the grounded chassis a few mm away.
Ionization potentials
I just looked up some gas ionization potentials in the "Handbook of Physics"
Hydrogen: 13.598 eV
H2: 15.427 eV
OH: 13.18 eV
H2O: 12.60 eV
CO2: 13.79 eV
He: 24.587 eV
N: 14.534 eV
N2: 15.51 eV
O: 13.618 eV
O2: 14.01 eV
Ar: 15.759 eV
Kr: 13.99 eV
Xe: 12.127 eV
Rn: 10.745 eV
Neon bulbs and VR tubes strike over around 65V up to 100+V or so .
Maybe the important parameter is when the ions have sufficient energy to actually cause damage to the cathode coating. Which might be more like the voltages necessary to cause secondary emission or higher.
On page 1085 there is a secondary-electron emission table for the elements,
and these seem to be roughly 300 V to 600 V. No listing for any oxides though.
Don
I just looked up some gas ionization potentials in the "Handbook of Physics"
Hydrogen: 13.598 eV
H2: 15.427 eV
OH: 13.18 eV
H2O: 12.60 eV
CO2: 13.79 eV
He: 24.587 eV
N: 14.534 eV
N2: 15.51 eV
O: 13.618 eV
O2: 14.01 eV
Ar: 15.759 eV
Kr: 13.99 eV
Xe: 12.127 eV
Rn: 10.745 eV
Neon bulbs and VR tubes strike over around 65V up to 100+V or so .
Maybe the important parameter is when the ions have sufficient energy to actually cause damage to the cathode coating. Which might be more like the voltages necessary to cause secondary emission or higher.
On page 1085 there is a secondary-electron emission table for the elements,
and these seem to be roughly 300 V to 600 V. No listing for any oxides though.
Don
Here's an article addressing cathode stripping and other mythical failure modes:
http://ozvalveamps.elands.com/standby.htm
http://ozvalveamps.elands.com/standby.htm
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.