Cathode poisoning is completely unrelated, and takes the form of contamination of the cathode's active area with impurities found in the cheaper tubular nickel cathode sleeve. It occurs only in indirectly heated tubes when they are run at zero or very small anode currents, with heater hot, and a long interval passes.
But, There is still more than one effect at work, even under the heading of 'cathode stripping'.
The kind found in transmitting tubes running 2kV or more, where cold cathodes may be damaged by high-voltage collisions between electrons and stray gas molecules, is not believed to have much effect at receiving-tube voltages.
however, the other effect, described by van de Weijer, and posted earlier, is very relevant to low voltage amplifier valves, because distortion will be increased when it occurs. To add to the words of van de Weijer, try the Deutsche version Wikipedia:
Elektronenröhre ? Wikipedia
under the heading Lifetime (Lebensdauer), see this:
Das emissive Material der Kathode kann sich mit der Zeit langsam ablösen. Zum
einen kann das durch sehr starke Überlastungsmomente geschehen, die zum Beispiel
auftreten können, wenn die Röhre mit bereits angelegter Anodenspannung
aufgeheizt wird. Zum anderen findet in der Röhre eine stetige Abdampfung statt.
Dieser Vorgang verläuft unter normalen Betriebsbedingungen jedoch sehr langsam
und macht sich erst nach einigen zehntausend Betriebsstunden bemerkbar. Bereits
vor Taubwerden der Kathode kann als Folge Gitteremission auftreten.
The emissive material of the cathode can wear off slowly with time. In the First
place, this can occur through very strong overload pulses that can occur for
example if the tube is heated with instantaneously applied anode voltage. The
other effect takes place in a tube as constant evaporation. This process takes
place under normal operating conditions but is very slow and becomes apparent
only after tens of thousands of hours of operation. Before that, end-of-life of
the cathode can occur as a consequence of grid emission.
Also:
Eine Folge von der Kathode verdampfenden und sich auf dem Steuergitter
niederschlagenden Materials kann Gitteremission sein. Dabei emittiert das
Steuergitter Elektronen, was dazu führt, dass es positiver wird, sich dadurch
der Anodenstrom erhöht und sich der Arbeitspunkt soweit verschiebt, dass
Verzerrungen und/oder thermische Überlastung auftreten. Dadurch heizt sich das
Steuergitter noch weiter auf und emittiert umso mehr Elektronen. Besonders
anfällig sind Schaltungen, bei denen die Gittervorspannung über einen besonders
großen Widerstand (1 Megaohm oder größer) zugeführt wird. Dann reichen bereits
wenige Mikroampere Gitteremissionsstrom, um einen Ausfall zu verursachen.
One consequence of cathode material evaporation is felt when material settles on
the control grid - leading to grid emission. Control grid emission leads to a
positive grid potential, which increases the anode current and the operating
point moves to the extent that distortion occurs and / or thermal overload. This
makes the control grid heats up even further on and emits the more electrons.
Particularly vulnerable are circuits in which the grid bias is supplied via a
very large resistance (1 megohm or greater). Then just a few microamperes due to
grid emission can cause failure.
This is the effect described by van de Weijer. German readers, please improve my translation!
Grid-1 emission certainly spoils a high-quality amplifier, because of the increase in distortion.
Whether you care or not about the effect is up to you - it's your money and your valves! But if you preheat your valves to protect them from this, you must also design a slow ramp up of B+, to prevent inrush current damage. I use 6CJ3 damper diodes, which do a beautiful job of this.
But, There is still more than one effect at work, even under the heading of 'cathode stripping'.
The kind found in transmitting tubes running 2kV or more, where cold cathodes may be damaged by high-voltage collisions between electrons and stray gas molecules, is not believed to have much effect at receiving-tube voltages.
however, the other effect, described by van de Weijer, and posted earlier, is very relevant to low voltage amplifier valves, because distortion will be increased when it occurs. To add to the words of van de Weijer, try the Deutsche version Wikipedia:
Elektronenröhre ? Wikipedia
under the heading Lifetime (Lebensdauer), see this:
Das emissive Material der Kathode kann sich mit der Zeit langsam ablösen. Zum
einen kann das durch sehr starke Überlastungsmomente geschehen, die zum Beispiel
auftreten können, wenn die Röhre mit bereits angelegter Anodenspannung
aufgeheizt wird. Zum anderen findet in der Röhre eine stetige Abdampfung statt.
Dieser Vorgang verläuft unter normalen Betriebsbedingungen jedoch sehr langsam
und macht sich erst nach einigen zehntausend Betriebsstunden bemerkbar. Bereits
vor Taubwerden der Kathode kann als Folge Gitteremission auftreten.
The emissive material of the cathode can wear off slowly with time. In the First
place, this can occur through very strong overload pulses that can occur for
example if the tube is heated with instantaneously applied anode voltage. The
other effect takes place in a tube as constant evaporation. This process takes
place under normal operating conditions but is very slow and becomes apparent
only after tens of thousands of hours of operation. Before that, end-of-life of
the cathode can occur as a consequence of grid emission.
Also:
Eine Folge von der Kathode verdampfenden und sich auf dem Steuergitter
niederschlagenden Materials kann Gitteremission sein. Dabei emittiert das
Steuergitter Elektronen, was dazu führt, dass es positiver wird, sich dadurch
der Anodenstrom erhöht und sich der Arbeitspunkt soweit verschiebt, dass
Verzerrungen und/oder thermische Überlastung auftreten. Dadurch heizt sich das
Steuergitter noch weiter auf und emittiert umso mehr Elektronen. Besonders
anfällig sind Schaltungen, bei denen die Gittervorspannung über einen besonders
großen Widerstand (1 Megaohm oder größer) zugeführt wird. Dann reichen bereits
wenige Mikroampere Gitteremissionsstrom, um einen Ausfall zu verursachen.
One consequence of cathode material evaporation is felt when material settles on
the control grid - leading to grid emission. Control grid emission leads to a
positive grid potential, which increases the anode current and the operating
point moves to the extent that distortion occurs and / or thermal overload. This
makes the control grid heats up even further on and emits the more electrons.
Particularly vulnerable are circuits in which the grid bias is supplied via a
very large resistance (1 megohm or greater). Then just a few microamperes due to
grid emission can cause failure.
This is the effect described by van de Weijer. German readers, please improve my translation!
Grid-1 emission certainly spoils a high-quality amplifier, because of the increase in distortion.
Whether you care or not about the effect is up to you - it's your money and your valves! But if you preheat your valves to protect them from this, you must also design a slow ramp up of B+, to prevent inrush current damage. I use 6CJ3 damper diodes, which do a beautiful job of this.
Thank you.
Can you please post a schematic of the way you use 6CJ3 as damper diodes?
best way is to just use the diodes as part of a hybrid bridge:
Lundahl Transformers - Hybrid power supply
I use UF4007 (Fairchild) diodes in the solid-state positions, and a 6CJ3 in the other two places. The HT comes up over many seconds, a very gentle rise. This is very kind to the valves.
use a capacitor-resistor snubber across the trafo secondary (47n/1500v - LCR capacitors type PC/HV/S and 50R/2W).
I have used a pair of JJ 300B with this supply since february 1999 (constant use, as my main amp), and they still sound really fine - but the filaments are current regulated and set at 2% of nominal voltage as well.
Coul also just use a solid state bridge, and insert a 6CJ3 in series with the dc supply.
Lundahl Transformers - Hybrid power supply
I use UF4007 (Fairchild) diodes in the solid-state positions, and a 6CJ3 in the other two places. The HT comes up over many seconds, a very gentle rise. This is very kind to the valves.
use a capacitor-resistor snubber across the trafo secondary (47n/1500v - LCR capacitors type PC/HV/S and 50R/2W).
I have used a pair of JJ 300B with this supply since february 1999 (constant use, as my main amp), and they still sound really fine - but the filaments are current regulated and set at 2% of nominal voltage as well.
Coul also just use a solid state bridge, and insert a 6CJ3 in series with the dc supply.
Rod, what I get is that you use a pair of 6CJ3 as HT rectifier (plus a pair of SS diodes). Is there any benefit using 6CJ3 over conventional rectifier, like 5R4?
Also, when the filaments are current regulated (which is very good in itself for prolonging the tubes' life) – the HT will be there long before the tubes will be fully heated.
Also, when the filaments are current regulated (which is very good in itself for prolonging the tubes' life) – the HT will be there long before the tubes will be fully heated.
Just to add to the grid-current issue, it is interesting to note the following article:
Metson, Wagener, Holmes, Child. The life of oxide cathodes in modern receiving valves. Proc. IEE, 1952, vol 99, part 3.
They found that:
"more than 98% of the reverse grid current observed under ko conditions [i.e., after a a few hours burn in] was due to X-ray irradiation [from the electrons bombarding the anode] and less than 2% to true gas ions."
This was for small receiving pentodes with 2W heaters, at anode voltages up to 300V. This would suggest that unusually high grid currents must be due to some really enormous residual gas or grid contamination! Not just some minor effect.
They apparently had nothing to say about pieces of cathode becoming loose.
Considering the biggest names in cathode technology admitted the enormous difficulty of proving or even understanding the many processes taking place in an ageing valve, Weijer's musing can be taken as completely idle guesswork.
Metson, Wagener, Holmes, Child. The life of oxide cathodes in modern receiving valves. Proc. IEE, 1952, vol 99, part 3.
They found that:
"more than 98% of the reverse grid current observed under ko conditions [i.e., after a a few hours burn in] was due to X-ray irradiation [from the electrons bombarding the anode] and less than 2% to true gas ions."
This was for small receiving pentodes with 2W heaters, at anode voltages up to 300V. This would suggest that unusually high grid currents must be due to some really enormous residual gas or grid contamination! Not just some minor effect.
They apparently had nothing to say about pieces of cathode becoming loose.
Considering the biggest names in cathode technology admitted the enormous difficulty of proving or even understanding the many processes taking place in an ageing valve, Weijer's musing can be taken as completely idle guesswork.
I tried to condense this down to the relevant statements. Please let me know if I omitted something important.
Talking about cathode material evaporating:
"this can occur through very strong overload pulses that can occur for
example if the tube is heated with instantaneously applied anode voltage."
I don't believe this can happen by simply heating up the cathode while B+ is applied and all the electrode voltages are stable. Where do these "very strong overload pulses" come from?
More likely IMO due to slamming B+ and other voltages onto a hot tube and having the voltages bounce around for a while ;-)
Clearly if cathode material does make it to the grid it will cause grid emission but how does this happen when applying B+ to a cold tube?
I haven't seen anything compelling yet, just speculation.
OK if folks want to build in delays but I think smooth ramp-up with stable control voltages and stay below Vao is all that is needed to assure healthy cathodes.
Cheers,
Michael
PS Cathode poisoning, cathode stripping, and thermal shock are not even remotely "related effects" and are only related by being things that affect a cathode.
The 5R4 yields HT that appears almost immediately (<1 second) and ramps very quickly.
The 6CJ3 yields no HT for ~ ten seconds, then applies the HT at a very slow ramp, taking about 20 seconds. REALLY gentle.
You are looking at effects that occur 'after a few hours burn-in'.
van de Weijer's note examines effects that occur 'after many cycles of heating and cooling'.
THese are completely different operating conditions! You would not expect to see any cathode particles on the grid after so short an interval.
hello Michael,
Maybe I can clarify -
See van de Weijer's note again:
"The gradually "powdered" ceramic cathode emission surface may keep
minute amounts of electrical charge stored after cooling down; the
surface in cold state remains nonconductive. Minute amounts of these
cathode particles, either with remaining charge or electrically
polarized upon sudden apply of anode voltage, may "dust off" and clog
onto the most nearby "sieve" i.e.: the control grid; cathode stripping
has happened, and here it is that this less heard of tube
degradation/aging mechanism .. occurs."
THe important parts are:
1. The surface of the cathode does not conduct when cold.
2.If HT is still present when the heater is switched OFF (even a fraction of the HT) some charge may remain stuck to cathode particles, and not drain away (which they would if the heater is left on for a while after removing HT).
3. If you reapply the HT, this isolated pocket of charged material gets sucked towards the anode. Maybe it reaches the anode, or maybe it gets stuck en-route onto the grid.
4. Then you have degraded the noise and grid leakage of the valve.
THe Deutsche Wiki article reports the same effect, albeit without discussing the mechanism in such detail. This is too much smoke for there to be no fire, in my view.
Hmmmm...?Can you explain a bit?
I use current regulation in the 300B filaments, and adjust the current limit until 4.90V appears across the filament. I should have said "within -2% of nominal filament voltage".
Hi Michael;
I wanted to post a very terribly sarcastic remark about overload pulses through non-conducting cold tube, but decided to read before. And found your posting.
Still, my question is unanswered: what did he smoke? Do you have an educated guess?
Ah yes, that makes more sense. Thought maybe you meant you ran them 2% under nominal. OK. I had some Neuman tube mics that ran 5V on 6.5V heater. Schoeps used the same tube, but ran the full 6.3V.
I heard that some believe low heat means low noise. Did the Neumann perform quieter?
I run 6J32P (EF86) in guitar and mic preamps from 5.5V. They were quieter such a way.
As I remember, yes. The difference did not seem huge. But they were all wonderful mics. Needed 30 minute warm up to come into the correct sound.
Which tube is that? The only tube I know of that Neumann and Schoeps used in common is the AC701k, and at $700 each no one experiments with the filament voltage ;-) which is betweeen 3.8V and 4.2V. I own several tube mics both Schoeps and Neumann that use the AC701k and they are quiet enough at 4V... Besides, the biggest noise source in a condenser mic is the capsule if it's well designed and uses decent parts.
Neumann uses the EF86 in their U67 (of which I am fortunate enough to own an example) and the heater voltage is definitely 6.3V. And yes indeed the U67 does take at least 30 minutes for the sound to stabilize. Then don't fog up the capsule with your breath or it changes again... So unpredictable, just like some kind of musical instrument or something...
But I have heard of so-called starved heater or starved filament (essentially cold cathode) operation in the quest of superior sound and/or lower noise.
So now we have another tube life factor, low filament voltage. How does that affect tube life?
Last edited:
Some data sheets warn that low filament voltage gives shorter life.
But if you are getting lower noise - maybe that's OK, you get a better mic preamp. Exchange is no robbery...
The paper is a broad discussion of tube life over time, not only early effects. I just mentioned that bit to illustrate how much grid current is due to the unavoidable anode current (rather than to guess work).
Well, you said:
"more than 98% of the reverse grid current observed under ko conditions [i.e., after a a few hours burn in] was due to X-ray irradiation [from the electrons bombarding the anode] and less than 2% to true gas ions."
If the paper says something useful about long term effects, please post it.
The serious point about grid emission is not simply excess current, but that you are creating unwanted cathode-like emission centres on the grid. Since the grid-to-cathode voltage is changing during the signal cycle, you will get current contributions from the grid that will add to anode current in a nonlinear way across the signal cycle.
"more than 98% of the reverse grid current observed under ko conditions [i.e., after a a few hours burn in] was due to X-ray irradiation [from the electrons bombarding the anode] and less than 2% to true gas ions."
If the paper says something useful about long term effects, please post it.
The serious point about grid emission is not simply excess current, but that you are creating unwanted cathode-like emission centres on the grid. Since the grid-to-cathode voltage is changing during the signal cycle, you will get current contributions from the grid that will add to anode current in a nonlinear way across the signal cycle.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.