I said "likely", not 'possible'. However, it is true that RF oscillator circuits are in general far more fussy about valve characteristics than amplifiers. Rh-k problems which would present no problem to an RF amp could ruin frequency stability in an oscillator. Resistive 'swamping' is the last thing you do in an oscillator, where tuned circuit Q is all important; if swamping is needed then it is done with high quality capacitors.Merlinb said:
Hum sent to the cathodes of an LTP will be balanced out, to a significant extent, so the problem is not as great as might be expected.
General operational recommendations Philips SQ-tubes, jan.1975:
To minimise the influence of variation and spread of the leakage current between heater and cathode the resistance of the external heater to cathode circuit should not exceed 20kohm in R.F. circuits where frequency stability or preservation of wave form is required and in A.F. circuits with low signal level.
However, when the D.C. value of Vkf is at least 3 times the RMS value of the heater voltage an external resistance between heater and cathode of max 220kohm can be used provided that the hum voltage wich may then occur across the cathode resistor can be accepted for the application considered.
To minimise the influence of variation and spread of the leakage current between heater and cathode the resistance of the external heater to cathode circuit should not exceed 20kohm in R.F. circuits where frequency stability or preservation of wave form is required and in A.F. circuits with low signal level.
However, when the D.C. value of Vkf is at least 3 times the RMS value of the heater voltage an external resistance between heater and cathode of max 220kohm can be used provided that the hum voltage wich may then occur across the cathode resistor can be accepted for the application considered.
Last edited:
This may be also of interrest to some of you, and is from the same source (Philips, SQ-tubes):
Limiting values of Vkf apply to the positive and negative D.C. component of the voltage between the cathode and any of the heater terminals.
The limiting peak value is 2 times the rated D.C. value with a maximum of 315V
At the published values only the risk of breakdown is considered. No conclusion with respect to hum should be drawn from this figures.
For those, looking for a cathodefollower tube with high Vkf look at the PC92, it needs 300mA or 3,15V on the heater and can usually be bougth quite cheap, it was used in 220V serial heated TVs, Vkf 250V/315V.
Limiting values of Vkf apply to the positive and negative D.C. component of the voltage between the cathode and any of the heater terminals.
The limiting peak value is 2 times the rated D.C. value with a maximum of 315V
At the published values only the risk of breakdown is considered. No conclusion with respect to hum should be drawn from this figures.
For those, looking for a cathodefollower tube with high Vkf look at the PC92, it needs 300mA or 3,15V on the heater and can usually be bougth quite cheap, it was used in 220V serial heated TVs, Vkf 250V/315V.
I am not sure what a 20k Ohm resistor does if it is connected from the filament to the cathode.
1 pF of capacitance has Xc of 20k Ohms at 7.9 MHz.
Parallel that with 20k Ohms of resistance.
10 pF of capacitance has Xc of 2k Ohms at the same 7.9 MHz.
I will make a wild guess and say that filament to cathode capacitance is >1pF and < 10pF.
What good will a 20k Ohm resistor do for most RF tubes filament to cathode interface?
When I remember treatment of RF tube filaments, I remember one thing:
For grounded grid RF amplifiers, I have seen chokes on each lead of the filament.
That allows the filament RF voltage to more or less follow the cathode RF voltage, and prevents RF
currents from being sent to ground there.
That technique can be applied both to Power grounded grid tubes, and to Receiver Input RF amplifier tubes.
1 pF of capacitance has Xc of 20k Ohms at 7.9 MHz.
Parallel that with 20k Ohms of resistance.
10 pF of capacitance has Xc of 2k Ohms at the same 7.9 MHz.
I will make a wild guess and say that filament to cathode capacitance is >1pF and < 10pF.
What good will a 20k Ohm resistor do for most RF tubes filament to cathode interface?
When I remember treatment of RF tube filaments, I remember one thing:
For grounded grid RF amplifiers, I have seen chokes on each lead of the filament.
That allows the filament RF voltage to more or less follow the cathode RF voltage, and prevents RF
currents from being sent to ground there.
That technique can be applied both to Power grounded grid tubes, and to Receiver Input RF amplifier tubes.
Last edited:
Much of what I learned as a kid about vacuum tubes was from the 1956 Radio Amateurs Handbook. A lot of the rest of it was from early Popular Electronics magazines.
And then there were kits, and tinkering.
But even at that, I later found out I was only scratching the surface.
Never stop learning, and pass on what you find out.
And then there were kits, and tinkering.
But even at that, I later found out I was only scratching the surface.
Never stop learning, and pass on what you find out.
Datasheet Philips E83CC (nearly equivalent to ECC83)
Range values for equippement design:
Voltage between cathode and heater Vkf = 100V
Leakage current between cathode and heater Ikf = 5uA max
My guess would be that Philips arrived at 20kohm because !/1000part of the leakage resistance seemed safe enough (except for applications that have been clearly put forward in the "General operational recommendations" )
Range values for equippement design:
Voltage between cathode and heater Vkf = 100V
Leakage current between cathode and heater Ikf = 5uA max
My guess would be that Philips arrived at 20kohm because !/1000part of the leakage resistance seemed safe enough (except for applications that have been clearly put forward in the "General operational recommendations" )
In sensitive audio applications just make sure the heater is more positive than the cathode so leakage current flows under all signal-levels from the cathode to the heater and not the other way around.
With Rf on the cathode (heater fed through usually quarter-wave chockes) the cathode is normally directly heated.
The heater has to be shunted with a suitable capacitor that can support the current and has a cap. value that will tune the heater-capacitor resonance to a safely lower frequency than the lowest operating frequency to avoid additional heating do to Rf.
At the same time, harmonics of the heater/capacitor resonace circuit should under no circumstances come close to any operating frequencies or theyr harmonics.
I just pointed this cap-thingy out because this can make a huge difference in Rf-power tube operational life and I have fare to often seen this done wrong.
With Rf on the cathode (heater fed through usually quarter-wave chockes) the cathode is normally directly heated.
The heater has to be shunted with a suitable capacitor that can support the current and has a cap. value that will tune the heater-capacitor resonance to a safely lower frequency than the lowest operating frequency to avoid additional heating do to Rf.
At the same time, harmonics of the heater/capacitor resonace circuit should under no circumstances come close to any operating frequencies or theyr harmonics.
I just pointed this cap-thingy out because this can make a huge difference in Rf-power tube operational life and I have fare to often seen this done wrong.
Last edited:
Yes. As I said in an earlier post, if the heater-cathode impedance is an RF problem then the solution is to swamp it with capacitance, not resistance. The other thing to remember is that for a signal on the cathode any added resistance appear in series with the h-k impedance, not in parallel; that means that swamping would require a high resistance, not a low one.6A3sUMMER said:
By January 1975 most of the people who really understood valves will have either left Philips or become managers. Info from the 1950s might be more reliable.
For your information and only slightly off topic: there is an interesting article about valve reliability in the latest Elektor, at least in the Dutch version. It is a summary of an even more interesting article that appeared in the Philips Technisch Tijdschrift in 1956, an English translation of which is available for free online:
http://www.extra.research.philips.c...ve/PTechReview/PTechReview-18-1956_57-181.pdf
http://www.extra.research.philips.c...ve/PTechReview/PTechReview-18-1956_57-181.pdf
True, but irrelevant, because as you surely understand the content of this book has not been written in 1975, but taken over from a very long line of databooks concerning SQ tubes and miscellaeous devices. The book contain mainly maintaince types and obscolent types.
I also have the Philips databook "Receiving tubes" published march 1975. In it, even the EL503, one of the last tubes developed, is categorized as maintance type.
So, nothing new in those books.
I worked at Philips from 1970 till I started my own company in 1973. When I left Philips offered me to take over the Rf-heating business in Finland. Many years later, Philips cleaned out their library and since I was remembered as being interested in everything regarding tubes (which was unusual at the time) my former boss offered me to take what ever I want.
It was a treasure of databooks, application notes, scientific writings, laboratory reports a. s. o.
Last edited by a moderator:
Thrilled to have some of that posted here, and really appreciate the recent link.
So, why doesn't a constant current sink imperil a valve? Is it something about the current being set by the sink preventing some sort of thermal runaway?
I've been looking at some totem pole situations and wondering if the resistance of the lower tube should be subject to the same kind of limits.
A constant current sink where this resistance is maximized well beyond 220k seems not to pose a hazard in practice- why not?
Sadly, it was, most of it got destroyed when huoligans burned down my house and the firebrigade finished the destruction.
I was able to safe some of the books but all the much more interresting stuff is gone (except for what I still can remember, but my memory is not what it used to be)
Last edited:
Good question, but consider this:
A: At the time this was written tubes where about the only current sinks
avaiable.
B: with 220kohm the current will have to be below 1mA and at such low currents you
propably would not use a constant current source anyway (no current sink has the low
and stable capacitance a resistor has)
C: with a constant current source you will propably use higher current together with lower than
maximum allowed cathode-heater voltage. Lower voltage means lower leakage current.
D: You propably will not use a constant current sink in a application
that is sensitive to hum.
Last edited:
It can be much more than just an impedance-thing, the cathode is a good input to get multiplicative mixing.
Please explain how do you swamp a unpredictable leakage current (and the with aging increasing isolation degrading effects do to electrolyse) between cathode and heater with a capacitor?
Do you heat the heater with Rf to get rid of all adverse 50/60Hz "leakage" effects possible?
Last edited:
Maybe we should define "leakage" first, I see it as the unwanted total current between cathode and heater (except for ac-current do to capacitance between heater and cathode).
Currents do to:
Positiv voltage on cathode with respect to heater (cathodes inside surface acting as anode in respect to the higher temp. of the tungsten filament).
Same as above but halfwave rectified AC (this could explain the waveform distortion mentioned by Philips).
hk-voltage acting on the real resistance of the isolation (alumina, mica, glass, ceramic or whatever exist between hk).
I would also include the effects do to electrolytic events with DC voltage.
Would you agree ?
Currents do to:
Positiv voltage on cathode with respect to heater (cathodes inside surface acting as anode in respect to the higher temp. of the tungsten filament).
Same as above but halfwave rectified AC (this could explain the waveform distortion mentioned by Philips).
hk-voltage acting on the real resistance of the isolation (alumina, mica, glass, ceramic or whatever exist between hk).
I would also include the effects do to electrolytic events with DC voltage.
Would you agree ?
Last edited:
The heater cathode impedance can be represented by a parallel RC network (and some fixed series inductance if you are pushing up in to rf regions).
The simple RC representation is likely to work quite well when there is a substantial voltage offset, as the resistive component tends to be quite high, with little variation even for heater AC voltage variation of the offset level.
The resistive component when offset is near or sweeps through 0V is not so constant, and is in general a much lower magnitude compared to when there is substantial offset, and that magnitude can become quite low for poorly performing valves. Not only is the resistive component then quite variable with small changes in offset voltage, but other slow time frame effects can also be observed around 0V offset (when taking measurements).
The simple RC representation is likely to work quite well when there is a substantial voltage offset, as the resistive component tends to be quite high, with little variation even for heater AC voltage variation of the offset level.
The resistive component when offset is near or sweeps through 0V is not so constant, and is in general a much lower magnitude compared to when there is substantial offset, and that magnitude can become quite low for poorly performing valves. Not only is the resistive component then quite variable with small changes in offset voltage, but other slow time frame effects can also be observed around 0V offset (when taking measurements).
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.