Out of interest to see some measured characteristics of heater-cathode DC conduction resistance, as a function of heater-to-cathode DC voltage, I quickly set up a simple test circuit and measured a small number of 12A-7 tubes.
http://dalmura.com.au/projects/Heater-cathode conduction plots.pdf
Klemperer from 1936 are the only results I've come across (http://dalmura.com.au/projects/Heater cathode insulation performance.pdf), although it appears the indirect heaters back then used different insulator material.
The sample of valves show a wide variation in resistance characteristics. Some are symmetric around nominal zero, whereas others are quite asymmetric - with some showing higher resistance of one polarity or the other.
In general, resistance increases with bias magnitude. In general, voltage saturation occurs for bias levels above about 10-20V.
Behaviour close to 0V bias is quirkier, and may be due to measurement technique - no measurement was taken at 0V bias, with 2V being the lowest level.
As Klemperer indicated, there was sometimes latency in settling.
I haven't seen any measurements of heater-cathode capacitance, but I'd be guessing it would be well above 1pF (which is about 3,000 Mohm at mains frequency) for the whole heater. Even though AC capacitive heater leakage is not uniform across the whole heater, capacitive leakage of AC heater current is likely to dominate DC conduction, although capacitance may well vary substantially with voltage as well. Maybe someone has access to one of those nice HP LCR meters with DC bias capability?
Ciao, Tim
http://dalmura.com.au/projects/Heater-cathode conduction plots.pdf
Klemperer from 1936 are the only results I've come across (http://dalmura.com.au/projects/Heater cathode insulation performance.pdf), although it appears the indirect heaters back then used different insulator material.
The sample of valves show a wide variation in resistance characteristics. Some are symmetric around nominal zero, whereas others are quite asymmetric - with some showing higher resistance of one polarity or the other.
In general, resistance increases with bias magnitude. In general, voltage saturation occurs for bias levels above about 10-20V.
Behaviour close to 0V bias is quirkier, and may be due to measurement technique - no measurement was taken at 0V bias, with 2V being the lowest level.
As Klemperer indicated, there was sometimes latency in settling.
I haven't seen any measurements of heater-cathode capacitance, but I'd be guessing it would be well above 1pF (which is about 3,000 Mohm at mains frequency) for the whole heater. Even though AC capacitive heater leakage is not uniform across the whole heater, capacitive leakage of AC heater current is likely to dominate DC conduction, although capacitance may well vary substantially with voltage as well. Maybe someone has access to one of those nice HP LCR meters with DC bias capability?
Ciao, Tim
Last edited:
Interesting data! Would be interesting to see a few modern valves on there too.
Heater-cathode capacitance is sometimes given on dataheets.
Heater-cathode capacitance is sometimes given on dataheets.
Very interesting.. I agree with DF96, a couple of Sovteks, JJs and sino tubes would be an interesting addition.
you might want to try your testing with both AC heater supply Vs. DC heater supply. as there is a electromagnetic reaction between the heater and the other tube componates.
Yes, I only have olde tubes! I'll see what I can borrow or steal.
The test technique is quite robust, but does require the use of battery power sources and a good lab meter to ensure rejection of stray mains. I'll use a linear regulator to generate the 6V for a few EF86's over the weekend.
I looked for heater-cathode capacitance specs, but only found grid-heater, which was likely aimed at defining the stray capacitance related to lead wiring.
I also incorrectly said in my first post that heater insulation material may be different for early valves, but I just rechecked, and they all appear to use alumina, but differences in quality/impurities seem significant. This link (http://dalmura.com.au/projects/SOME EXPERIMENTS ON THE BREAKDOWN OF HEATER-CATHODE.pdf) talks about that insulation breakdown and valve failure mechanisms which adds some insight to reasons why variations between tubes may occur.
I wanted to start without AC on the heater, to try and differentiate the DC effect before bringing in an AC effect. I'll start playing with an AC test jig soon to see what I can make robust.
The test technique is quite robust, but does require the use of battery power sources and a good lab meter to ensure rejection of stray mains. I'll use a linear regulator to generate the 6V for a few EF86's over the weekend.
I looked for heater-cathode capacitance specs, but only found grid-heater, which was likely aimed at defining the stray capacitance related to lead wiring.
I also incorrectly said in my first post that heater insulation material may be different for early valves, but I just rechecked, and they all appear to use alumina, but differences in quality/impurities seem significant. This link (http://dalmura.com.au/projects/SOME EXPERIMENTS ON THE BREAKDOWN OF HEATER-CATHODE.pdf) talks about that insulation breakdown and valve failure mechanisms which adds some insight to reasons why variations between tubes may occur.
I wanted to start without AC on the heater, to try and differentiate the DC effect before bringing in an AC effect. I'll start playing with an AC test jig soon to see what I can make robust.
I haven't chased up any new valves yet, but have measured the heater-cathode capacitance at 15-16pF for a few of the valves used for the conduction tests. Capacitance variation was negligible with heater on or off, or DC bias between heater and cathode, or measurement frequency (1-20kHz) or applied AC voltage (1-10Vrms).
That's about 180 Megohm impedance at mains frequency, which is significantly lower than most of the DC conduction levels measured so far. This would indicate that once hum was minimised with a humdinger pot, then getting any further improvement with elevated DC heater is probably a rarity. It also suggests that amplifiers hum less in Australia ;-)
That's about 180 Megohm impedance at mains frequency, which is significantly lower than most of the DC conduction levels measured so far. This would indicate that once hum was minimised with a humdinger pot, then getting any further improvement with elevated DC heater is probably a rarity. It also suggests that amplifiers hum less in Australia ;-)
It depends on the intended use whether they give Chk as a separate spec, or simply include it with a general "Ci". VHF types like the 12AV7 (Chk= 4.0pF) or the 6BQ7 (Chk= 2.6pF) include this figure in case grounded grid/cascode operation is used. At high VHF frequencies you need to isolate the heater with bifilar heater chokes to prevent the heater from short circuiting the input.
Audio types typically don't bother speccing Chk since it's presumed the signal frequencies won't be anywhere near high enough for it to make any meaningful difference.
Tnx for that interresting paper, please notice that in it also the rectifying effect is mentioned.
- Status
- Not open for further replies.