• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Advantage of beam triode over simple triode

Regarding measurement of small capacitance - I recall measuring the dielectric constant of air in a physics class lab (around 1969). We had an air capacitor in a bell jar with vacuum pump, cap was in an RF oscillator circuit. Measured delta F before / after evacuating and calculated capacitance change, proportional to change in dielectric constant which is about 1.00059. We measured frequency with a WWII vintage BC-221 frequency meter.
 
  • Like
Reactions: 1 user
Pentode grid-to-plate capacitance is so low because there is screen between them, which is grounded for AC signal. The residual capacitance is due to holes in the screen. Would be zero if screen was solid film instead of wire mesh.

Unlike triode's, pentode's gp capacitance cannot be directly measured.
Why not?

Jan
 
Regarding measurement of small capacitance - I recall measuring the dielectric constant of air in a physics class lab (around 1969). We had an air capacitor in a bell jar with vacuum pump, cap was in an RF oscillator circuit. Measured delta F before / after evacuating and calculated capacitance change, proportional to change in dielectric constant which is about 1.00059. We measured frequency with a WWII vintage BC-221 frequency meter.
Yes, there are many ways to skin a cat. It reminds me of a great story I read in a book by, iirc, Richard Feynman. He wanted to make the case for creative thinking in science and related the following.

A student was given the question of how he would determine the height of an apartment building by means of a barometer.
The answer: I will walk up the stairs and mark off 'barameter lengths'; on the wall. Then when I get at the top, I will count the number of marks and multiply that with the barameter length which gives me the apartment building height.
The Prof flunket him and told him to try again.
The student wrote: I will place the barometer on the ground so that its shadow ends at the same point where the shadow of the building roof ends. Then I have two congruent triangles and the ratio of the shadow lengths on the ground gives me the ratio of the building height to barometer height. I can measure the barometer height directly, and apply the ratio to give me building height.
No, the Prof said, try again.
The student wrote: I will climb to the roof of the building, and drop the baromter over the edge. I will time how long it takes for the barometer to crash on the surface. This time, and the equation for acceleration by gravity a = 2d/t^2, will tell me the apartment height d.
Now the Prof gets pissed and tells him - try again.
Undeterred, the student writes: I will go to the concierge of the building and tell him: 'My good man, if you tell me the height of this building I will give you this fine barometer'.
At which point the Prof calls in the student and says: 'You DO know which answer I am after, don't you?'.
Of course, the student says, but that's sooo boring, the same question, the same answer for decades, nobody learns anything from that.
He was passed.

Great story from a great man!

Jan
 
Last edited:
  • Like
  • Thank You
Reactions: 5 users
Yes, there are many ways to skin a cat. It reminds me of a great story I read in a book by, iirc, Richard Feynman. He wanted to make the case for creative thinking in science and related the following.

A student was given the question of how he would determine the height of an apartment building by means of a barometer.
The answer: I will walk up the stairs and mark off 'barameter lengths'; on the wall. Then when I get at the top, I will count the number of marks and multiply that with the barameter length which gives me the apartment building height.
The Prof flunket him and told him to try again.
The student wrote: I will place the barometer on the ground so that its shadow ends at the same point where the shadow of the building roof ends. Then I have two congruent triangles and the ratio of the shadow lengths on the ground gives me the ratio of the building height to barometer height. I can measure the barometer height directly, and apply the ratio to give me building height.
No, the Prof said, try again.
The student wrote: I will climb to the roof of the building, and drop the baromter over the edge. I will time how long it takes for the barometer to crash on the surface. This time, and the equation for acceleration by gravity a = 2d/t^2, will tell me the apartment height d.
Now the Prof gets pissed and tells him - try again.
Undeterred, the student writes: I will go to the concierge of the building and tell him: 'My good man, if you tell me the height of this building I will give you this fine barometer'.
At which point the Prof calls in the student and says: 'You DO know which answer I am after, don't you?'.
Of course, the student says, but that's sooo boring, the same question, the same answer for decades, nobody learns anything from that.
He was passed.

Great story from a great man!

Jan
Related to it is saying by Robert Wood that a good experimenter should be able to drill with a saw and to saw with a drill.
 
  • Like
Reactions: 1 user
When measuring Cag of a triode, you connect grid and plate to a capacitance measuring bridge. What do you connect to capacitance meter when you measure Cag of a pentode?
I would think connect G2 to anode, G3 to cathode, then measure between grid and anode?
Isn't that the way it would be in actual operation?

Jan
 
In normal operation both g2 and g3 are effectively coupled to the cathode, so both are shielding. To measure effective Cag1 with a meter I would first measure g1 to plate capacitance, with the other electrodes floating. Then I would measure plate to g3 capacitance and g3 to g1 cap. (other electrodes floating) Form the series cap equiv. Then measure plate to g2 capacitance and g2 to g1 cap (other electrodes floating). and form the series cap equiv. That tells you how much g2 and g3 are blocking the field coming from the plate going to g1. Subtract those two series equiv. caps from the original plate to g1 cap to get the net effective Cag1 capacitance. Should be accurate enough as a linear interpolation hopefully (not counting space charge effect).

Might try the test on a tube that has Cag1 listed to see if they agree.

I suppose one could send some current thru the tube while doing these measurements, if you want to include space charge effect. Would have to sort out the resistive effects possibly, but maybe a current source drive could keep that sorted. I think there are some heuristic factors given for space charge in some textbooks, maybe Spangenberg or Beck or Deketh.
 
Last edited:

PRR

Member
Paid Member
2003-06-12 7:04 pm
Maine USA
www.diyaudio.com
What do you connect to capacitance meter when you measure Cag of a pentode?
Or we could look in the book. From RDH4th:
 

Attachments

  • cgp-1.gif
    cgp-1.gif
    61.1 KB · Views: 34
  • cgp-2.gif
    cgp-2.gif
    43 KB · Views: 32
There is a tiny error in the scheme I mentioned above. (post 48) A little bit of plate field gets blocked by BOTH g2 AND by g3 on the way to g1, and is getting counted twice in the correction calc'd above. (a blocked portion, getting blocked again, doesn"t count again)
So one would want to subtract out the g3 blocking effect on g2 from the plate to g2 cap measurement used above.

Need to do an extra measure of cap between g3 an g2 to do that (all others floating). A similar series corr calc. ( Cg3g2 series'd with Cag3 gets subtracted from the Cag2 used in the above calc's ) Small effect I think. But with some pentodes listing 0.003 pF Cag1, probably has to be figured in, so as to not get a negative effective Cag1.

The problem with the RDH4 schemes is:
Where is ground with a two lead L meter?
Probably at a center tap between the active leads.
 
Last edited:
Looks like one has to correct for the grid 2 re-blocking of grid3 blocking on grid 1 also. So Cg3g1 used in post 48 gets the series combo of Cg3g2 and Cg2g1 subtracted out of it. Hopefully that's it.
Geez, so many corrections to do. Maybe you don't -really- want to know the effective Cag1!

So you need to measure these: (other electrodes floating)
Cag1
Cag3
Cag2
Cg2g3
Cg3g1
Cg2g1

and calc. this:
effective Cag1 = measured_Cag1 - ( Cag3 series [ Cg3g1 - (Cg3g2 series Cg2g1 ) ] ) - (Cg2g1 series [ Cag2 - (Cg3g2 series Cag3) ] )

"Series" means 1/(1/Cx+1/Cy) formula for series capacitances

Start with the innermost parenthesis and work outward.

and you thought Relativity was complex. This is right up there with US tax forms.
 
Last edited:
Or we could look in the book. From RDH4th:
That's not very helpful for a pentode.
I went to VA4's author and he advises:

" I wouldn’t attempt to measure pentode Cag directly (although Blumlein would have used a ratio bridge to do so). Just making the connections adds stray capacitances, so although the suggestion would work in theory, I think it would be very difficult to apply in practice because you’re subtracting one large number from another in the hope of determining a small number accurately. It’s much easier to measure Cag by configuring the pentode as an amplifier, measure its low frequency gain from a low source resistance, then add a 100k series resistor and plot high frequency response. All measurements taken from the anode. The Miller effect multiplies Cag to make it measurable by forming a low-pass filter at the input, so (knowing the f-3dB point) you can calculate the input capacitance, subtract Cgk and calculate Cag. I’d probably set up a spreadsheet to fit to the CR filter response to obtain f-3dB accurately. What you would need to be careful about would be end cap capacitance across your series resistor. It’s somewhere around 0.1pF for MRS25, but you could probably reduce it by connecting a string of resistors in series. Even the Miller capacitance method is going to need considerable care in practice. Low capacitance measurements are tricky. I felt I had done really well in measuring 8fF to an uncertainty of +/-60%!

Another problem you will find is that Cag is not constant with frequency. If you use the Miller capacitance method on a VHF triode (like 417A/5842) you will find that you get a much larger low frequency Cag than stated on the data sheet. Being intended for VHF, the manufacturer makes their measurements at VHF. In the end, all parameters are models and at high frequencies, models tend not to conform to the simplistic single component model that is valid at low frequencies."

Jan
 
In fact, my first 'real' job was with Philips in Eindhoven, The Netherlands, developing inductive components for color TV.
The high voltage flyback was indeed 25kV as I remember, and we used PL500 as secondary side regulator. Didn't know you Yanks did it at the primary side!
I guess it's a typo and you meant the PD500, see the Philips K7 CTV chassis schematics for example. The so called ballast triode acts as a shunt regulator to keen the CRT acceleration voltage as constant as possible, for resons explained above. With a black screen, the PD500 or 510 had to dissipate all the HV power, sometimes causing red-plating which coukld be observed by a close look through the HV cage holes.

Best regards!
 
You're absolutely right! We worked on K7 and later K9, if memory serves.
The whole HV stack indeed had its own cage to screen the ionizing radiation from such high voltage high speed electrons.
There always seemed to be some corona discharge, you could smell whether the lab setup was powered up or not.

Jan
 
  • Like
Reactions: 1 user
I guess it's a typo and you meant the PD500, see the Philips K7 CTV chassis schematics for example. The so called ballast triode acts as a shunt regulator to keen the CRT acceleration voltage as constant as possible, for resons explained above. With a black screen, the PD500 or 510 had to dissipate all the HV power, sometimes causing red-plating which coukld be observed by a close look through the HV cage holes.

Best regards!
You are correct! I meant PD500. PL 500 /504 where used as primary side switch running with picture line frequency for BW tv sets, PL509 / 519 for colour sets.
 
I took some pF measurements on a 17KV6A beam pentode, using the touch/no touch method:
Cpg1 4.3 pF
Cpg2 5.5 pF
Cpg3 8 pF
Cg3g2 7.7 pF
Cg3g1 5.3 pF
Cg2g1 11.8 pF
Cg1k 12.2 pF
Ckp 3.4 pF

datasheet values are:
Cin 22pF
Cout 9 pF
Cag1 0.6 pF

Since this is a BEAM power pentode, my above derived formula will not work due to g3 being a beam plate, which does not "filter" the plate field going thru g2.
This however makes it far simpler to calculate Cag1.
Measured Cag1 of 4.3 pF minus the series C shielding of g2:
Cpg2 series with Cg2g1 = 1/(1/Cpg2 +1/Cg2g1 )
So 4.3 - 1/(1/5.5 +1/11.8) = 4.3 - 3.75144 = 0.55 pF

0.55 pF comparing with 0.6 pF on the datasheet.

Some direct P to g1 capacitance in the lead wires would not get shielded by g2, so a small discrepancy.
A plate cap tube should give spot on results.
But tiny RF/IF tubes will probably have lead wire effects comparable to the grid effects.

So it's very easy to calculate Cag for Beam Power tubes.
 
Last edited:
  • Like
Reactions: 1 user