cathode follower calculations

[William Weinstein]

Hello All,

I realize this may be outside the current design Architecture, I was hoping
someone could lend assistance with some math. I'm using a 5840 miniature Pentode
in a triode configuration for noise reasons, further its application is cathode
follower in a True capicator microphone preamp. The Plate voltage is 105
regulated at ~3mA, the grid cap is .01uF@630VDC(its what I had on hand) the
Cathode resistor is 3.3K into 100K to Ground and the Grid leak resistor is 10M.
my output capacitor is 4.4uF (2X 2.2uF) into a 3:1 transformer (other options
are 7:1 and 10:1 on same transformer). Polarization works (41.2V derived from
plate voltage), my output is too low and has almost no bottom end, My best guess
is the output impedance matching is wrong, and or my input cap is too small.
Some quick measurements yeilded these results: with primary loaded ar 1.2K (.25w
MF resistor), secondary L=16.8H measurement taken with B&K L-C-R meter portable
model not bench model; with same load secondary Z=~70K, measurement taken with
TOA portable Impedance meter. All assistance is welcome.

Sincerely,
Bill

 

[Adrian Tuddenham]

[...]

... the Grid leak resistor is 10M.

...my output is too low and has almost no bottom end

A low signal could be due to a number of causes, but lack of bottom end
suggests that there is too short a time constant formed by the
microphone capacitance and the resistance effectively across it. Is the
grid leak going directly to earth or is it bootstrapped to the cathode?
How are you feeding in the polarising voltage?

For a capacitor mic you need hundred of meghohms between grid and earth
(and the mic and HT+). Either you need special high-value grid leak
and feed resistors or you need to bootstrap them so that the 'earthy'
ends follow the cathode signal and there is effectively no signal
voltage change across them, so they draw no signal current. Even with a
bootstrap. 10 megohms is a bit low.

Circuit 1 at:

http://www.poppyrecords.co.uk/other/CathFollower.gif

....shows bootstrapping for the grid and the HT+ with some typical
component values.

An alternative system which uses fewer components is shown in Circuit 2;
but this has the disadvantage that the microphone backplate (and
possibly casing) will be live. It also demands a very well smoothed and
decoupled HT+ supply.

If you have plenty of HT to spare and you only need 40v across the
microphone (which seems very low to me, as I would have expected about
300v), Circuit 3 shows a way of polarising the mic with the backplate
earthed, by increasing the voltage dropped across the cathode 'tail'.
(I should add that I have never seen it done this way before.)

 

[Robert Baer]

[...]

... the Grid leak resistor is 10M.

...my output is too low and has almost no bottom end

A low signal could be due to a number of causes, but lack of bottom end
suggests that there is too short a time constant formed by the
microphone capacitance and the resistance effectively across it. Is the
grid leak going directly to earth or is it bootstrapped to the cathode?
How are you feeding in the polarising voltage?

For a capacitor mic you need hundred of meghohms between grid and earth
(and the mic and HT+). Either you need special high-value grid leak
and feed resistors or you need to bootstrap them so that the 'earthy'
ends follow the cathode signal and there is effectively no signal
voltage change across them, so they draw no signal current. Even with a
bootstrap. 10 megohms is a bit low.

Circuit 1 at:

http://www.poppyrecords.co.uk/other/CathFollower.gif

...shows bootstrapping for the grid and the HT+ with some typical
component values.

An alternative system which uses fewer components is shown in Circuit 2;
but this has the disadvantage that the microphone backplate (and
possibly casing) will be live. It also demands a very well smoothed and
decoupled HT+ supply.

If you have plenty of HT to spare and you only need 40v across the
microphone (which seems very low to me, as I would have expected about
300v), Circuit 3 shows a way of polarising the mic with the backplate
earthed, by increasing the voltage dropped across the cathode 'tail'.
(I should add that I have never seen it done this way before.)

Practice indicates at least 30V and not much more than 100V across
the capacitive mike element.
However, theory indicates that a large voltage bias will cause more
non-linearity than a smaller bias; in the limit (zero bias) no
non-linearity - but then no signal either!

Most capacitive mikes i have seen use 300 megs, and i think i still
have a lot of those, so if you need one or two i would need a name and
address..

 

[Mark]

check the transformer, if you drive a transformer from a source Z that
is too high, you will loose the bottom end. What Z is the transformer
designed for?


Mark

 

[josephkk]

Sorry, made a mistake in the schematic. Input to U2 should be taken
from the top of R2.

It wouldn't be a follower if you did that. Just pull a smidgen of grid
current in U2 and it works fine.

?-)

 

[josephkk]

How loose? Will it fall off? ;-)

I suspect that it rolled off.

?-)