I have yet to see any data sheets that give data on manufacturing
tolerances. In simple terms all I need to know at the moment is the spread
on Vg1 for, say 1mA anode current, for the two halves of an ECC83 from any
manufacturer. Am I asking too much?

Graham Holloway

I have found that Tungsol, Raytheon and Telefunken data sheets often have
tolerances stated. Philips also sometimes lists tolerances. I had a look
through what I have here and I found a Philips spec sheet for 12AX7S which
states that with 250 V on the plates and -2V on the grids, the plate
currents on each side will be 1.2 mA nominal, min 0.75 to max 1.75, and
within 0.6mA of each other. Now I know that is not exactly what you asked
for, but it is a measure of balance. At a nominal transconductance of 1.6
mA/V that works out to a spread in Vg of about .6 volts to achieve equal
currents.

As to wether you are asking too much, well, it depends on what you want to
do with the information. Most spec sheets dont bother with this stuff, and I
take the view that tube radios and amps in the old days were generally built
with 10% or worse components, tested with 20000 ohm VOMs, and designed using
slide rules, so the best thing IMHO is to just make sure to use cathode bias
etc and dont expect great precision.

I can send you the spec sheet if you like.

"Graham Holloway" wrote in message
...
I have yet to see any data sheets that give data on manufacturing
tolerances. In simple terms all I need to know at the moment is the spread
on Vg1 for, say 1mA anode current, for the two halves of an ECC83 from any
manufacturer. Am I asking too much?

Graham Holloway

The other source for tolerance information I almost forgot is the Mil specs.

http://www.dscc.dla.mil/Programs/Mil...cDoc=MIL-PRF-1

although I cant seem to find 12AX7.


"Robert McLean" wrote in message
.. .
I have found that Tungsol, Raytheon and Telefunken data sheets often have
tolerances stated. Philips also sometimes lists tolerances. I had a look
through what I have here and I found a Philips spec sheet for 12AX7S which
states that with 250 V on the plates and -2V on the grids, the plate
currents on each side will be 1.2 mA nominal, min 0.75 to max 1.75, and
within 0.6mA of each other. Now I know that is not exactly what you asked
for, but it is a measure of balance. At a nominal transconductance of 1.6
mA/V that works out to a spread in Vg of about .6 volts to achieve equal
currents.

As to wether you are asking too much, well, it depends on what you want to
do with the information. Most spec sheets dont bother with this stuff, and
I take the view that tube radios and amps in the old days were generally
built with 10% or worse components, tested with 20000 ohm VOMs, and
designed using slide rules, so the best thing IMHO is to just make sure to
use cathode bias etc and dont expect great precision.

I can send you the spec sheet if you like.

"Graham Holloway" wrote in message
...
I have yet to see any data sheets that give data on manufacturing
tolerances. In simple terms all I need to know at the moment is the
spread
on Vg1 for, say 1mA anode current, for the two halves of an ECC83 from
any
manufacturer. Am I asking too much?

Graham Holloway

Simon

My first task in a design is to determine the smallest possible value for an
inter-cathode potentiometer for balancing a long-tailed (differential) pair
without reducing the open loop gain significantly. I was planning on having
to derive the information from whatever limits are given.

Yes, the data you have would be useful as I have not found any other on the
internet yet.

Kind regards

Graham Holloway

"Robert McLean" wrote in message
.. .
I have found that Tungsol, Raytheon and Telefunken data sheets often have
tolerances stated. Philips also sometimes lists tolerances. I had a look
through what I have here and I found a Philips spec sheet for 12AX7S which
states that with 250 V on the plates and -2V on the grids, the plate
currents on each side will be 1.2 mA nominal, min 0.75 to max 1.75, and
within 0.6mA of each other. Now I know that is not exactly what you asked
for, but it is a measure of balance. At a nominal transconductance of 1.6
mA/V that works out to a spread in Vg of about .6 volts to achieve equal
currents.


OOPS, that should say .6 / 1.6 = .38 volts spread in Vg.

As to wether you are asking too much, well, it depends on what you want to
do with the information. Most spec sheets dont bother with this stuff, and I
take the view that tube radios and amps in the old days were generally built
with 10% or worse components, tested with 20000 ohm VOMs, and designed using
slide rules, so the best thing IMHO is to just make sure to use cathode bias
etc and dont expect great precision.


Don't forget that tubes wear out after a while, and that some
parameters will change as they age. A good design would work
in spite of this. And throw in loose tolerance parts in production,
and the engineers spend a lot of design effort to make the design
as robust as costs permit. Robust enough to outlive the
warranty period at least.... What they want is a design that
will work well without having to make adjustments on the
production line (no trimpots for example). Such adjustments
would probably drift anyway, so it's best to make the design
not need the trimpot.

Hi Robert,

Don't forget that tubes wear out after a while, and
that some parameters will change as they age.

I am sure you know this, but I might like to add that for
the original poster: At the other end of tube lifetime,
AKA brand new tubes, often show considerable parameter
changes during their first dozen hours of usage, too.

I especially found that old stock small signal tubes,
stored for decades, often need at least a day or two
of duty to stabilize. I am not talking about sonical
bla bla here, but just cathode current at given plate
and g1 voltages.

But the worst contenders I ever had were a batch of
old stock 807 which needed about 100 hours (or five days)
of burning in at about 15 watts Pd until they stopped
drifting. But since then, they stay nailed to the point
in my 807PP/UL amp since about two year of daily duty.

Tom

--
fnord (now you see it, now you don't)
- R.A. Wilson / FZ