"Jim Candela" wrote in message
. ..

"Phil Allison" wrote in message
u...

"Me" wrote in message
...

I recently had an experience using the Electro-Harmonix 7591 tubes in
a
McIntosh 1500 that I feel warrants a comment.


** I suspect the cause is just internal temperature ( due to high
plate
dissipation ) - at a high enough temp the ( imperfect) mica insulation
from
G2 to G1 becomes leaky and thermal runaway sets in.

If the 7591 grids were supplied with negative bias through a
lowish
resistance (or cathode drive were used) this sort of problem is
avoided.




............. Phil


I think Phil is correct. I would suggest lowering the idle current to
30ma. The heat will go way down, and the Plate B+ will go up a little. I
don't think the maximum power will drop very much....

I thought I would do a little more investigation. The thermal runaway due
to Mica insulation intrigued me, so I took one of my NOS 7591's and compared
it to one of the new E-H's on my Vacu-trace, while measuring actual screen
current. Results as follows:

In both case, plate was 380 volts, screen at plate potential and the tubes
were biases to 70 ma (I wanted to put some thermal load on the tubes and try
to spot a trend).

E-H 7591 Screen current: NOS 7591screen
current:
0 secs: 17.5 ma 6.5 ma
30 secs: 20 ma 6.8 ma
120 secs: 12.8 ma 6.9 ma
300 secs: 10.0 ma 6.8 ma

As you can see, the NOS tube stayed stable, with fairly constant screen
current flow. The E-H, on the other hand, immediately showed higher screen
current (?exposed more to the electron flow?), and then screen current began
to decrease, which I theorize is the screen beginning to emit electrons as
it heated (positive current initially measured was current flow into the
screen due to electron impingement....the only reason can think of for flow
to decrease is that the screen began to emit as it heated and this
screen -to-cathode flow (vs. the normal cathode -to-screen flow direction)
reduced the apparent current inflow). This could also be a conduction
effect across the mica to the screen spacers, of course, so maybe I haven't
proven screen emission, but I'll try to think of a better test.

No runaway, but the highest plate voltage I can get with the Vacu-trace is
380 volts, and I have no doubt that if I could push it to 450, I would get
screen current going thru 0, then negative followed by a runaway. I may try
that on my HV power supply soon.

The E-H tube is usable within standard 7591 specs ( 400 V on the screen),
but the Mac 1500 pushes it's tubes very hard at spec setup (460V plate, 450V
screen) that they apparently can get away with on the NOS tubes.

Interesting.











able to hear the slight rise in distortion (open to debate).