Time ago I was testing the vacuum tubes for leakage current and
breakdown voltage using the focus connections from CRT monitors,
so the max. available voltage was 7500V from the very high output
impedance FBT's focus source. The leakage current for different
tested NOS 3 kV tubes was beginning to rise at HV levels between
6 and 7.5 kV and because of high output impedance, I could never
get a real breakdown between tube's electrodes.
I have bought few GU43 power tubes, '79 and '82 manufacture
date and found that their data sheets specify the allowed storage
period of time as being 8 years. Found by others local hams
also, this info was the source for some controversial discussions
here. For trying to find the right answer, few weeks ago I have built
a better HV source, using a switching AC HV supply after which I
added a tripling output rectifier block from an old TV set.
The tripling rectifier include capacitive filtering and the resulted
output impedance is much lower now. By varying the PS's output
switching transformer DC voltage, I can get an output DC HV from
about 300V to over 17 kV.
The measurements were made between the anode and closest
element (screen/suppressor grid), without any other previous
conditioning of the tubes cathodes/gettering.
HV was measured directly, by a 30 kV 1000:1 probe and the
leakage by inserting a 10k 2W resistor on the B- lead (plus
the needed protective parallel 5.1 V Zener), using two DMMs.
Now the interesting part...
I tested again some of the NOS tubes I had, plus others.
I have tested 3 GU46, 8 pcs of GU74, a GU78, a 4CX1500B,
SRL460 (4CX250), RS1009, two used 4S040 (4-400C), a used
At some high voltages, the leakage got high and then few discharges
could be heard from inside of each of the tested tubes (sparks could
be seen inside for the low power RS1009 glass tube). After few
discharges, the tube's breakdown voltage was increasing and the
leakage went low. For every important increase of the HV level
(for every few kV up), this phenomenon was repeating for two or
three times up to very high voltage when no more improvement in
breakdown HV level could result.
This was the case for most of the verified vacuum tubes, except for
the used 4CX1000 which presented no improvement and for the
used 4S040 (4-400) for which the improvement was less important
than for the other tubes (the all others were NOS).
I did not know what was happening inside the tubes, but I suspected
that somehow a similar process to a "gettering" took place.
One friend here confirmed my findings from its experience regarding
the use of NOS low power glass tubes (SRS551/RS1003) for his RF amp.
Searching the reflector again I have found the following messages:
These are explaining what was happening and my findings seems to
confirm the respective statements also. The tripling rectifier's
filtering seemed to help more for "debarnacling" or "conditioning" the
tubes (because of higher spark/discharge energy available?!)...
Most interesting was that two of the NOS GU74 (1985 manufacture)
could withstand very high voltages of over 13 kV. More than this, they
did not breakdown internally, but one did it externally at about 14 kV,
the other one at 15 kV (also outside of the tube), directly from the
cooler to screen's connection ring. The GU74's data sheet specify
the limit anode voltage of only 2kV DC and kV peak. For the
two NOS SRL460 (4CX250B) which have similar specs, (2 kV) the
breakdown voltages were found to be at about only 50% from
the GU74's measured ones. These two GU74s are the only tubes
for which I have found (until now) that are limited by the external
(outside of the tube) rather that internal breakdown.
NOS 4CX1500B at 10 kV and leakage not high until close to top HV
two of the GU43's and the GU78 breaks at about 14 kV,
third GU43 at 9 kV,
one GU46 at 12 kV,
other two GU46 at 14 kV,
others GU74 at 12 to 14 kV,
used 4CX1000 - high leakage: 170 uA at 9 kV and no improvement by the
500pF/5kv vacuum variable cap at 6 kV and no improvement by
My findings seems to confirm the results reported by some other
amateurs which are using the GU74 beyond the specs, at 3kV,
and the GU84 and GS35s at 4kV with no big problem, and this may be
a proof for the respective tubes ruggedness...
Of course, the tube's cathode conditioning and their emission
may be the subjects for other stories...
Obviously, some questions arising...
The conditioning is usually needed for large tubes, but the lower power
(1k) ones may need it too?!
May these tubes to need for similar "conditioning" again, after some
long time of unuse?
(I think they may do, at least after a long period of unuse, even if
another test after two weeks has shown here no detectable deterioration
during this time).
Can some NOS tubes to arc when first high RF level is aplied because of
missing such previous "conditioning"?
I don't want to start a controvercy, but any other such
All the Best from Bucharest,
Amps mailing list