Some might wonder about the reliability of Russian made output tubes,
notably the KT90EH sold by New Sensor in New York.
I have a positive story to tell.
In 1996, I built an 8585 amp shown at
http://www.turneraudio.com.au/8585-a...ober-2006.html
The initial tubes I used were new GE6550A which were regarded as the
best 6550 available.
I sold the amp in about 2003, and by 2004, some GE 6550 had
considerable reverse grid current, and showed signs of wear. The owner
had me upgrade the circuit to that shown at the link above after a
routine service, and with the EL84 drivers the Rg to each 6550 was
reduced to 100k, and this limited reverse grid current at idle, and
helped keep the bias voltage stable. 6550EH were also fitted to
replace the aged GE6550. The owner was very happy with the mod, but
had the desire to try other tubes and the 6550EH were replaced with
KT90EH. In about 2005, the amp was sold to its present owner who
drives 3 stacked pairs of fully re-engineered Quad ESL57. He's been
delighted with the sound, and has used the system every day since
2005. John Hall of Melbourne had re-engineered the ESL.
Now last week, the amp stopped working, and was sent to me for a fix,
its first breakdown since 1996. It turned out all four 5 Watt "75V"
zener diodes to shunt regulate the screen voltage had all changed
their zener voltage to lower voltages and so Eg = +220V instead of a
normal +330V, thus turning off anode current to all OP tubes, and no
collateral damage occurred. I've replaced the zeners with a far better
shunt reg based in a pair of series BU208P + MJE340 darlington
connected bjts on a heatsink because it seems that particular brand of
5W zeners is subject to failure after passing years. In fact the 75V
zeners had a zener voltage of around 82V when new. Their working
condition was with about 20mA in each, and 82V, and that's 1.64Watts,
and I'd cramped them between plate of aluminium and the aluminium
chassis with lots of heat paste to keep 'em cool as cucumbers.
Didn't stop them frying it seems.
The condition of the 8 x KT90EH was the same as if I had plugged in a
set of new tubes, but these have been used a lot since 2005, about 6
years, at least each day says the owner, who was surprised the amp
didn't need new tubes, and that it needed such little work.
The KT90 conditions are :- fixed bias, Ea = +475V,
Ia = 32mA, Eg2 = +330V, Ig2 = 2mA, and Eg1 = -37V.
The differences between applied Eg1 and Ia were less than 5%, and
between Ig2, less than 10%, and the Vdc across each 8 x 100k Rg less
than 0.03V, with little noise, ie, the tubes measured just like a new
set. I think they are lasting better than GE6550A.
The EL84EH in triode are also showing zero signs of wear. All other
NOS Australian made input tubes are also working as new.
Manwhile, the 6550EH which were replaced by KT90EH were supplied to
other owners at a low cost and appear to be working just fine.
I do hope the Russians are keeping up their good work, and If I could
send a bottle of Vodka, I would.
Patrick Turner.