------------------------------
 
Date: Fri, 24 Jun 94 11:49:22 EDT
From: jas@proteon.com (John A. Shriver)
Subject: Re: KT88

The Richardson Electronics KT88 is expensive because production runs
of 20,000 tubes are very small.  They bought the KT88 tooling from GEC
in England, shipped it over, and set it up.  Also, Richardson is not a
low-cost operation.  They primarily make tubes for the military and
aerospace contractors.

The New Old Stock ones are expensive because they are rare.  I've
NEVER seen a NOS KT88 for sale, anywhere, at any ham radio flea.  I've
seen (and bought) Tung-Sol 6550's and Mullard EL34's, even seen
Genalex/GEC KT66's.  But never a KT88!

The reason that the KT88 is so rare and sought after is that it really
is a unique tube.

It combines the high power capability of the 6550 with the easy drive
requirements of the 6CA7/EL34.

The 6550 has a maximum grid circuit resistance of 50 kohms.  Very low.
Hard to drive.  The KT88 allows 100 kohms or higher, depending on the
exact internal dissipation.

This allowed amp designers to design high power amps with the KT88 in
the conventional (read: cost-effective) way, driving the KT88 grids
without cathode followers, capacitor-coupled.  Indeed, Hafler (in the
Mark III) drove the KT88 from the split-load phase splitter.

There are a LOT of amplifiers out there that are designed to take
advantage of these capabilities of the KT88.  You could run them with
6550's, but they would have to be biased down, to prevent grid
emission induced runaway.

Now, many modern amplifier designers have realized how important the
direct-coupled cathode follower drivers are for output tubes.  They
offer major improvements in bias stability in case of output stage
overload (positive grid current).  They also can overcome grid
emission problems (negative grid current) caused by gassy output
tubes.  (6550 class tubes tend to fail by going gassy.  Note how much
of a getter Tung-Sol used on their 6550, to try and control the gas.
Then look at how little is left on a used-up 6550.)  Another advantage
is that you can drive into Class AB2 and get a bit more power out.

This is really nothing new, you find designs back into the late 1940's
using direct-coupled cathode follower drivers.  (Eg. the Childs'
Amplifier.)  Later on, a classic example is the Heathkit W-6M.

Thus, I'd recommend any home-brewer use the direct-coupled cathode
follower driver, perhaps with a constant current load for better
linearity.  The available output tubes aren't getting any better, and
this stage can drive inferior ones just fine.

Moreover, today, the added cost of the cathode followers is far less
then the added cost of using KT88's instead of 6550's.