"N_Cook" wrote in :

1.8mm slabs of porcelain under TO220 devices.

(This is continuation of Ampeg BA600-115, can of worms.
I've got mains back on it and is in working order , now to try with
high power.)

This is my educated guess as to what happened at manufacture as absent
but distinctive 20mm long bolt not found around here. The 4 output
devices should be held to the heatsink via a cross-bar over all 4 of
them and 2 bolts thru it into the heatsink. The insulators are ceramic
and the 2 central ones, butted up together , do not clear the
retaining bolts on their outer edges. After graunching the thread the
assembler decided to leave it out and trust to luck one off-centre
bolt would hold. Coming to reassembly I found a 3mm screw for
replacement but it would not hold, as tapped thread is stripped, so I
tapped that one hole out to 4mm. Tried grinding a clearance notch on
the side of one pad but got nowhere as ceramic.
Not only that but I had no trouble holding the slab in fingers while
trying to grind a slot , it was barely getting warm - good thermal
insulator or small grinding wheel not generating heat as not cutting
into the material ? So ,lateral thinking, decided to grind down the
screw thread where it interferes with the ceramic , on one bolt.
I wonder if this was a one off at manufacture or a whole batch like
this, all that was needed was shift a tapped hole 1mm, plenty of room
there for that. Or one ceramic pad 1 or 2 mm narrower.

ps
anyone use "chain mail glove" for small part in-hand grinding? At
80GBP and
upwards unlikely , firstly on cost, second need fingers for gripping
items. As an experiment wound some kevlar (ex fibre optic leads)
around finger of old glove and held in place with hot melt string.
Simulated grabbing/skittering of Dremmel and grinding disc into this
on a piece of dowel. It did not penetrate to the dowel, I expected the
kevlar to snag and stall the motor but cut clean through.




they may be -alumina- slabs,not porcelain;TEK used to use ~5/8" sq. alumina
slabs under the LVPS pass transistors on the 5000 series o'scope
mainframes.

--
Jim Yanik
jyanik
at
localnet
dot com