So the main suspect is the S-4, but I have no idea how to troubleshoot it.
I can see the tiny sampling pulses being fed to it by 7S12, but nothing "notable" on the preamp
output
on pin 4. The manual says not to test the sampling bridge diodes using an ohmmeter, but a curve
tracer. How could an ohmmeter (actually DMM, <1mA test current) be more destructive to these
diodes than a curve tracer? (provided that the test is carried out using forward bias only). How to
test
the snap-diode? Is there any other extra-sensitive component I should be aware of?
The reason to use a curve tracer is that it is controlled. You operate it very carefully within the
safe operating area of the tiny diodes and check correct operation both in positive and reverse bias.
I would not even think about deviating from Tek's measurement procedure.
Now sure, the S4 was introduced in the 1960's (and what a design tour de force it was!) when modern
DVM's weren't even a twinkle in the measurement eye. So a meter was something with a 15V battery and a
needle - and 15V in reverse bias would certainly kill a travelling wave sampling diode. But even now
you need to check what the open circuit voltage of your meter is when checking resistance. It needs to
be an absolute maximum of 5V to prevent damage in reverse bias. And it is reverse bias you need to
check - it is that that will particularly reveal whether you have a fried diode. You need to do all
this very, very carefully.
You do know that the input to the S4 (and S6) is exceptionally prone to static damage? Even the static
charge built up on a length of coax can kill an S4. Tek even recommended handling the S4 and S6 on a
static mat with wrist strap.
Mine live with an SMA grounding cap on the inputs until I use them, and then I short circuit every bit
of coax I use when making a connection.
Try and find a copy of the Tek Applications Note 42W-5850 "Preventing Sampling Head Overdrive and
Static Damage".
Craig
