I picked up a Tek 067-0554-00 tunnel diode pulser.
(There's one for sale on eBay now, it you'd like
to see a photo).
It is a little box that goes between a 100 volt square wave
generator and the scope's vertical input, to test the risetime.
The effect of the tunnel diode is to "sharpen" the edge
of the square wave.
Does anyone have the specs for this unit?
I'd like to know the specified risetime of the tunnel
diode pulser output and the type of tunnel diode used.
There are only a few parts in the box, and should be easy to home-brew.
I think a Tek TU-5 tunnel diode pulser is the same thing,
except it uses UHF type connectors instead of BNC.
The manual for the 1A1 (Nuvistor input) plug-in, page 6-9
shows a TU-5 in use to check risetime.
Well, I'd like to say that I had the answer to that, having compiled lists
cross referencing TD's to Instruments to Specifications - but I draw a blank
on that one. I *think* that a later version was made for the PG506
calibration generator - that is called an 067-0681-01 and is specified at
<=125ps. The tunnel diode for this is a 152-0177-02 or -03. The -02 is
10mA, 3pF germanium. The switching time for a Ge TD is about C/(2I) ps so
in this case 3/0.02 = about 150ps, consistent enough with Tek's
specification. The -03 version of the TD might have lower capacitance, but
I don't have figures for this.
The fastest TD's that Tek used were the 152-0383-00, at 50mA, 1.5pF used in
the S51 and S52 TDR pulse heads - which should be capable of 15ps, but
actually do a bit slower than that, because circuit capacitance adds a pF or
two.
Another fast one was the 067-0513-00 pulser, which used a 152-0254-01 diode
with 100mA, 6pF characteristics. Theory says it should do 30ps, which is
what the pulser is capable of (I've measured mine with an S4 head, and even
that involves care and calculation to unravel what you see from the S4's
25ps rise).
As for making a roll-your-own, problem 1 is finding a TD with the right
capability. You could cannibalise an old 7B92 to harvest the right TD.
Next is making sure that the layout doesn't compromise the switching time -
stay capacitance and inductance is a killer.
Cheers
Craig
