T
Tim Williams
Today holds two firsts for me. First of all, it occurred to me that hey, I
do have copper board in stock. It was perfboard that I mistakenly bought as
solid copper instead of pad-per-hole, which I usually prefer. So I decided
to build a schottky bridge / sampler head thing on the stuff. Science nut
that I am, I have plenty of ferric chloride solution in jars, so etching is
no problem. Resist is also no problem, as I have a sharpie marker at my
bench. So let's see here, place the parts...
http://webpages.charter.net/dawill/Images/Sampler_1.jpg
Draw around a bit, drop in the brown goo, rinse, sand off the paint, what
have we here? Add the parts, some solder and a healthy dose of rosin and
it's done!
http://webpages.charter.net/dawill/Images/Sampler_2.jpg
(Not shown, I later added jumpers around the top and side pads, making the
ground "pour" more continuous. I also put the balanced pulse connection on
the bottom.)
Following the flow chart of my Analog Sampler circuit, using a 555 and CCS
for sweep, LF353 for 2 x followers and LM360 and LM319 for high-speed
comparators (trigger, delay (with CCS) and retrigger holdoff, respectively),
and the previously discussed avalanche pulse generator (plus a 1:1 balun I
made, which seems to work just well enough), for the S&H one-shot.
After much puttering, I think I found an edge in all of this! The actual
waveform is very low duty cycle, with ~50ns edges from the output and ~10ns
edges from an internal tap. (I had to replace the output transistors in my
function generator with slow types, so I patched in a connection to a
testpoint from its preamp.)
You can see here the exponential slope of the output proper:
http://webpages.charter.net/dawill/Images/Sampler_3.jpg
And here, the preamp output:
http://webpages.charter.net/dawill/Images/Sampler_4.jpg
The waveform sucks Eeyore balls, chock full of jitter and irregularity, but
in all of it, there nonetheless lies an edge on the order of 10ns! Yippie!
*Dances a jig*
Next, I need to mess around with the timing: the displayed region seems to
be ~0.5us after the trigger event, so I need some serious delay, or un-delay
as the case may be. Not to mention do something about that damn jitter.
And why is it vertically erratic? Not triggering at the right point (with
regard to the voltage at that instant)? Not coupling cleanly? The pulses
should have identical effect and therefore cause the same error, whatever
that may be. (I have a common mode choke on the twisted pair, so there
shouldn't be much charge coupling, though into ~5pF and whatever
differential capacitance the diodes have, a lot could certainly get skewed.)
Tim
do have copper board in stock. It was perfboard that I mistakenly bought as
solid copper instead of pad-per-hole, which I usually prefer. So I decided
to build a schottky bridge / sampler head thing on the stuff. Science nut
that I am, I have plenty of ferric chloride solution in jars, so etching is
no problem. Resist is also no problem, as I have a sharpie marker at my
bench. So let's see here, place the parts...
http://webpages.charter.net/dawill/Images/Sampler_1.jpg
Draw around a bit, drop in the brown goo, rinse, sand off the paint, what
have we here? Add the parts, some solder and a healthy dose of rosin and
it's done!
http://webpages.charter.net/dawill/Images/Sampler_2.jpg
(Not shown, I later added jumpers around the top and side pads, making the
ground "pour" more continuous. I also put the balanced pulse connection on
the bottom.)
Following the flow chart of my Analog Sampler circuit, using a 555 and CCS
for sweep, LF353 for 2 x followers and LM360 and LM319 for high-speed
comparators (trigger, delay (with CCS) and retrigger holdoff, respectively),
and the previously discussed avalanche pulse generator (plus a 1:1 balun I
made, which seems to work just well enough), for the S&H one-shot.
After much puttering, I think I found an edge in all of this! The actual
waveform is very low duty cycle, with ~50ns edges from the output and ~10ns
edges from an internal tap. (I had to replace the output transistors in my
function generator with slow types, so I patched in a connection to a
testpoint from its preamp.)
You can see here the exponential slope of the output proper:
http://webpages.charter.net/dawill/Images/Sampler_3.jpg
And here, the preamp output:
http://webpages.charter.net/dawill/Images/Sampler_4.jpg
The waveform sucks Eeyore balls, chock full of jitter and irregularity, but
in all of it, there nonetheless lies an edge on the order of 10ns! Yippie!
*Dances a jig*
Next, I need to mess around with the timing: the displayed region seems to
be ~0.5us after the trigger event, so I need some serious delay, or un-delay
as the case may be. Not to mention do something about that damn jitter.
And why is it vertically erratic? Not triggering at the right point (with
regard to the voltage at that instant)? Not coupling cleanly? The pulses
should have identical effect and therefore cause the same error, whatever
that may be. (I have a common mode choke on the twisted pair, so there
shouldn't be much charge coupling, though into ~5pF and whatever
differential capacitance the diodes have, a lot could certainly get skewed.)
Tim
