Jim Williams describes that too in an appnote, just a
450 Ohm and a 50 Ohm cable. I have experimented with that
together with a friend and it works great. Even managed to
improve it by adding a small trimmer from centre coax to
ground right behind the resistor.
Also tried lts of resistors and it is to bad I did
loose my notes, because there were some strange unexpected
result. If I remember well a big carbon composite or a
wire wound gave a very good result ( but we only tested on
on signal, it was just for fun, a 50MHz squarewave with
1ns risetime so it is possible those bizar resistors
showed just the right parasitics for only this frequency.
The simple carbonfilm 450 Ohm alone I used before and
goes well over a huge range. I did not used the probe
termination at the probeside ( williams did not do this
too)
But it could be better. Downside is the probe lenght
adds more induction and strays as a resistor at the cable
end. But I would not do it like in the linkt. I would use
a inline termiator to fit the bnc made probe. And if i
would use smd like he did, a better way is two 100 Ohm
resistors opposite from each other. I have done tests to
make a good vna load, but there is also lecture about
that, even Williams talks about it somewhere ( not probe
reated) i know also some other people who experimented and
2 resistors placed oposit is the most ideal. The
inductance added cancels the added capacitance. A 4
resistor configuration is more capacitive in total. A one
resistor is more inductive.?
Upto 1 ns and for scope use you will probbably not
notice much difference between 2 or 4 resistors. We used
them to find the " ideal" VNA load.
The 1 M input of a scope degrades by the input
capacitance. I measured it direct with 50 Ohm cable to a
vna. At 10 MHz it dropped allready to in the hundered of
Ohms. A Rigol dropped much faster as a 2445 or 7A26. At
100 MHz they are between 10 and 100 Ohm. That is why you
need a good 10X double compensated probe. There is a lot
of junk under probes.
my littleprobetest, from a Williams appnote, comparing
several probes from an old R&S probe and 1GHz active
probes to a modern 10X Tek and a 10X Rigol probe. I used a
50 MHz (500 ps) squarewave generator.
Fred PA4TIM
Op 13 feb. 2013 om 04:47 heeft "Dennis Tillman" <
dennis@...>
het volgende geschreven:
?
Another
option is to make your own 50 ohm probes. They
are not as hard as you might imagine and a lot
less expensive than buying a Tek probe. I have
made lots of 10X, 20X and 100X probes for my 50
ohm amplifiers. One big, big advantage of the
DIY probes is that you can solder them right
onto the spot in your circuit you want to
measure. This eliminates grounding problems and
insures good response from the probe because
there is virtually no ground lead inductance to
cause ringing or distortion.
?
Here is a
link to a DIY DC to 1GHz 50 ohm probe design.
This is a good place to start because he has
gotten so many questions from his ¡°students¡±
that he knows how to explain the details
clearly. The risetime from his design is much
faster than most vertical amplifiers except the
7A29 so the probe will not be the limiting
factor in your experiments.
?
Take a look
at the photographs I attached to this email I
sent to the forum (it is in the archives) to
explain it all:
From: Dennis Tillman
[mailto:dennis@...]
Sent: Friday, February 01, 2013 12:55 PM
To: 'TekScopes@...'
Subject: Resolving "undershoot" on 74LS logic
?
From:
Dave Daniel, Sent: Tuesday, February
12, 2013 5:52 PM
Folks,
Thanks for the information. Without going
through the TDS 'scopes' specs, I was a little
confused, but you've cleared that up. I've
also done what I should have done in the first
place, which is to go through in more detail
the catalogs, in particular the "7000 SERIES
OSCILLOSCOPE SYSTEMS/PROBE SELECTION GUIDE".
It's clear that the P6056 probe would be the
all-around best choice for use with my 7904A
and 7104 for high-BW measurements using
amplifiers with 1 Mohm input impedance. It
looks like my trusty P6106s will be adequate
for use with 50 ohm input amplifiers.
I've always liked the 7xxx series 'scopes, but
choosing a system (vert/horiz plug-ins,
special-function plug-ins, sampling heads and
probes) certainly requires a lot of homework!
As Rob would say, thanks for the bandwidth!
Dave
