<azonenberg>
about to start assembling the PCBs, will link some photos shortly
<azonenberg>
The mixed rogers+fr4 construction looks weird
<monochroma>
:o
<azonenberg>
Just finished reflow, waiting for it to cool before hand soldering the tip
<azonenberg>
Too much solder. I need smaller paste apertures for the flip chip resistors
<azonenberg>
Not enough to wreck it but i dont like the look
<monochroma>
solder wick!
<azonenberg>
First things first. Probe tip and ground lead
<azonenberg>
Which is proving nontrivial to solder
<azonenberg>
Might need to make some kind of jig for this in the future
<azonenberg>
Hmmmmm
<azonenberg>
So, my probe loads less than the Pico 10x robe but has a slightly higher overshoot on the rising edge
<azonenberg>
And has three distinct humps in the TDR trace that i am almost certain are caused by the three resistors. Two big ones then a small one. Which makes sense since i have 200/200/50R
<azonenberg>
interestingly, when i probe at the end of the DUT line instead of passing through to the scope i see two humps instead of three. And i think less overshoot?
<azonenberg>
also i seem to have damaged the tip while pinching it with tweezers to solder it... it won't pogo
<monochroma>
:<
<azonenberg>
But i have plenty of boards and tips
<azonenberg>
This is just the first one
<azonenberg>
Good news is, it seems to work decently with the zs1500 ground leads and the overall mechanical fit and finish seems pretty good
<azonenberg>
Re the tdr humps, they're at 12.8 ns from the rising edge
<azonenberg>
Which i find quite interesting
<azonenberg>
rg188 has ~0.7C propagation velocity
<azonenberg>
whiiich gives a distance of 2.69 meters
<azonenberg>
the coax from the scope to the probe is ~1m long
<azonenberg>
So 'm wondering if this is a reflection off the scope or something?
<azonenberg>
i've seen some z0 probes with 50R termination to ground at the near end, which i think is intended to prevent this issue
<azonenberg>
i'm thinking of reworking this board with that to see if it helps
<azonenberg>
This is definitely my best probe yet
<azonenberg>
So its a start
<lain>
:D
<azonenberg>
The back termination turned the 3 humps into 1
<azonenberg>
So that was definitely it
<azonenberg>
Still not thrilled about the overshoot but i think it's going to be usable
<monochroma>
should be able to calibrate it out in measurements, right?
<azonenberg>
Yes, basically phase shift the waveform by that amount, scale, and subtract
<azonenberg>
It should be easily de-embeddable once i add probe de-embedding support
<monochroma>
:D
<azonenberg>
wait
<azonenberg>
thats a test fixture artifact
<azonenberg>
it happens with the lecroy active probe too
<monochroma>
:O
<monochroma>
huh!
<azonenberg>
Gimme a few mins more to collect some data...
<azonenberg>
I'm gonna probe at the end of a terminated line now
<azonenberg>
TLDR it looks like this probe is a lot better than i thought
<azonenberg>
There is a bit of ringing on incredibly fast edges that i'm still not thrilled about
<azonenberg>
I think it might be peaking at higher frequencies. Going to pop it on the VNA next
<azonenberg>
the prove moved during the measurement
<lain>
lol
<azonenberg>
ok reuploaded
<azonenberg>
That looks a lot less bad
<azonenberg>
18-22 dB out to about 2.4 GHz
<azonenberg>
If we call it -20 dB nominal we have 1.5 dB of peaking around 1.2 GHz and starting to roll off to -2 dB at 2.5 GHz
<azonenberg>
i'm not sure why there's so much variability in the frequency response, or if i can even trust the xaVNA to be this accurate
<azonenberg>
This is definitely my best probe to date by far though
<monochroma>
oh yeah, nice!
<lain>
flat is justice
m4ssi has joined #scopehal
<m4ssi>
hi, azonenberg i was trying to build cmake-scopenhal as describred on gslcope-manual.pdf....git clone https://github.com/azonenberg/scopenhal-cmake.git --recurse-submodules gave me some troubles
<m4ssi>
it tourned out that git try to update submodules via git protocol (ssh) instead https, so i've edited .gitmodules replacing git://..... with the relative https://
m4ssi has quit [Quit: Leaving]
<adamgreig>
you can put relative paths for submodules which are relative to the url you checked out the main repository, i.e. ssh if you used ssh and https if you used https
m4ssi has joined #scopehal
m4ssi has quit [Remote host closed the connection]
m4ssi has joined #scopehal
m4ssi has quit [Remote host closed the connection]