<azonenberg>
S-parameter view is coming along niecly
_whitenotifier-d has joined #scopehal
<_whitenotifier-d>
[scopehal] azonenberg pushed 2 commits to master [+0/-0/±3] https://git.io/J0OWU
<_whitenotifier-d>
[scopehal] azonenberg 9fb33f2 - SParameters: added GetNumPorts() method as preparation for supporting >2 port Touchstone files in the future
<_whitenotifier-d>
[scopehal-apps] azonenberg opened issue #386: scopesession files do not preserve X or Y axis units of waveforms - https://git.io/J0OBE
<_whitenotifier-d>
[scopehal-apps] azonenberg labeled issue #386: scopesession files do not preserve X or Y axis units of waveforms - https://git.io/J0OBE
<_whitenotifier-d>
[scopehal] azonenberg pushed 1 commit to master [+0/-0/±4] https://git.io/J0OEU
<_whitenotifier-d>
[scopehal-apps] azonenberg pushed 2 commits to master [+0/-0/±4] https://git.io/J0OEG
<_whitenotifier-d>
[scopehal-apps] azonenberg fc35eda - Can now load Touchstone files via command line arguments
<_whitenotifier-d>
[scopehal-apps] azonenberg b29c811 - Added initial support for offline scopes. So far only mock scopes can be offline and there's no support for changing between online and offline states.
<_whitenotifier-d>
[scopehal-apps] azonenberg pushed 1 commit to master [+0/-0/±1] https://git.io/J0ODn
<_whitenotifier-d>
[scopehal-apps] azonenberg 885265e - No need for line continuation
someone-else has joined #scopehal
someone-else has quit [Quit: Connection closed]
Guest3062 has joined #scopehal
Guest3062 has left #scopehal [#scopehal]
someone-else has joined #scopehal
someone-else has quit [Quit: Connection closed]
<Stephie>
that's simulated TDR? can scopehal do TDR/TDT to S11/S21?
someone-else has joined #scopehal
<azonenberg>
Stephie: scopehal can do channel emulation on all 4 s-parameters of a 2-port network on an arbitrary signal
<azonenberg>
to do TDR/TDT you use the "step" filter to generate a unit step
<azonenberg>
then apply an s-parameter channel for the appropriate port to it
<Stephie>
ahhh, makes sense
<azonenberg>
so TDR on port 1 would be step -> channel emulation of S11
<azonenberg>
etc
<azonenberg>
you can then optionally apply the TDR filter to that to re-scale the reflected voltage waveform to ohms or reflection coefficient
<azonenberg>
You can also of course do channel emulation (or de-embedding) with the same network on an arbitrary waveform including stuff acquired from a scope
<azonenberg>
it's fast enough to do interactively
<Stephie>
doing fft from a TDR/TDT measurement to S11/S21 would be an interesting feature though
<Stephie>
once i get a good tdr i might try
<azonenberg>
That has been a wishlist item for a while
<azonenberg>
so far i havent been able to get useful enough s-parameters from it
<azonenberg>
But i have a 16 GHz TDR setup and my VNA only goes to 6 GHz
<azonenberg>
so getting even poorish quality s-params from a 16 GHz scope and pulse generator would be very useful
<Stephie>
yeah, i was wondering how accurate it was in practice
<Stephie>
yeah
<Stephie>
i might end up with one of the old HP boatanchor 30+ghz tdr/sampling scope test sets so
<azonenberg>
the lecroy wavepulser 40ix is a 40 GHz TDR that has fft transform to give s-parameters
<azonenberg>
it's very high on my wishlist
<azonenberg>
unless i get a 20 GHz copper mountain VNA instead
<azonenberg>
both are about the same price and 4 ports, its a question of more freq vs more dynamic range
<Stephie>
yeah
<azonenberg>
Leaning toward the TDR as most RF stuff i would be likely to do would be sub 6GHz and the picovna would handle that
<azonenberg>
i mostly do higher bw s-param stuff for digital interconnect characterization
<Stephie>
interesting
<azonenberg>
which needs less dynamic range but a ton of bandwidth for high speed serdes
<Stephie>
i actually have a TDR already but it's an old tek CRT "cable tester" unit i picked up for £80
<Stephie>
it's still really good though, you can see into the 0.1ohms, and sub-cm
<azonenberg>
this is the kind of data i get with the SDA 816Zi and a LeoBodnar pulse generator
<azonenberg>
100ps/div
<Stephie>
whats the rise time on that pulse gen?
<azonenberg>
i forget if i used the 30 or 40 ps unit for this test
<Stephie>
ah
<azonenberg>
the bnc is 40 and the sma is 30
<azonenberg>
the scope's rise time is i think 22?
<azonenberg>
or 28
<mxshift>
azonenberg: I've got a handful .s80p files that I can try when you get high port count parsing working
<azonenberg>
dont have specs handy
<Stephie>
the HP one im looking at is <18ps
<azonenberg>
mxshift: s80p? o_O
<azonenberg>
now i'm curious, is there a file format for sparse measurement of s-parameters?
<mxshift>
DDR signals between die and pin
<azonenberg>
i.e. where you want thru and reflected paths on each port plus crosstalk between adjacent ones
<azonenberg>
but maybe not crosstalk from one side of the package to the other
<azonenberg>
where the coupling is insignificant
<mxshift>
this appears to have full matrices
<azonenberg>
yeah touchstone only seems to support full matrices
<mxshift>
80MB files
<azonenberg>
that's why i'm wondering if there is one that isn't
<azonenberg>
side note, do you use zero prefixed port numbers when doing s-params of >9 port networks?
<azonenberg>
i.e. how do you tell whether S312 is the path from port 12 to port 3, or port 2 to port 31?
<azonenberg>
not a problem i've ever had to deal with given my wimpy little 2 port vna :p
<azonenberg>
also ooh
<azonenberg>
if anyone needs some fairly high speed s-param characterization needs in the future...
<azonenberg>
i just got a quote from TRS RenTelCo for a WavePulser 40ix
<azonenberg>
It's only $1350 to rent for a month
<azonenberg>
Until I get budget to actually buy one (which I would like to do eventually) I'm seriously considering renting one to do a bunch of test fixtures, test various SMA connector launches, and get serialized s-params for all of my cables so i can de-embed them to the full bandwidth of my scope
<azonenberg>
a month would be more than enough for that
<azonenberg>
No idea on timeline but $1350 is cheap enough i could probably swing it in the fairly near term
<azonenberg>
and more importantly it would get me a chance to play with the thing and decide if i want to buy it
<azonenberg>
mxshift: to load your s80p files the biggest change we would need, other than fixing the parser
<azonenberg>
would be making the import logic not creating a view for all 12,800 channels :p
<azonenberg>
(mag and phase for 80*80 ports)
<azonenberg>
by default for a s2p it opens everything
<azonenberg>
also just as important the dialog for selecting a channel to display, or as input to a filter, would become a nightmare
<azonenberg>
like, it would be technically possible to open the file as soon as i fix the parser
<azonenberg>
but it would be very hard to do anything useful with it for a while
<mxshift>
I figured. Hence why I mentioned I have them.
<azonenberg>
Yeah i think i will start with some more reasonable s4p's :p
<mxshift>
Using → to delineate ports makes sense to me: S21→4
<azonenberg>
mxshift: except that's backwards from the usual convention
<azonenberg>
it would have to be S4 <- 21
<azonenberg>
Legitimately curious how the conveniton got to be that way
<azonenberg>
Were S-parameters invented by a guy whose native language was arabic, hebrew, or something else that goes right to left?
<mxshift>
There's ← as well
<miek>
it comes from matrix notation
<azonenberg>
miek: what i mean is, having the matrix the way it is