azonenberg changed the topic of #scopehal to: libscopehal, libscopeprotocols, and glscopeclient development and testing | https://github.com/azonenberg/scopehal-apps, https://github.com/azonenberg/scopehal, https://github.com/azonenberg/scopehal-docs | Logs: https://freenode.irclog.whitequark.org/scopehal
Bird|otherbox has quit [Remote host closed the connection]
Bird|otherbox has joined #scopehal
maartenBE has quit [Ping timeout: 265 seconds]
maartenBE has joined #scopehal
juli965 has quit [Quit: Nettalk6 - www.ntalk.de]
<_whitenotifier-f> [scopehal] azonenberg edited issue #307: Retool Multimeter API to have GetMeterValue() and GetMeterUnit() methods - https://git.io/JTIZH
<_whitenotifier-f> [scopehal] azonenberg opened issue #309: Add secondary value support to multimeters - https://git.io/JTLNG
<_whitenotifier-f> [scopehal] azonenberg labeled issue #309: Add secondary value support to multimeters - https://git.io/JTLNG
<_whitenotifier-f> [scopehal] azonenberg labeled issue #309: Add secondary value support to multimeters - https://git.io/JTLNG
Degi has quit [Ping timeout: 260 seconds]
Degi has joined #scopehal
<_whitenotifier-f> [scopehal] azonenberg closed issue #272: Support for Tek 6 series DVM option - https://git.io/JUVqR
<_whitenotifier-f> [scopehal] azonenberg closed issue #307: Retool Multimeter API to have GetMeterValue() and GetMeterUnit() methods - https://git.io/JTIZH
<_whitenotifier-f> [scopehal] azonenberg pushed 3 commits to master [+0/-0/±19] https://git.io/JTLx5
<_whitenotifier-f> [scopehal] azonenberg 23b3d7e - Multimeter: now have GetMeterValue() rather than separate Get() functions for each type of measurement. Added method to query current unit. Fixes #307.
<_whitenotifier-f> [scopehal] azonenberg 70f4391 - Added significant digit support to Unit::PrettyPrint
<_whitenotifier-f> [scopehal] azonenberg c58522c - Added significant digit support to Multimeter class. Finished Tek DMM support. Fixes #272.
Pretzel4Ever has joined #scopehal
Pretzel4Life has quit [Ping timeout: 244 seconds]
electronic_eel has quit [Ping timeout: 246 seconds]
electronic_eel has joined #scopehal
<_whitenotifier-f> [scopehal] azonenberg pushed 1 commit to master [+0/-0/±1] https://git.io/JTteL
<_whitenotifier-f> [scopehal] azonenberg 4d3268e - Updated to latest graphwidget
<_whitenotifier-f> [scopehal-apps] azonenberg pushed 1 commit to master [+0/-0/±6] https://git.io/JTteq
<_whitenotifier-f> [scopehal-apps] azonenberg 80f4f60 - Initial multimeter display support. Fixes #232.
<_whitenotifier-f> [scopehal-apps] azonenberg closed issue #232: Oscilloscope multimeter support - https://git.io/JTInP
<azonenberg> test on tek mso6, about to test on the wavesurfer at the ioa lab remotely
<azonenberg> doesn't yet work with standalone DMMs but that's a pretty easy fix now that i have the UI
<azonenberg> i just have to make glscopeclient's command line arg parsing understand things that are meters but not also scopes
<lain> nice
<_whitenotifier-f> [scopehal] azonenberg opened issue #310: Support for Rigol MSO5000 integrated multimeter - https://git.io/JTtUW
<_whitenotifier-f> [scopehal] azonenberg labeled issue #310: Support for Rigol MSO5000 integrated multimeter - https://git.io/JTtUW
azonenberg_work has quit [Ping timeout: 256 seconds]
azonenberg_work has joined #scopehal
<azonenberg> lain: holy moley
<azonenberg> this is really cool
<lain> :o
<azonenberg> So i've been taking a lot of photos and collecting characterization data on my PP023 passive probe. The 10M 10:1 probe that came with my 2 GHz scope
<azonenberg> getting set up to write an article for xdevs about it (and eventually one for each of my other probes)
<azonenberg> One of the things I did was compensate the probe to get the best possible data off it
<azonenberg> So in addition to using the low freq compensation i decided to do the high freq comp too (two additional tiny trimmers)
<azonenberg> This uses a higher frequency signal than the normal 1 kHz cal output on the scope so i switched to the "fast edge" output on the aux port
<azonenberg> And decided to try out, for the first time, a handy dandy adapter that came with the probe
<azonenberg> basically it treats the probe as a male coax connector and adapts that directly to a push-on BNC
<azonenberg> so you can push the adapter onto a BNC signal generator output and stick a probe right into it
<azonenberg> lain: make sense?
<azonenberg> Guess what i found
<lain> oh hey I have those adapters for my agilent probes
<lain> what did you find :o
<azonenberg> no matter how much i tuned the compensation that little hump at about 13 ns didn't go away
<lain> waiiit a minute
<lain> is that
<azonenberg> So i started doing some math
<azonenberg> yes, the probe cable is 1.3 meters long
<lain> ehehehehe
<azonenberg> assuming 0.66*C velocity factor that gives you 6.57 ns one way or 13.14 ns round trip
<lain> the math checks out
<lain> so it's reflecting off the scope input..
<azonenberg> Yep. I just tried on a second probe, which i had never compensated
<azonenberg> its harder to see because this is undercomped, the edge is pretty round
<azonenberg> but if you zoom in it's there
<lain> ffs lol
<azonenberg> either the scope input or the compensation box
<azonenberg> (can you tell i dont use 1M probes much? lol
<lain> hehe
<azonenberg> But yes. Something at the scope side of the link, either the scope frontend or the probe compensation box/BNC termination, has a mismatch
<azonenberg> And when you probe a fast signal you see it
<azonenberg> I can think of no other plausible explanation for the data
<lain> I guess testing a simple coax between the two ports is a good test
<azonenberg> This is the same probe on a different test setup
<azonenberg> leobodnar pulse gen -> 3" semi rigid coax -> short bit of GCPW -> 50 ohm terminator
<azonenberg> (SMT 50 ohm resistor, not at the far end of the fixture)
<azonenberg> i used the probe w/ spring ground from end of the line to ground
<azonenberg> and that was the waveform. a bit of ringing on the edge from the spring inductance but the same reflection
<azonenberg> this is my curse, i'll never look at a scope probe the same again lol
<azonenberg> now that i've been working on this project i'm seeing every little non-ideality
<azonenberg> reuploaded
<lain> lol
<NeroTHz> moninb
<azonenberg> NeroTHz: o/
<azonenberg> I just finished writing an article on my 500 MHz LeCroy passive probe for xdevs
<azonenberg> and found an impedance mismatch in the probe and/or scope frontend while doing it
<azonenberg> (see above)
<NeroTHz> okay will look in a second
<azonenberg> Lots more notes and docs are on the xdevs docsite FTP under the LeCroy_PP023 directory. Waiting for TiN to look at it and hopefully it will go live shortly
bvernoux has joined #scopehal
<NeroTHz> just spent 2 days setting up cadence
<NeroTHz> but as a result now, I have a good workflow
<NeroTHz> my colleagues actually manually export layouts at every stage of the design and import them into the next tool
<NeroTHz> but now all of my tools can just read and interpret the same libaries, so I can just design and layout a transformer in ADS, simulate it in momentum, open it in HFSS/Clarity, hook it up to my big layout in Cadence, and run DRC, without manually ever having to deal with layermappings or exporting layout files
<azonenberg> Nice
<NeroTHz> less nice
<NeroTHz> my 50 gbit/second link paper got rejected because someone submitted a 100gbit/s link to the same conference
<azonenberg> :(
<monochroma> :<
<NeroTHz> the frustrating thing
<NeroTHz> is that my current design also aims at 100 gbit
<NeroTHz> so it´s already not state-off-the-art by the time it´s done
<NeroTHz> so now I either have to make the choice to push and go wild, or to really just focus on novelty (and try to get my hands on an early-access version of that paper to see what they do so I d something different)
<azonenberg> hey, on the plus side it was rejected for a good reason
<azonenberg> I've had papers rejected for absolutely nonsensical reasons
<azonenberg> like "too technical" or some other nonsense
<azonenberg> like, if you want to criticize my work on technical merits, or believe it's flawed, that's one thing
<NeroTHz> I mean I kinda would prefer that it was a nonsensical reason because then I could go ´those damn shitty reviewers!´ *angry fist shake*
<NeroTHz> :p
<NeroTHz> but yeah you are right
<azonenberg> but if you submit a review that makes it clear you don't understand the paper, or barely read it
<azonenberg> that tells me nothing about how to improve it for next time
<azonenberg> That was one of my favorite things about submitting my thesis work to CHES
<azonenberg> Some of the reviews were pretty critical (although it was ultimately accepted)... but
<azonenberg> it was clear from the content that they UNDERSTOOD my work
<azonenberg> at several previous cons i had been rejected from, it was clear that it went over their head
<NeroTHz> LOL
<NeroTHz> I was typing ´substrate=1mm´ into clarity EM
<NeroTHz> and it was just ´nope´ and crashed, no error message or nothing, just gone
<azonenberg> NeroTHz: so guessing you didnt get a chance to look at the probe data i linked earlier?
<azonenberg> lol
<NeroTHz> oh that is interesting
<NeroTHz> so stuff not matched
<NeroTHz> hmm
<NeroTHz> man this is such a frustrating bug in clairty
<azonenberg> NeroTHz: yeah. It's subtle
<azonenberg> but it's measurable and consistent across two different probes
<azonenberg> this is my curse, as soon as you start seeing non-idealities in probes you can't ever unsee them
<azonenberg> a year or two ago my brain would have just blamed that on the test signal not being an ideal squarewave
<NeroTHz> Now imagine what I go through having to simulate 200 GHz stuff based on models that, on a good day, got extrapolated from DC-10 GHz measurements
<azonenberg> now, i immediately recognize it as a reflection and within seconds i had guesstimated the propagation delay of the cable and determined the likely cause of it as a mismatch at the probe-scope interface
<NeroTHz> but yes, once oyu start noticing those things you can´t unsee them
<azonenberg> and lol, yeah i imagine PDKs probably aren't characterized that high
<azonenberg> This is going to be a fun little writeup series
<NeroTHz> I have the same with some stuff in (totally unrelated field) photography and videography
<azonenberg> I'm doing reviews on xdevs of all my probes, both 1M and transmission (and eventually, active)
<NeroTHz> where I get why certain reality TV things are cut a certain way or whatever
<azonenberg> lol
<azonenberg> i remember watching a movie at one point and seeing visible MPEG artifacts during a pre-roll advertisement video
<azonenberg> the 8x8 pixel boundaries were noticeable
<azonenberg> as was ringing around some of the sharp edges
<NeroTHz> (or the frustrating thing where the host of the show visits the subject of that episode at home, and you see the subject open the door and be surprised and go ´Oh I wasn´t expecting you, mr&mrs tv host!´ */from a camera perspective within the fucking house*/
<azonenberg> Lol
<monochroma> XD
<NeroTHz> Also got really worked up on continuity error in some house-hunter program
<NeroTHz> where you could see, as the host and the guests went through the house tour, in the background the outside went from sunlight to streetlights to sun to streetlights
<azonenberg> lol
<NeroTHz> (My girlfriend also got annoyed at my need to point out that in the hunger games movies, you could tell when a background was real or CG by the fact that the real movie was shot on anamorphic, but the fake-background-blur they added to the CG shots was spherical
<monochroma> :P
<NeroTHz> (I was genuinly proud that I could tell lol)
<NeroTHz> also gets well into cinematography-nerd-level before people even get why I think that is cool
<azonenberg> lol
<azonenberg> NeroTHz: btw, any idea if there's a feasible way to VNA a 1M ohm probe?
<azonenberg> do they make VNAs with 1M inputs? :P
<NeroTHz> what do you mean feasible way?
<azonenberg> Given a 50 ohm VNA if I want S21 of a 10:1 passive scope probe
<NeroTHz> you can just probe with regular 50 ohm VNA, and then do a port-translation in simulation
<azonenberg> That works fine for things like 40 or 60 ohm lines
<azonenberg> my gut feeling is that 1M is far enough off you'll have dynamic range issues
<azonenberg> i could be wrong
<azonenberg> the other issue is i think those probes expect a specific capacitance to ground at the scope side
<NeroTHz> could be. I don´t have any direct reference for accuracy/dynamic range requirements
<azonenberg> since it's a R||C divider
<NeroTHz> you should even be able to include that by adding that cap to your s2p file in simulation
<azonenberg> i'm pretty sure you can't just throw a 50 ohm termination on it and de-embed
<NeroTHz> you shuold do a 2-port measurement
<azonenberg> hmm, i guess if you did a full sim of the scope frontend on the far side
<azonenberg> it would be possible
<azonenberg> I'll try it in a bit i guess and see if i get something somewhat plausible
<bvernoux> hello
<bvernoux> finally I have bought the MSO5074 ;)
<bvernoux> thanks to Batronix ;)
<bvernoux> (they offered a 6% discount)
<bvernoux> as I have bought other stuff with it
<bvernoux> finally my list is
<bvernoux> 1x Rigol MSO5074
<bvernoux> 1x Sensepeek 4016 PCBite 2x SP200 Kit
<bvernoux> 1x Sensepeek 4x SP10 Probe Kit 4005
<bvernoux> 1x TekBox TBPS01-TBWA2/40dB
<bvernoux> I have not found a good Amplifier up to 40dB so I have take the expensive full version of TekBox ;)
<bvernoux> like that I could wait Tek MSO64B refurbished with 3 year warranty is available ;)
<bvernoux> it is the next step ;)
<bvernoux> also asked to Lecroy if they plan to provide something equivalent (in specification) to the Tek MSO64B
<bvernoux> IIRC they do not have anything equivalent so far
<azonenberg> bvernoux: i believe the closest competitor they have is the WavePro HD
<azonenberg> Which is 4 channels max, 8 GHz bandwidth, 20 Gsps
<azonenberg> it would not surprise me if they introduced a 6-8 channel version at some point
<azonenberg> There is no digital downconversion/spetrcum mode other than the software FFT, i believe because tek has a patent on that
<bvernoux> yes but it is 8bits ADC also
<bvernoux> not 12bits
<azonenberg> No
<bvernoux> you shall not check the ENOB anyway ;)
<azonenberg> The WavePro HD is 12 bits
<bvernoux> ha ok
<azonenberg> But it tops out at 20 Gsps and 8 GHz
<bvernoux> so yes it is the nearest competitor
<azonenberg> It's their fastest 12 bit scope
<bvernoux> but MSO 64B is 50GSPS ;)
<bvernoux> so 50GSPS vs 20GSPS is a big improvement
<azonenberg> the fastest 8 channel scope is the WaveRunner 8000HD which is 2 GHz 10 Gsps max but 8 channels
<bvernoux> especially they can upgrade the BW from 1GHz to 10GHz only by SW license
<azonenberg> if you want more sample rate you have to move to the WaveMaster line, which goes up to 30 GHz 80 Gsps, but is 4 channels only and 8 bit
<azonenberg> then LabMaster you're already familiar with is the crazy high end line
<azonenberg> There has not been a refresh on those models in a couple of years. Hopefully a new one comes out soonish
<bvernoux> on my side I'm not really interested by 8chan just 4chan is perfect ;)
* azonenberg sends bvernoux off to 4chan to hang out with the trolls there
<bvernoux> hehe
<bvernoux> anyway more is always better to push them to innovate ;)
<bvernoux> next ADC is maybe with saphire substrate ;)
<NeroTHz> fking all this microwave design software using a ideal ground
<NeroTHz> I can´t get my head wrapped around it
<NeroTHz> how can you do anything if you can´t take into account return current
* NeroTHz angry-fist-shakes
<bvernoux> we need that for future oscilloscope https://twitter.com/everythingrf/status/1316046254674259968
<bvernoux> ;)
<miek> and it's samtec, so it'll probably cost the same as the scope :P
<bvernoux> yes clearly ;)
<bvernoux> it is for 70GHz badnwidth oscilloscope up to 16chan ;)
<bvernoux> bandwidth
<bvernoux> for the 6G ;)
<bvernoux> I still do not see how they will sell 5G to people which have already too much with 4G ;)
<bvernoux> the most funny is in some country they have 3G/4G with the speed of 2.5G ;)
<bvernoux> so much BS marketing when operator cannot even warranty a minimum speed ...
<bvernoux> the only great things (to be checked) is latency of 5G which is defined as Typical Latency 1ms
<bvernoux> it is theoretical ;)
juli965 has joined #scopehal
bvernoux has quit [Quit: Leaving]
lain has quit [Ping timeout: 260 seconds]
vup has quit [Ping timeout: 260 seconds]
vup has joined #scopehal
lain has joined #scopehal