azonenberg changed the topic of #scopehal to: libscopehal, libscopeprotocols, and glscopeclient development and testing |,, | Logs:
Degi has quit [Ping timeout: 256 seconds]
Degi has joined #scopehal
<_whitenotifier-b> [scopehal] azonenberg opened issue #206: Allow signal generator IBIS file name / model name to be configured -
<_whitenotifier-b> [scopehal] azonenberg opened issue #207: Allow signal generator sample/symbol rate to be configured -
electronic_eel has quit [Ping timeout: 260 seconds]
electronic_eel has joined #scopehal
electronic_eel has quit [Ping timeout: 264 seconds]
electronic_eel_ has joined #scopehal
<_whitenotifier-b> [scopehal] azonenberg opened issue #208: When a channel is renamed, filters that use it as an input should also be renamed if they are using default names -
<_whitenotifier-b> [scopehal-apps] azonenberg labeled issue #154: Allow hidden filters to be added to display -
<_whitenotifier-b> [scopehal-apps] azonenberg opened issue #154: Allow hidden filters to be added to display -
<azonenberg> New demo video - showing off de-embedding and channel emulation
<azonenberg> lain, monochroma, kliment: ^
<lain> :3
<Kliment> azonenberg: embedding and channel cancellation :P
<azonenberg> Kliment: lol
<Kliment> Seriously nice though
<Kliment> I'm learning to VNA for an antenna design
<azonenberg> Kliment: for reference, this is the sort of thing that scope vendors normally charge four or five digits for
<azonenberg> lol
<Kliment> I know :)
<Kliment> ohnoes, we have to do math! Math costs money!
<Kliment> I mean, I get their logic - they can't sell base models for a high price because competition, so they try to split out as much as possible into addons
<azonenberg> Yeah
_whitelogger has joined #scopehal
katharina has joined #scopehal
<katharina> o/
<azonenberg> o/ katharina
<azonenberg> glad to see you back
<azonenberg> Hope you're recovering well?
<katharina> its slow but steady, im now at a point where i can take steps without breaking into a massive sweat
<azonenberg> Sounds unpleasant but at least heading in the right direction
<katharina> exactly, but on the other hand you can notice daily improvements
<azonenberg> Yay
<azonenberg> as you can see from my tweets we've been busy here :)
<katharina> yes! also from all the new contributors fixing issues in the protocols etc
<katharina> awesome how the project has grown
<azonenberg> Yeah it's definitely building momentum
<azonenberg> I was chatting with Shahriar the other day about having him do a signalpath video on it at some point once it's a bit more polished
<katharina> oh wow, i love his channel
<azonenberg> That should be good advertising, but it's not to that point yet
<azonenberg> My scope probe will be getting a review there soonish though
<katharina> thats awesome! ill be definitely watching that video
<azonenberg> i'm starting the production assembly of the kickstarter probes on monday
<katharina> does the software still build on windows?
<azonenberg> all of the boards and components are here and i did one
<azonenberg> but waiting for my lab tech to be available to help me with the big run
<azonenberg> No. There's a few regressions in some of the recent code, but it should be trivial to fix
<azonenberg> i actually have a working copy with some of those fixes
<katharina> i think that would be a good thing for me to get back into programming if there are fixes left to do
<katharina> i also have been writing some verilog yesterday and that has been going ok
<azonenberg> Not sure if you saw my tweet, but I sold my waverunner 8104 back to LeCroy as a trade-in
<azonenberg> and replaced it with a shiny new (to me) WaveRunner 8404M-MS, which is the same family but quite the upgrade on specs
<katharina> oh i saw the one where you had glscopeclient running on it
<azonenberg> 10/20 to 20/40 Gsps, and 1 -> 4 GHz
<azonenberg> and yes
<azonenberg> i got glscopeclient to build and run on it. Which is quite the milestone :D
<azonenberg> But i had to comment out a few things and patch a few tidbits
<katharina> jesus that are some beefy spec upgrades
<azonenberg> this is what happens when our windows maintainer gets sick :p
<azonenberg> katharina: yeah but for the kinds of work i do it's totally necessary - in fact, not even enough
<azonenberg> but closer to what i need than what i had before
<azonenberg> it's my only scope for the next week or two
<azonenberg> My other LeCroy, the HDO9204, is the same sample rate but only 2 GHz bandwidth (10 bit resolution though). It's out sick at the moment
<azonenberg> Channel 4 had a really weird malfunction at high gain settings in 50 ohm mode where it appears to be AC coupling instead of DC
<azonenberg> I sent it back to the dealer i bought it from, they confirmed it failed their internal cal and that whatever was wrong was beyond their ability to diagnose
<katharina> i hope ill manage to have some time to get into RF some day, it seems like such an interesting field
<azonenberg> So it's now back at LeCroy's lab for further debug and repair
<katharina> im just sitting here with my FPGA solutions at work running at 100 MHz tops lmao
<azonenberg> Lol. I'm mostly a high speed digital guy, i got dragged kicking and screaming into RF
<azonenberg> because i got into high speed networking and such, and *sob* my wires aren't wires anymore
<azonenberg> and my signals don't transition instantly and travel down ideal conductors anymore
<azonenberg> in zero time
<azonenberg> So i'm being forced to learn all of the stuff i need to make my stuff faster, and then getting into test equipment is making me learn even more analog and DSP
<azonenberg> but ultimately the goal is still building fast digital stuff
<katharina> :)
<azonenberg> I just need to do analog stuff to debug it :p
<azonenberg> Anyway, so if you're feeling up to it priority #1 for you is the handful of tweaks to get my recent changes to compile on windows again
<katharina> sure i think i wanna do that right about now after i devoured this cold pizza from yesterday
<azonenberg> Longer term a windows CI system that logs to this channel might not be a bad idea, so i can get immediately notified if a commit breaks the build on a platform other than the one i'm developing on
<azonenberg> I might still expect you to *fix* the problem in that case, but at least we'd know about it right away
<azonenberg> Your #2 priority is to get the preferences dialog and surrounding stuff to the point it's ready to get merged
<azonenberg> We have quite a few pending things that will require creation of preferences
<azonenberg> More so than the startup dialog and other ui-dev work. While very nice, that work isn't blocking anyone else and the preferences dialog is
<azonenberg> So i want that in master ASAP so we can all start building on it :)
<katharina> gotcha
<katharina> ill get the windows build working today, lets see how exhausting that is for me
<azonenberg> Good luck
<katharina> i think its a good start to get back into doing actual work
<azonenberg> Glad to have you back
<noopwafel> \o/
<azonenberg> ok so, does anybody have pressing priorities for me to work on tonight?
<azonenberg> Probably going to be spending most of the weekend getting MAXWELL ready to go, and doing non-test-equipment related work around the house etc. And cable plant work in the lab
<katharina> azonenberg: could you send me a diff of the windows issues you fixed in your local working copyß
bvernoux has joined #scopehal
<azonenberg> katharina: it was one usleep and some #include paths not resolving right on my msys
<azonenberg> i think that was it?
<azonenberg> it should be pretty trivial if you just pull latest and compile
<azonenberg> iirc my msys wanted ffts/ffts.h not ffts.h
<azonenberg> but that may have just been me installing it wrong :p
<bvernoux> azonenberg, very nice your latest video with DeEmbedded
<bvernoux> and Emulation
<bvernoux> will be nice to extend glscope to an other tool glvna ;)
<azonenberg> VNA will come after specan support
<bvernoux> ha great
<azonenberg> if ever
<azonenberg> While i want to add vna support to scopehal it will likely be a separate UI
<azonenberg> similar to power supplies and DMMs
<bvernoux> the deEmbedded stuff will be welcome for the XYY view in glvna ;)
<bvernoux> also a way to compare sxp file
<bvernoux> like PocketVNA
<azonenberg> but i guess you could use glscopeclient for that
<azonenberg> yes there is a lack of good s-parameter viewers out there
<azonenberg> like i mentioned i mostly use sonnet for that
<bvernoux> so far there is not any good GUI for VNA SxP files
<bvernoux> all suxx ;)
<azonenberg> Re specan support, the big holdup is that i dont own a specan :p
<bvernoux> yes but sonnet is ugly for sxp files ;)
<azonenberg> what's ugly about it?
<bvernoux> the display
<azonenberg> i like it quite a lot, more than anything else i've used
<bvernoux> and the GUI is awful to zoom in/out a...
<azonenberg> i mean i wouldn't buy it for the viewer but if i have it, i'll use it as my general sxp viewer until i find something better
<bvernoux> test PocketVNA you will understand ;)
<bvernoux> it is the best so far
<bvernoux> but it does not support DeEmbedding ...
<bvernoux> It is why glVNA will be welcome to rule them all
<bvernoux> adding relatime and offline mode
<bvernoux> realtime
<bvernoux> especially if it can support GPIB instrument over USB...
<bvernoux> or in worst case over Ethernet ;)
<bvernoux> I will check If I can buy a GPIB Ethernet bridge ;)
<azonenberg> A SCPIGPIBTransport driver is on the to-do list
<azonenberg> tverbeure was talking about making one then dropped off the channel
<azonenberg> if anybody here has a usb gpib dongle and a gpib-capable instrument, feel free to work on it
<bvernoux> something using libusb done from scratch ?
<bvernoux> as everything is proprietary ...
<bvernoux> same issue as using PicoScope over USB 3 ...
<noopwafel> pico libraries are easy enough ;p
<bvernoux> all shall be reversed and rewritten with libusb which is a big task
<azonenberg> is there no standard linux driver stack for gpib?
<azonenberg> like usb serial
<azonenberg> i figured it would be a commodity character device interface that we could just use
<bvernoux> I doubt a linux driver stack for gpib exist but who know
<bvernoux> in theory it is just a serial port
<noopwafel> it does, and there is sw too
<bvernoux> noopwafel, there is no open source lib for Picoscope
<bvernoux> Picoscope lib are binary blob
<noopwafel> bvernoux: I mean for gpib
<bvernoux> ha ok ;)
<noopwafel> for picoscope, you can just use picoscope libs :)
<bvernoux> so it is a very good point for gpib as tons of instrument use it
<bvernoux> even if it seems obsolete I cannot live without that on my old instrument ;)
<bvernoux> noopwafel, using picoscope libs was a no go for glscope ..
<azonenberg> bvernoux: why?
<bvernoux> as it will maybe break the license of glscope
<bvernoux> to be checked
<azonenberg> as long as it's not copyleft, it shouldn't be a problem
<azonenberg> and they're blobs
<bvernoux> if it does not break the license it break the concept to have something fully open source ;)
<azonenberg> Yes, that much i agree with
<azonenberg> but i'd rather have a half-blob solution than none
<bvernoux> IIRC it is something like copyleft ...
<azonenberg> i mean i would love to eventually reverse the protocol and write a fully open driver stack
<noopwafel> so I have a hacky SCPI bridge for now
<azonenberg> yeah that is a good solution in general to keep it at arms length
<bvernoux> azonenberg, A must is to use libusb I think for all usb stuff
<bvernoux> azonenberg, it is portable and work fine on all platform
<noopwafel> I was going to try adding libusb for the vds1022 yesterday
<noopwafel> but instead I spent the afternoon trying to work out why my vds1022 no longer works
<bvernoux> noopwafel, it is more problematic if you hw does not work ...
<bvernoux> noopwafel, it seems the vds1022 is a sort of Picoscope clone ;p
<noopwafel> yeah
<bvernoux> 10M record length is short
<bvernoux> does it support streaming ?
<noopwafel> the vds1022 itself is a toy: it is 100msps and it has no RAM, so you are limited to block RAM on the FPGA, which is like 5k
<bvernoux> I love the Picoscope3000 for that over USB3
<bvernoux> I have bought it for that ;)
<bvernoux> realtime streaming at up to 250MSPS over USB3
<bvernoux> using API of course ;)
<noopwafel> and no, it is no picoscope replacement :p
<bvernoux> I imagine it is not the same price too
<bvernoux> Pico3000 is only 1GSPS too
<bvernoux> and only 8bits ;)
<noopwafel> but the vds1022 is ~$60 sometimes, and it has hw trigger, so you can leave it permanently attached to long-term setups which is convenient sometimes
<bvernoux> But good for small signal stuff not exceeding 20MHz
<noopwafel> the 1gsps ones are way more serious devices, but they cost real money, and I don't think they are worth it
<bvernoux> what is missing on Pico3000 is 50Ohms input
<bvernoux> or a way to switch from 50Ohms to 1MOhms like on lot of scope ...
<noopwafel> but then what would you do with all your pico 50ohm terminators :)
<bvernoux> yes I use one ;)
<bvernoux> but I will prefer to have it inside ;)
<noopwafel> well, if you buy a fancy picoscope with the 50ohm input, I am interested to hear how they are
<bvernoux> Waiting a good offer for a "cheap" 20GSPS scope ;)
<bvernoux> something like what azonenberg have ;)
<bvernoux> it is a must to have warranty for manufacturer with a calibrated scope especially at that price / performance
<bvernoux> -for+from
<bvernoux> as Tek6 Series for less than 10KUSD is not for tomorrow ;)
<bvernoux> especially the Tek6 with 8GHZ BW ;)
<bvernoux> hehe simulation of Tek6 with 8GHz BW option +add 75KEuros ;)
<bvernoux> at the end it is >100KEuros ;)
<bvernoux> RackMount option is >1Keuros ;)
<bvernoux> Option to have Removable SSD with WIndows License 993Euros !!
<bvernoux> crazy ...
<bvernoux> each option are crazy
<azonenberg> bvernoux: well that is basically targeting the defense world
<azonenberg> when you want to remove the drive before sending it in for service
<bvernoux> 8B10 protocol decoder => 2720Euros ;)
<azonenberg> so of course it's expensive, the customer is the military
<azonenberg> lol
<azonenberg> and this is why i wrote one :p
<bvernoux> yes
<bvernoux> but they should have price for normal lab ;)
<bvernoux> price for military big industry and for others ;)
<bvernoux> I plan to do that for my products
<bvernoux> Price for students/reasearcher/very small company) and price for others
<bvernoux> woo the only free things is Power Options => Option A1 - Universal EURO (add 0 Euros) ;)
<bvernoux> Calibration Data report also is 9Euros ;)
<bvernoux> at the end of the quote they do not want to say price ;)
<bvernoux> as it exceed 100KEuros ;)
<bvernoux> I will need to do a Flipper Zero Kickstarter to buy it ;)
<azonenberg> lol
<azonenberg> If i ever had $10K to spend on a scope
<bvernoux> they are near to reach 1MEuros now
<azonenberg> $100k*
<bvernoux> totally crazy
<azonenberg> it better have waaaaaay more than 8 GHz bw
<bvernoux> and the trends is 16MUSD for Flipper Zero ;)
<azonenberg> like. at that price i could almost buy a new labmaster
<bvernoux> azonenberg, yes clearly for 100KUSD it will be more than 25GSPS & 8GHz BW ;)
<bvernoux> Especially when you know it is just SW
<azonenberg> bvernoux: The 20 GHz 80 Gsps LabMaster is MSRP $114.3K
<bvernoux> when you have bought the 25KUSD version 1GHz BW it is the same HW as the 8GHz BW ;)
<azonenberg> So $100K for just the base scope is a bit nuts
<azonenberg> and of course, secondhand you can get waaaaay more for your dollar
<bvernoux> ha yes the LabMaster is clearly more intersting vs Tek6
<bvernoux> But it does not support mixed domain ;)
<azonenberg> if i was going to buy a labmaster i'd buy the 36 GHz one on ebay for $52K :p
<azonenberg> Not a new one for $114K
<azonenberg> with only 20 GHz bw
<bvernoux> yes but it is a big risk for 52KUSD
<bvernoux> if something wrong happen you have loose 52KUSD ...
<azonenberg> Of course. And I don't have $52K to spend on it either
<bvernoux> like it was the case for me with my Agilent E44405B ;)
<bvernoux> but it cost less ;)
<bvernoux> but I was not far to loose 3KUSD
<bvernoux> I was lucky it was directly defective
<bvernoux> and not after 2month of use
<bvernoux> also I have received some SOuthWest conenctors
<bvernoux> Brand NEW !!
<bvernoux> 2.92mm
<bvernoux> the packing is so awfull half are dead ...
<azonenberg> lol
<bvernoux> they was unscrewed
<bvernoux> and center pin is dead
<bvernoux> !!
<azonenberg> wow
<bvernoux> what a shame on this vendor
<azonenberg> i wonder just how much expensive lab equipment is destroyed in shipping each year by clueless ebay sellers
<bvernoux> yes clearly just because of very bad packing all in same bag unswrewed
<azonenberg> (i assume you complained and asked for a refund?)
<noopwafel> best not to ask these questions
<bvernoux> when you see the center pin for PCB trace it is so fragile
<bvernoux> not protected
<bvernoux> Maybe I can repin them as the connectors are very nice
<bvernoux> in fact it was item lost ...
<bvernoux> which I have received 3 month after
<azonenberg> also ooooh rigol is stepping up their game
<bvernoux> else I will have paid the guy but all is broken
<bvernoux> azonenberg, ha ther is some new rigol scope ... ?
<azonenberg> Not a scope
<bvernoux> SA ?
<azonenberg> The DG5352 is 350 MHz, 14 bit, 1 Gsps
<azonenberg> AWG
<bvernoux> woo
<bvernoux> nice
<azonenberg> With I/Q modulation including QSPK and QAM
<bvernoux> 10KEuros ;)
<azonenberg> 14.1K list price actually
<bvernoux> yes very nice features
<azonenberg> I want to build something better :p
<noopwafel> also slightly more affordable?
<bvernoux> it is very nice
<bvernoux> not like the new scope
<bvernoux> seems very compact too
<azonenberg> noopwafel: I want to put an AD9154 and an xc7k160t in a 1U system
<azonenberg> up to 1 Gsps data rates with interpolation to 2.4 Gsps, on four channels at 16 bits
<noopwafel> right, this would be great
<bvernoux> the fun stuff look the family ;)
<bvernoux> DG5071
<bvernoux> is also 1GSA/S
<bvernoux> most interesting is 2 chan
<bvernoux> DG5072
<bvernoux> which can probably hacked to be a DG5352 ;)
<bvernoux> with exactly same HW
<azonenberg> noopwafel: it would basically be maxwell's fpga and ram subsystem, but only gig-e since i'd be using all of the GTX's as JESD204B
<azonenberg> i think i could hit 500 MHz nyquist bandwidth although i dont know if performance would be great that far
<azonenberg> also it's very affordable. i wouldnt need rogers since it wouldnt have a lot of long high speed traces to worry about, probably just fr408hr
<bvernoux> it is challenging to reach 500MHz
<azonenberg> the pcb wouldnt be nearly as wide as maxwell either
<azonenberg> so i could probably build it in house
<bvernoux> yes would be an amazing instrument
<azonenberg> The AD9154 is only $121 at digikey too
<bvernoux> a must will be an option to stream data over Ethernet Gb....
<azonenberg> the fpga will be more than that, probably the same one as maxwell
<azonenberg> bvernoux: yes but only 1g
<azonenberg> no 10g
<azonenberg> because the fpga i'm looking at only has 8 serdes lanes and i need all of them for the DAC
<azonenberg> anyway i think that will be my next project after maxwell
<azonenberg> with the scope side of things running in parallel with the non-scope efforts
<bvernoux> see you
<katharina> damn it, i have to reinstall windows for some reason, it crashed during update and now its broken, why is that still ä thing
<azonenberg> lol
<azonenberg> because windows never had a proper package manager?
<katharina> yea
<azonenberg> it doesnt even have kernel versioning for fallbacks etc
<katharina> when this happens in my linux, i can just chroot into it and fix the packages that failed to install
<azonenberg> you have one ntoskrnl.exe
<azonenberg> it either works or doesnt
<katharina> yea its super annoying
<katharina> one thing that is super annoying is that on my thinkpad, windows cant use USB3.0 devices during install
<katharina> i assume its missing the drivers, but i havent seen something like that before
<azonenberg> you mean not just "dont work at 5 Gbps" but "don't work at all"?
electronic_eel_ is now known as electronic_eel
<azonenberg> monochroma: hey when you get a chance can you tell me the legal memory depths your DDA has on the upper end?
<azonenberg> i see 24M on 4ch or 48M points on 2 as the max in the datasheet
<azonenberg> but normally i would expect there to be a 20M and/or 40M point option
<azonenberg> is that present or not?
<azonenberg> basically anything past 10M points i'm curious about
<monochroma> azonenberg: there are different depth options and i'm not 100% sure what all is available
katharina has quit [Ping timeout: 240 seconds]
<azonenberg> monochroma: that's why i said "when you get a chance" meaning actually turn it on and poke the settings :p
<azonenberg> Heads up everyone, two new virtual methods being added to Oscilloscope for configuring channel combining
<azonenberg> i'll be supplying default implementations in all of the drivers we have so far in master
<_whitenotifier-b> [scopehal] azonenberg pushed 1 commit to master [+0/-0/±20]
<_whitenotifier-b> [scopehal] azonenberg ad28e35 - Initial channel interleaving support. Only implemented for LeCroy so far. Fixes #23.
<_whitenotifier-b> [scopehal] azonenberg closed issue #23: Add APIs for configuring channel interleaving -
<_whitenotifier-b> [scopehal-apps] azonenberg closed issue #11: Add UI for interleaving channels -
<_whitenotifier-b> [scopehal-apps] azonenberg pushed 1 commit to master [+0/-0/±4]
<_whitenotifier-b> [scopehal-apps] azonenberg 42ba0eb - Added UI for turning channel interleaving on/off. Fixes #11.
<azonenberg> So according to my scope, the rise time of this pulse generator is 154.39 attoweeks
<azonenberg> or 1.08 femtodays
<azonenberg> those are actually units it can be configured to display, lol
<_whitenotifier-b> [scopehal] azonenberg opened issue #209: Rename TouchstoneParser class to reflect that it's really a generic S-parameter class -
<_whitenotifier-b> [scopehal] azonenberg labeled issue #209: Rename TouchstoneParser class to reflect that it's really a generic S-parameter class -
<_whitenotifier-b> [scopehal] azonenberg labeled issue #203: Rewrite header comments to say "libscopehal" and not "Antikernel" -
<monochroma> azonenberg: well what i mean is there are different memory options, and i do not know what they all are, iirc i ran into an issue of lack of access to old sales materials/documentation for what the max memory option was
<monochroma> i will look again though
<monochroma> i can obviously tell you what's in my scope, but that's not the only memory option for these
<_whitenotifier-b> [scopehal] azonenberg pushed 2 commits to master [+0/-0/±4]
<_whitenotifier-b> [scopehal] azonenberg 4ef437a - ProtocolDecoder: fixed file load/save of "filenames" parameter type. Fixes #205.
<_whitenotifier-b> [scopehal] azonenberg d9413b2 - DeEmbedDecoder: don't crash with null deref if S2P file couldn't be found
<_whitenotifier-b> [scopehal] azonenberg closed issue #205: Channel emulation / de-embed filters segfault when loading from file (null touchstone data) -
<azonenberg> monochroma: yeah i wanted to use your config as a starting point
<_whitenotifier-b> [scopehal] azonenberg edited issue #187: Add CTLE filter -
<_whitenotifier-b> [scopehal] azonenberg opened issue #210: Add DFE filter -
<_whitenotifier-b> [scopehal] azonenberg labeled issue #210: Add DFE filter -
<_whitenotifier-b> [scopehal] azonenberg opened issue #211: Add FFE filter -
<_whitenotifier-b> [scopehal] azonenberg labeled issue #211: Add FFE filter -