azonenberg changed the topic of #scopehal to: libscopehal, libscopeprotocols, and glscopeclient development and testing |,, | Logs:
<azonenberg> lain: sooo i just did some *very* interesting measurements
<azonenberg> This is input impedance of each of my probes after properly de-embedding the test fixture
<azonenberg> Starting with the worst, the brown one is a stock 10M ohm 500 MHz passive probe
<azonenberg> At 500 MHz it has an impedance of 30 ohms :p
<azonenberg> Next worst is pink, the pico TA061. At 930 MHz it's 54 ohms
<azonenberg> Removing those because they were cluttering the view, we have
<miek> ooh nice, those plots are looking less confusing :)
<azonenberg> Cyan is the lecroy zs1500 active probe
<azonenberg> Black is the kickstarter AKL-PT1 v1.0 which, in retrospect, sucks :p
<azonenberg> But more interestingly, red is AKL-PT1 v1.3 and blue is the pico 921
<azonenberg> We're *very* closely matched
<lain> ooo
<lain> indeed
<azonenberg> except for that weird hump in the 921 around 2.5-4.5 GHz
<azonenberg> which is right about where the S21 curve tanks
<azonenberg> But more to the point look at the improvement from the black to the red
<azonenberg> lol
<azonenberg> I'm also going to be recharacterizing some of my probes for S21 with the fixture de-embedded, because now i have a reasonably good model for the loss of the fixture
<azonenberg> And on wednesday when the 4 GHz active diff probe comes in of course i'll be slapping that on the vna too
<Degi> Oh that's nice
deltab has quit [Ping timeout: 264 seconds]
deltab has joined #scopehal
Degi has quit [Ping timeout: 258 seconds]
Degi has joined #scopehal
_whitelogger has joined #scopehal
Pretzel4Ever has quit [Ping timeout: 256 seconds]
<_whitenotifier-f> [scopehal] azonenberg pushed 3 commits to master [+0/-0/±6]
<_whitenotifier-f> [scopehal] azonenberg e657d31 - CANDecoder: implemented parsing of extended frames. See #32.
<_whitenotifier-f> [scopehal] azonenberg da307e5 - CANDecoder: Added detection of FD mode bit. Can't actually decode FD-mode frames, but we can at least identify them. See #334.
<_whitenotifier-f> [scopehal] azonenberg bfd990c - Added protocol analyzer view for CAN packets. Fixes #32.
<_whitenotifier-f> [scopehal] azonenberg closed issue #32: Add CAN protocol decoder -
<azonenberg> Ok, i think we have functional CAN decoding now of both base and extended frames
<azonenberg> no CAN-FD but that shouldnt be too hard to add down the road
_whitelogger has joined #scopehal
electronic_eel has quit [Ping timeout: 260 seconds]
electronic_eel_ has joined #scopehal
electronic_eel_ is now known as electronic_eel
juli966 has joined #scopehal
katharinawork has joined #scopehal
<katharinawork> o/
katharinawork has quit [Ping timeout: 268 seconds]
<_whitenotifier-f> [scopehal] azonenberg opened issue #335: EtherCAT protocol decode -
<_whitenotifier-f> [scopehal] azonenberg labeled issue #335: EtherCAT protocol decode -
lukego has quit [*.net *.split]
lukego has joined #scopehal
bluecmd[m] has quit [Ping timeout: 240 seconds]
promach3 has quit [Ping timeout: 240 seconds]
jevinskie[m] has quit [Ping timeout: 240 seconds]
bvernoux has joined #scopehal
bluecmd[m] has joined #scopehal
jevinskie[m] has joined #scopehal
bluecmd[m] has quit [Quit: Bridge terminating on SIGTERM]
jevinskie[m] has quit [Quit: Bridge terminating on SIGTERM]
bluecmd[m] has joined #scopehal
bvernoux1 has joined #scopehal
bvernoux has quit [Ping timeout: 244 seconds]
jevinskie[m] has joined #scopehal
promach3 has joined #scopehal
<azonenberg> Next iteration AKL-PT2 prototoype is en route
<azonenberg> Enclosure, I mean
<azonenberg> not a new PCB spin
juli966 has quit [Quit: Nettalk6 -]
<_whitenotifier-f> [scopehal] azonenberg pushed 1 commit to master [+0/-0/±1]
<_whitenotifier-f> [scopehal] azonenberg 50150eb - CANDecoder: fixed handling of frames with RTR=1
yootis has joined #scopehal
<yootis> First time installing here. CMake seems to want me move all sorts of scopehal stuff to various places in the scopehal-apps tree. Is that right?
<_whitenotifier-f> [scopehal] azonenberg opened issue #336: CAN decode: don't give garbage results if waveform starts halfway through a frame -
<_whitenotifier-f> [scopehal] azonenberg labeled issue #336: CAN decode: don't give garbage results if waveform starts halfway through a frame -
<_whitenotifier-f> [scopehal] azonenberg labeled issue #336: CAN decode: don't give garbage results if waveform starts halfway through a frame -
<azonenberg> yootis: You shouldnt have to move anything around
<azonenberg> Did you follow the build instructions in the manual?
<azonenberg> If not, start there
<yootis> Ah, didn't see the docs package
<azonenberg> as far as actually installing to a location outside the build tree, that's not yet possible
<azonenberg> you have to run it from the cmake build directory for now. A proper install is on the to-do list and we have two people working on various build infrastructure issues, including that
<azonenberg> but we've got some higher priorities to deal with first
<yootis> Ah, the old --recurse-submodules -- that was the bane of the new user's existence when we did it in GNURadio.
<yootis> explains the moving of directories
<azonenberg> For all of their problems submodules are still the least bad solution to this problem right now
bvernoux1 has quit [Ping timeout: 272 seconds]
bvernoux has joined #scopehal
bvernoux1 has joined #scopehal
bvernoux has quit [Ping timeout: 260 seconds]
<yootis> ffts compiled and installed fine, but when building scopehal, I get a bunch of messages of the form "/usr/bin/ld: /usr/local/lib/libffts.a(ffts.c.o): relocation R_X86_64_32S against `.text' can not be used when making a shared object; recompile with -fPIC"
<yootis> I'm on Fedora 32
<azonenberg> yootis: Did you build ffts as a shared lib?
yootis has quit [Ping timeout: 245 seconds]
<_whitenotifier-f> [scopehal] azonenberg pushed 1 commit to master [+0/-0/±1]
<_whitenotifier-f> [scopehal] azonenberg a865fb1 - CANDecoder: require 7 UIs idle before a start bit. Fixes #336.
<_whitenotifier-f> [scopehal] azonenberg closed issue #336: CAN decode: don't give garbage results if waveform starts halfway through a frame -
<_whitenotifier-f> [scopehal-apps] azonenberg pushed 2 commits to master [+1/-0/±1]
<_whitenotifier-f> [scopehal-apps] azonenberg 7988956 - Updated scopehal
<_whitenotifier-f> [scopehal-apps] azonenberg f935d2c - Added 10 meter (positive) CAN eye mask
yootis has joined #scopehal
<azonenberg> yootis: in case you missed this before you left
<azonenberg> (14:57:29) azonenberg: yootis: Did you build ffts as a shared lib?
<yootis> got everything built now that I actually read the docs :)
<yootis> when I go to connect I get an error with GLEW and something about graphics drivers.
<azonenberg> What graphics card are you using?
maartenBE has quit [Ping timeout: 264 seconds]
<azonenberg> We need OpenGL 4.x support which generally means geforce 400 series or newer, an AMD card of about the same era, or an intel skylake or newer integrated GPU. Not sure which AMD integrated GPUs work, i don't have any experienec with them
<yootis> built in intel graphics on a Thinkpad
<azonenberg> How old a thikpad?
<yootis> ooh, I probably am older than skylake
<azonenberg> Yeah. Unfortunately that is a problem because we need compute shaders. All rendering is done in compute shaders and we need 64-bit integer support on the card
<azonenberg> eventually we'll be doing DSP on the GPU too
<azonenberg> Semi-old laptops are pretty much the edge case there
<azonenberg> any <10 year old discrete desktop gpu is fine, even a super cheap one
<azonenberg> Do you have a more recent machine you can test on?
<yootis> yeah, I'm on Broadwell
<azonenberg> Yeah you're just one generation too old
<yootis> oh well, will upgrade at some point.
<azonenberg> Yeah i'm not expecting to be bumping system requirements much in the future
<azonenberg> actually hmm
<azonenberg> wikipedia says opengl 4.3 as old as Haswell
maartenBE has joined #scopehal
<azonenberg> and broadwell is supposed to have 4.4 on windows and 4.6 on linux
<azonenberg> but i'm not sure if they have 64-bit integer support
<yootis> oh well. If you want to include fedora dependencies in the docs, here is what the user needs to install:
<yootis> sudo dnf install gtkmm30-devel cmake pkg-config glm-devel texlive libyaml-devel yaml-cpp-devel glew-devel catch-devel
<azonenberg> yootis: can you pastebin me a glxinfo so i can have a look and double check?
<azonenberg> i see a post on phoronix saying that broadwell supports ARB_gpu_shader_int64
<azonenberg> so i'm curious what the missing feature is and if it's something we can work around
<yootis> I suppose it could be a glew problem as well, I've never used that
<azonenberg> yeah gimme a minute i'm looking
<azonenberg> ok so i see gl 4.6, GL_EXT_blend_equation_separate, GL_EXT_framebuffer_object, GL_ARB_vertex_array_object, GL_ARB_shader_storage_buffer_object, GL_ARB_compute_shader, GL_ARB_gpu_shader_int64
<azonenberg> That looks like everything we need
<azonenberg> what's the error glscopeclient gave during startup?