azonenberg changed the topic of #scopehal to: libscopehal, libscopeprotocols, and glscopeclient development and testing | | Logs:
Degi_ has joined #scopehal
Degi has quit [Ping timeout: 260 seconds]
Degi_ is now known as Degi
bvernoux1 has quit [Read error: Connection reset by peer]
<_whitenotifier-5> [scopehal] azonenberg pushed 1 commit to master [+2/-0/±5]
<_whitenotifier-5> [scopehal] azonenberg 470dfa5 - Refactoring: pulled a lot of common code out of DDR3 decoder to a common base class so we can share it with DDR1/2/4 decodes
<d1b2> <j4cbo> QSFP+ (4 lanes of 10G) active optical cables are a couple hundred bucks, but that’s still only a few channels of 1gsps
Famine- has joined #scopehal
Famine_ has quit [Ping timeout: 240 seconds]
massi has joined #scopehal
Implant has joined #scopehal
Implant_ has quit [Ping timeout: 252 seconds]
Tost has joined #scopehal
bvernoux has joined #scopehal
<Degi> Why would you do QSFP+ for 1 GS/s?
<Degi> You can get one meter of QSFP+ for 90 € (without taxes)
<Degi> (AOC cable, DAC is much cheaper)
<Degi> Oh, you can get up to 7 m for that price
massi has quit [Remote host closed the connection]
<azonenberg> ADC over optics is just stupid when you could put fpga and ram next to it
<azonenberg> and run optics to a PC
<azonenberg> adc over optics means you still need an fpga on the other side
<azonenberg> and you've added the cost of optics to the mix
<azonenberg> also most sfps are bidir
<azonenberg> and you're wasting half the lanes with one way data
<_whitenotifier-5> [scopehal] mubes synchronize pull request #402: Changes to support SDS2000+ scope -
<azonenberg> mubes: so what's the status of this code
<azonenberg> are you happy enough that you're ready for me to review and merge to mainline?
<d1b2> <mubes> Its alpha, but I think it's time for you to cast an eye over it and poke at the things that must be fixed before merge
<d1b2> <mubes> I've got another patch against xptools for windows compilation but I'm unable to test it on a real system so slightly loath to push it
<azonenberg> Ok i'll review after work or over lunch today
<azonenberg> Send the PR for xptools and i'll look. If it passes CI testing (right now mostly just 'does it compile', we have very limited actual unit testing so far) i'll merge
<d1b2> <mubes> No prob. I've got one guy who is building the windows version but he's new to all this so not sure how far he will get. I can push that patch too?
<azonenberg> and we can address problems if users complain
<d1b2> <mubes> OK. The code is only used in Siglent, so no-one else will get upset
<d1b2> <mubes> (famous last words)
<azonenberg> Send everything. As a general rule, patches that don't break existing functionality are OK even if they're not fully complete
<d1b2> <mubes> 'tis done.
<azonenberg> i.e. the functionality of master after each commit/merge should always be a strict superset of what it was last commit
<d1b2> <mubes> Well, Siglent will be slightly less borked, so I guess that's some sort of positive? If someone has an old model Siglent scope it should be easy enough to add those back in if we want to.
<d1b2> <mubes> should now have everything. The scopehal patch won't cross-reference to the latest xptools one because I've tried to keep them decoupled, but you will need both for windows builds.
<azonenberg> ok
Tost has quit [Ping timeout: 260 seconds]
Tost has joined #scopehal
<azonenberg> Hmmmm
<azonenberg> So there seems to be a bug in the lecroy driver somewhere WRT skew between analog and digital channels
<azonenberg> Don't know what it is yet
<azonenberg> but looking at the scope screen vs displayed waveforms I see a *massive* shift
<azonenberg> i first noticed it by observing causality violations where analog channels showed effects before the causing event on the digital channel
<azonenberg> and looking at the scope display, they're correct
<azonenberg> In the scope display it appears that the analog capture starts 5us after the digital does
<azonenberg> no wait i'm misreading the time scales
<azonenberg> 100ns after
<azonenberg> But i dont know if this is a consistent offset or if it's related to these scope settings, or how to figure out, etc
<_whitenotifier-5> [scopehal] azonenberg labeled issue #404: LeCroy: digital and analog channels displayed out of sync with each other -
<_whitenotifier-5> [scopehal] azonenberg assigned issue #404: LeCroy: digital and analog channels displayed out of sync with each other -
<_whitenotifier-5> [scopehal] azonenberg opened issue #404: LeCroy: digital and analog channels displayed out of sync with each other -
<_whitenotifier-5> [scopehal] azonenberg labeled issue #404: LeCroy: digital and analog channels displayed out of sync with each other -
<_whitenotifier-5> [scopehal] azonenberg edited issue #404: LeCroy: digital and analog channels displayed out of sync with each other -
<Degi> I thought about PCIe over optics, that should work
<azonenberg> Might be a little tricky if you have power management enabled because of the OOB states (differential zero iirc?)
<azonenberg> once the link is up, it should
<Degi> Hm, I think as long as the optics look sufficiently like a PCIe receiver and you don't go to power save ever it should work fine
<azonenberg> Yes exactly
<azonenberg> i'm just not sure if you'd have to hack anything to correctly detect or something
<azonenberg> once the link is up it would work fine
<Degi> I mean I think getting it up should be fine as long as your SFP on the PC side is powered (at least on the RP64 you only need to send IIRC TS1 packets, not sure how more sophisticated devices handle it but I think it should be the same for most systems)
<Degi> So if you have your own FPGA it should be fine
<d1b2> <TiltMeSenpai> oh no of course that's a thing
juli969 has joined #scopehal
juli969 has quit [Quit: Nettalk6 -]
m4ssi has joined #scopehal
<Degi> Nice
<Degi> Would be good for isolated stuff
<Degi> I wonder why they have chips between PCIe and the SFP
m4ssi has quit [Remote host closed the connection]
<Degi> Probably to fully meet PCIe spec instead of just being "good enough"
m4ssi has joined #scopehal
bvernoux has quit [Quit: Leaving]
m4ssi has quit [Remote host closed the connection]
Tost has quit [Ping timeout: 246 seconds]
maartenBE has quit [Ping timeout: 246 seconds]
maartenBE has joined #scopehal
<azonenberg> also ooh my LPA-K-A adapters shipped ,coming early next week
<azonenberg> These are BMA to 2.92mm adapters but they have screw mount frames around them so you don't put any shear loads on the BMA
<azonenberg> which feels pretty soft and i don't think would like that long term
<azonenberg> I ordered two so if i wanted to use all four scope channels i'd need to use some combination of my BMA-SMA cables, the 13 GHz active probe, or the lower speed BNC inputs
<azonenberg> but for the near term i'm not expecting to need more than a handful of 16 GHz channels at a time