<azonenberg>
Soo i tried using the can decoder in the repo and it was pretty unusable :p
<azonenberg>
I'm going to probably end up rewriting most of it
<_whitenotifier-f>
[scopehal] azonenberg pushed 1 commit to master [+0/-0/±2] https://git.io/JT5Jt
<_whitenotifier-f>
[scopehal] azonenberg a6932eb - Better handling of differential signals with high common mode
Degi has quit [Ping timeout: 260 seconds]
Degi has joined #scopehal
Pretzel4Ever has joined #scopehal
Pretzel4Life has quit [Ping timeout: 240 seconds]
<azonenberg>
And i'm halfway through rewriting the CAN decode
<azonenberg>
Off to bed but hopefully can finish before work tomorrow morning
electronic_eel has quit [Ping timeout: 264 seconds]
electronic_eel has joined #scopehal
juli966 has joined #scopehal
bvernoux has joined #scopehal
<_whitenotifier-f>
[scopehal-apps] tarunik commented on issue #274: ASAN reported new/delete size mismatch (sliced object?) - https://git.io/JT5bJ
* azonenberg
stabs PicoVNA with a titanium spork
<azonenberg>
So the latest of their infuriating quirks of the stupid vb6 software
<azonenberg>
It is very picky with s-parameter files for de-embedding
<azonenberg>
specifically they have to use windows cr/lf line endings
<azonenberg>
Also it looks like my probe impedance is much higher than i thought it was
<azonenberg>
because i wasn't doing reference plane extension correctly in my VNA measurements
<azonenberg>
so now i have to re-measure lots of other stuff. yay :p
<azonenberg>
Apparently i hadn't quite grokked that port impedance depends very much on the *phase*, not just the *magnitude*, of the reflection
<bvernoux>
hello
<bvernoux>
I have finished my custom LA POD for the Rigol MSO5000 ;)
<bvernoux>
it works like a charm
<bvernoux>
Just tested all chans 1 by one so far will test in // later
<bvernoux>
anyway it works up to 250MHz signal as input
<bvernoux>
and it was a sin signal as I do not have good AWG only RF Gen ;)
<bvernoux>
I checked the level also it works down to about 1.2V
<bvernoux>
very nice video TSP #178 - Keysight MXR 8-Channel Real-Time Oscilloscope Review, Teardown & Experiments
<bvernoux>
This scope is amazing but I imagine the price ;)
<miek>
i'm very jealous of that fancy sampling scope/module in the previous video
<azonenberg>
miek: yeah i want that lol
<azonenberg>
i'm half an hour into #178 and had to go to work so couldnt finish
<bvernoux>
and it have a 10bits ADC with up to 9bits ENOB
<azonenberg>
16 Gsps is a little light for 6 GHz though
<azonenberg>
it's only a little bit past Nyquist
<bvernoux>
yes I was thinking the same 6GHz with only 16GSPS ...
<bvernoux>
Tek is more serious about that as their Tek 6 Series have 50GSPS ;)
<bvernoux>
but they have a very bad ENOB in fact ;)
<bvernoux>
less than 6 bits ENOB IIRC in their spec
<azonenberg>
Although the LeCroy 12 bit scopes are also relatively low sample rate. The WavePro HD series, for example, are 10 Gsps 4ch / 20 Gsps 2ch
<azonenberg>
in fact, the 6 and 8 GHz models cannot hit full bandwidth on all four cahnnels at once
<miek>
also quite impressed with the features on that newer sampling mainframe, like pattern lock and all the automatic jitter stuff - i thought they had been more neglected than that
<azonenberg>
you need to be in 2ch 20 Gsps mode to even get full bandwidth
<bvernoux>
the question is who needs 8 channels ;)
<azonenberg>
Which i'm not thrilled about
<azonenberg>
But for example my 8 bit WaveRunner 8404M-MS is 4 GHz bandwidth and does 20 Gsps on all channels and 40 Gsps on 2ch
<bvernoux>
I have tested the Bandwidth of my MSO5074 ;)
<miek>
bvernoux: gotta have more BNCs or it looks silly as the screens get larger!
<bvernoux>
with my RF SigGen ;)
<bvernoux>
each channels are not on par ;)
<azonenberg>
Which is what i would consider a more reasonable sample rate to bandwidth ratio
<azonenberg>
i.e. 10:1
<bvernoux>
Chan1 568MHz ;)
<azonenberg>
miek: also interesting that keysight is pushing to higher bandwidth on BNC
<bvernoux>
Chan2 547Mhz
<bvernoux>
Chan3 558MHz
<azonenberg>
Everyone these days seems to be starting to push BNC a lot further than it was originally designed
<bvernoux>
Chan4 580Mhz
<bvernoux>
it is in 1MOhms mode as there is no native 50Ohms :(
<bvernoux>
I'm pretty sure it will reacheasily 2GHz BW with 8GSPS
<bvernoux>
like the MSO8000 ;)
<bvernoux>
anyway for the price 1KUSD MOS5074 is just totally amazing
<azonenberg>
yeah, if we can just make glscopeclient stop crashing it :p
<bvernoux>
it does not crash with latest firmware
<bvernoux>
it is just not very usable too slow
<bvernoux>
it is fixed in latest FW
<azonenberg>
Oh
<bvernoux>
confirmed also by Degi
<bvernoux>
yes I have helped her to update it
<bvernoux>
she confirmed that like me there is no any crash
<bvernoux>
it is rock stable
<bvernoux>
I have just a strange things with an offset on channels depending on scale ...
<_whitenotifier-f>
[scopehal-apps] azonenberg commented on issue #146: glscopeclient hangs on Rigol MSO5000 - https://git.io/JTd6W
<bvernoux>
but glscopeclient freeze often on Windows
<azonenberg>
Great. Let's wait till Kenley and pepijndevos_ confirm then i'll close that ticket
<azonenberg>
If you have an unrelated reproducible crash, file a separate bug for it
<bvernoux>
the issue is everything is frozen it is not a crash
<azonenberg>
Well, still a bug
<bvernoux>
I will do more test later
<azonenberg>
Ok
<bvernoux>
also the driver really need to be rewritten
<bvernoux>
it is full of trap
<bvernoux>
it does not support trigger
<bvernoux>
all shall be ready in fact ;)
<azonenberg>
That's not a rewrite, that's just extending the trigger code
<azonenberg>
edge is the default trigger that's done first in a new driver
<bvernoux>
ha ok
<azonenberg>
there's dynamic enumeration code for a driver to advertise what kinds of triggers it supports
<azonenberg>
which may even depend on software options on the scope, etc
<bvernoux>
ha great
<azonenberg>
So right now the rigol driver only advertises edge trigger
<bvernoux>
I suspect the Rigol MSO5000 driver is very basic about that
<azonenberg>
This is the big todo for the tek mso5/6 driver too
<azonenberg>
i have a lot of trigger types i havent added
<azonenberg>
even lecroy has a few i haven't done yet like i2c and spi
<miek>
it's a good point though, i'm not sure that even the basic stuff on the mso5k driver was ever really verified/finished?
<miek>
someone with the scope really needs to take on development of it
<azonenberg>
miek: well now that we have several active users in the channel, we have people who can do that :p
<azonenberg>
degi, bvernoux, kenley, and pepijndevos_ all have mso5000s afaik
<bvernoux>
yes
<azonenberg>
between the four of them we should be able to make it a "tier 1" supported platform like tek mso6 and lecroy
<azonenberg>
i.e. almost everything works, it's stable, well tested, etc
<azonenberg>
It's definitely not there yet
<azonenberg>
My TODO for the next couple of days is to implement slew-rate triggering on tek mso5/6, fix the can bus protocol decoder, possibly write a sae j1939 protocol decode, and then a few random UI fix tickets that have been pending
<azonenberg>
also probably lecroy i2c/spi trigger as i've been putting those off for a while
<bvernoux>
for the Rigol MSO5000 the most mandatory is to find a way to be usable remotely
<bvernoux>
as it is like <1 frame per second
<azonenberg>
That may just be cpu performance limitations
<azonenberg>
we can do some optimizations to minimize unnecessary queries
<bvernoux>
clearly I prefer to use it natively than waiting 1s for a screen refresh
<azonenberg>
but we cant fix slow firmware
<bvernoux>
yes I know
<bvernoux>
To be checked with Rigol if they can update that for remote mode without screen update locally to improve speed to send the data ...
<miek>
some of that may also be hard limits on ethernet/usb bandwidth. i think that scope has quite a lot of memory?
<bvernoux>
I'm pretty sure it could push at least 10 frames/s
<bvernoux>
to be much more usable
<miek>
my agilent stuff certainly gets sluggish with full memory & all channels on
<bvernoux>
it has Gigabit Ethernet I doubt the limit is here
<miek>
doesn't mean it can utilise that :p
<bvernoux>
it have also memory especially if they cut the display to a special remote mode ...
<Degi>
I mean 10 FPS with like 10 kPoints should be doable
<Degi>
Ofc with 200 Mpoints its hard
<bvernoux>
yes not with 200 Mpoints of course ;)
<miek>
i would be extremely surprised if you can get them to develop a special remote feature for you
<bvernoux>
not for me ;)
<bvernoux>
to be checked
<bvernoux>
maybe it is not a so huge work
<bvernoux>
for them ;)
<bvernoux>
As I do not plan to do such development I have tons of things to do on my side
<bvernoux>
so far I'm on Ledger Donjon CTF 2020 ;)
<bvernoux>
the hsm to destroy I'm on last level ;)
<bvernoux>
remote glitching the secure_mcu(Atmega) from main MCU (Cortex-M3 STM32F2) using a remote exploit in main MCU ;)
<bvernoux>
It is first time I'm doing remote exploit on bare-metal MCU it is very fun
<bvernoux>
with less than 1Kbytes ram for the exploit ;)
<_whitenotifier-f>
[scopehal] azonenberg pushed 1 commit to master [+0/-0/±2] https://git.io/JTdPx
<_whitenotifier-f>
[scopehal] azonenberg 31deb55 - CAN: Complete rewrite of original decoding logic. Now works for base format frames, but not extended. See #32.
<_whitenotifier-f>
[scopehal] azonenberg opened issue #334: CAN FD protocol decode - https://git.io/JTdX4