Magnova driver

[fredzo]

Hi Andrew,

As discussed, this PR is for Batronix Magnova scope driver.
It's not yet ready to merge (I'll remove the [WIP] tag when it is): it's being beta tested by EEVblog forum members and Batronix team.
I'll have beta testers file their reports here.
I'll create a doc PR soon for the driver documentation.

Best,

Frederic.

[fredzo]

@azonenberg this PR is ready for review if you can.
Magnova driver is working pretty nicely.
Doc PR is on it's way.

[xyphro]

Unfortunately the new version also shows the same issues on waveform reads :-(
In addition to the ngscope trace, I also made a wireshark capture from the exact same session - so they match in content.

I could not see anything suspicious in the Wireshark trace pointing to significant issues / breaking tcpip connection or so. The scope simply doesn't respond to the data readout.
Packet 5880 shows a socket reset, which is at the moment I killed the application, so it is out of context and not part of the issue.

Here both captures:
ngscopeclientTrace_magnova.zip

Firmware version you can see back from the traces :-) OK, it's the latest greates 1.7.1 version...

UsbTmc has some issues on Magnova like transmitting no zero length packets [Batronix looks into that right now], would not be surprised if the TcpIp part has also some issues.

[fredzo]

Thanks for the feedback @xyphro .
Can you tell me the exact steps to reproduce, starting from the scope loaded with default settings to make sure everything is identical ?
On my side I am able to run a continuous acquisition for more then 10 minutes without a hitch.
I am developing and testing on a Windows 10 machine, and the scope is connected to the PC with wired Ethernet via a switch.
May I suggest to try with a different machine if you can ? Maybe an antivirus or something is getting in the way ?

[xyphro]

Hi Fredzo,

as I use the same machine to do long term stress tests of another embedded device ethernet stacks I doubt an issue, but never say never. Also control other instruments with tcpip quite fine from that PC.

Let me try another windows machine tomorrow, testing is much cooler than believing, right? :-)

From the wireshark logs it get's also visible that Magnova does simply not respond.

Antivirus/firewall? Nope. I understand that doubt but can exclude that from the list of potential issues.

Letme see with another machine and report out tomorrow. With direct lan connection.

In terms of steps: Just select the scope, then press once the play button. Sometimes it works, sometimes it gets stuck for me.
Had so often "it works on my machine scenarios" with stuff I developed too, sometimes things are tricky. Hope I can find potential cause.

[xyphro]

Not that it is also configiration related on magnova. Or e.g. caused by having a samba share open from it?. Let's see, will spend half an hour after a sleep.

[fredzo]

I would bet on the SMB share! I haven't activated it on my side.

[xyphro]

Short interim testing state: with another ethernet adapter and direct connection and without DHCP it works pretty well on the same PC.
(Apart of minor pull request unrelated ngscope related issues like getting out of memory on too long sample buffer reads or drawing issues showing vertical lines at random locations).

Will now investigate on samba share. That one will not work in that setup scenario. So will turn it off in the original failing scenario.

[xyphro]

WOW, my suspicion was right: Having an active Samba connection kills SCPI on Magnova :-/

The good news is, that this means what you did is working as good as it is supposed to be.

Will write this Andre.

[xyphro]

Even though it is unrelated to your magnova driver... would be super cool if you could try out samba share too to get a second confirmation.