RTO6 support

[kristof-1]

Hello,

I have a R&S RTO6 oscilloscope, and release of nscopeclient.

Right now the program is crashing while I want to connect with LAN 198.19.60.4:5025 or twinlan 198.19.60.4:5025:5026 .

Something is happening because it enters remote mode during connect.

I can support in debugging, sending you logs of the issue and so on.

Thank you and best regards,
Kristof

[azonenberg]

Twinlan is only for specialized instruments that have a separate data plane server. Most benchtop scopes, including the RTO6, use the normal lan transport.

To start, please try the latest development build (built from source or download the most recent GitHub Actions test package for your platform), run with --debug --trace SCPISocketTransport, and post the console output.

[kristof-1]

SCPI-log.log

[kristof-1]

just a small update. with rs driver it doesn't crash but there's a problem with plotting

[azonenberg]

The RS driver is for older scopes using a different command set. It's going to give you complete nonsense if you try to use it on an RTO6.

Let me have a look through the log and see what jumps out at me.

[azonenberg]

So it looks like everything works normally up to the point that we detect the function generator (option B6) and attempt to configure it.

[SCPISocketTransport::SendCommand] [198.19.60.4] Sending WGEN0:ENABLE?
Warning: Socket read failed (errno=0, No error)
[SCPISocketTransport::ReadReply] [198.19.60.4] Got 
[SCPISocketTransport::SendCommand] [198.19.60.4] Sending WGEN0:VOLTAGE?
Warning: Socket read failed (errno=0, No error)

At that point, we send a couple of commands which the scope doesn't reply to, and the reads time out. The crash is probably the result of calling std::stof() on the reply to WGEN0:VOLTAGE? in RSRTO6Oscilloscope::GetFunctionChannelAmplitude(), which isn't a valid floating point number.

It looks like the recv() call is returning SOCKET_ERROR, but instead of setting errno like happens on POSIX, Windows expects you to call WSAGetLastError() to see what the failure is. My guess is that it's WSAETIMEDOUT (read timeout) or WSAECONNRESET (socket closed by remote side) but I don't have enough data here to know why.

[azonenberg]

Also, this seems to be an issue specifically with the RTO6 driver on the back end and not a GUI bug, so I'm going to move this issue over to the scopehal repository rather than here.

[azonenberg]

According to the RTO6 user manual on page 2670, WGENx:ENABLE? seems to be the right command to send.

Do you actually have the waveform generator option installed? Does it work as expected if you use the scope front panel to control it (i.e. there's nothing actually wrong with the function generator hardware etc)?

[azonenberg]

Oh, wait a minute I think I see the problem. The indexes are 1 based, we should be sending WGEN1:ENABLE? not WGEN0:ENABLE? and ditto for voltage.

[azonenberg]

But the code here looks correct, which makes me wonder if there's a bug elsewhere.

Are you able to compile from source and attach a debugger? If so, please break on bool RSRTO6Oscilloscope::GetFunctionChannelActive(int chan) in scopehal/RSRTO6Oscilloscope.cpp and let me know the values of chan and m_firstAFGIndex.