Hi Chris,
I hope you are well. We are working with the 2024.05.29 version of MWorks and we are finding this error, where it seems the pump is responding, then we see that it is not completely connected.
What do you think may be happening?
Thanks for your help,
Nina
Thanks Nina, and adding:
It’s odd - it seems like the NE500 driver is returning a response. Earlier today, when we tried this, we got garbage in the RETURNED message. Then we power cycled the pump and unplugged and replugged the USB to serial connector (an FTDI device). Then we get this. I tried increasing the response_timeout, 300ms to 500ms, with no change.
(ps: the MWorks manual for response_timeout says it defaults to 100 microseconds - is that correct?)
best!
Mark
Hi all,
We are working with the 2024.05.29 version of MWorks
While it’s probably not the cause of this issue, I’d recommend using the 2024.05.06 build I saved for you instead (see this comment). The nightly build generally sees a lot of changes shortly after a release. Unless you’re specifically interested in testing or using those changes, you probably don’t want to rush them in to production setups.
we are finding this error, where it seems the pump is responding, then we see that it is not completely connected.
The “Did not receive a complete response” message indicates that MWorks didn’t receive a response from the pump within the response timeout. It looks the response did arrive shortly after the timeout, though.
Earlier today, when we tried this, we got garbage in the RETURNED message. Then we power cycled the pump and unplugged and replugged the USB to serial connector (an FTDI device). Then we get this.
Garbage in the response seems bad. Are you seeing this with just one pump or with multiple pumps? If it’s just one, then perhaps the FTDI device or the pump itself is failing. Can you try a different USB-to-serial connector?
I tried increasing the response_timeout, 300ms to 500ms, with no change.
That’s surprising, because, at least in the screenshot Nina sent, it looks like the response is coming in immediately after the timeout. Increasing the timeout ought to have an effect.
ps: the MWorks manual for response_timeout says it defaults to 100 microseconds - is that correct?
It’s 100 milliseconds, not microseconds. The manual has the correct value.
Cheers,
Chris
Ok, thanks! Quick responses
we will look to shift to the other nightly
thanks for confirming the response is coming after the timeout. That was even for a 500ms timeout. Does that suggest some other problem? We will try longer timeouts and let you know.
garbage in the response happened only once so it could have been due to the pump being on a very long time. Powercycling pump fixed it. If it happens again we’ll look to debug.
re: microseconds I was referring to this. I was confused but I think I understand now: this means that the unadorned value means microseconds, but the default is “100ms”.
Hi Mark,
thanks for confirming the response is coming after the timeout. That was even for a 500ms timeout. Does that suggest some other problem?
500ms seems like a totally unreasonable delay for communication via a USB-to-serial adaptor. It’s hard to imagine what could be going wrong in MWorks to cause this, which is why I suggested that it might be a hardware issue. I suppose the OS USB stack could also be messed up somehow. Have you tried restarting the Mac to which the problematic pump is connected?
I was confused but I think I understand now: this means that the unadorned value means microseconds, but the default is “100ms”.
Right. 100ms just means 100 * 1000, as described in the manual. If you set response_timeout to just a bare 100, then that would be 100 microseconds.
Chris
Hi Chris,
It looks like it may be a different hardware issue. There are two ethernet ports on the pump, and when we switch to the second port we get a connection. I think this is solved!
Nina
Hi Nina,
Interesting. According to the user manual, only one of those ports should work for connecting the pump to a computer. The other port is for connecting another pump to the first as part of a pump network. The fact that it even partially works on both of the ports is surprising.
In any case, I’m glad that fixed the problem!
Cheers,
Chris
Oh interesting, and thanks for the catch in the manual!
Its good to know that this is expected. Thank you.
Happy 4th,
Nina
