O900 drops connection

Open Perimetry Interface
#1

Hi all,

One of our researchers is having trouble with the O900 dropping the connection as soon as opiInitialise is called. Has anyone else seen this?

PXL_20201110_063946389
PXL_20201110_0639463894032×3024 2.97 MB

#2

Hi Andrew,

Yes. I faced this issue. I did the following things. I do not know which one of the following solved the issue. But the issue got fixed after did this.

  1. Reinstall the OPI.
  2. Disconnect cables and reconnect again.
  3. Have a dummy presentation before going into loop.
  4. Now, this is strange. I still sometimes get this error if i source the file from Rstudio. I never faced this issue if i run code via terminal. (have no idea why?)

Few extra tip:

  1. make sure there is network switch between installed between the computer and the OPI.
#3

Hi all,

I just came across this problem when specifying the ‘color’ of a stimulus incorrectly for the O900 in the latest OPI from CRAN. In particular, the color should be
color=.OpiEnv$O900$STIM_WHITE
and not
color=“white”.

I have added a warning in the code now, which is updated on github, but will take a few days to get to CRAN.

Thanks for Karam at Iowa for helping me find this bug.

Happy New Year!
Andrew

#4

And here I am again, making similar mistakes…

This time I had stim$x and stim$y as NULL. Perhaps this is why your ‘dummy presentation before the loop’ works. Perhaps the dummy presentation initialises a variable before the loop…

So the moral of the story - a dropped connection means you asked the O900 to present a bad stimulus. Looks like we should improve the error messages…