Troubleshooting
Software problems (e.g. Matlab, get_scope, Cadence)
Can't use get_scope from Matlab
Symptoms
I get an error in Matlab (please add details here).
Possible causes
1) Are you trying to acquire with get_scope function from a channel which is not switched on on the scope? Please before reading further check if the channel that you are acquiring is plotting any data on the scope. If not the gpib function will get stuck. The solution to this problem is to switch off and on the scope and acquire from the right channel.
2) The system didn't assign correct permissions to /dev/gpib*.
3) iv is not working. When you do 'which iv' you get a Mathworks function, not the local function to access GPIB. Your Matlab path is not set correctly. Add the path /home/avlsi/Documents/students/functions. This folder has been copied from linsrv SMB by the TAs to each machine's local home folder. (See thread below)
Solution (ask your TA)
From a root shell (type in su -s from avlsiuser) you can type:
# ssh -l avlsiadmin localhost
insert password
# sudo -s
insert password
# gpib_config
# chmod 777 /dev/gpib*
SMB mounting and linsrv/ Matlab PATH
On 22/09/14 16:14, Giacomo Indiveri wrote:
Hi Adrian.
We were trying to connect the student lab machines to linsrv02, but I
can't get samba to work.
In what way not work? What did you try? How did it fail?
We would like to (properly) set the path to the shared folder:
smb:linsrv02/misc/avlsi/share/matlab > There seem to be 2 issues here: 1. How to make a smb connection permanent (i.e. get reconnected whenever the user logs in). I currently don't know how to do this (properly), but can talk to Claudio about it if still desired. However, the TAs have now made local copies of what they would otherwise need from the server, and this seems like a good setup to me. 2. Making matlab remember the path from session to session. This works on 3 out of 4 machines. I'll continue to look at why it doesn't work on the 4th later, but for the moment, Chenghan told me he is happy to work around the problem by setting the path manually on that machine each time it's necessary (after the student logs in). > BTW, there used to be a /mnt/projects folder that was mounted to point > to linsrv02/avlsipub but that shared folder is not accessible anymore
Can you tell me what directory that share referred to?
At present I see the following under /projects/misc/avlsi:
drwxrws— 2 512 avlsi 0 May 6 2003 bin drwxrws— 8 512 avlsi 0 Feb 4 2009 book drwxrws— 3 1146 avlsi 0 Aug 10 2011 chbraen drwxrws— 2 512 avlsi 0 Apr 28 2009 circuits drwxrws— 5 512 avlsi 0 Nov 26 2012 class drwxrws— 6 512 avlsi 0 May 31 2011 docs drwxrws— 2 512 avlsi 0 Sep 20 2002 htdocs drwxrws— 10 512 avlsi 0 May 26 09:07 mpws drwxrws— 3 512 avlsi 0 Apr 28 2009 pcb drwxrws— 7 512 avlsi 0 Apr 28 2009 projs drwxrws— 8 1146 avlsi 0 Jul 3 07:38 sensors drwxrws— 7 562 avlsi 0 Feb 22 2012 share drwxrws— 2 9999 avlsi 0 May 15 2009 student drwxr-s— 3 525 avlsi 0 Oct 5 2011 svn-recovery drwxrws— 8 512 avlsi 0 May 27 11:40 sw drwxrws— 2 512 avlsi 0 Dec 4 2009 tobi
GPIB
See https://www.ini.uzh.ch/bugtracker/show_bug.cgi?id=3197 for the BugZilla on linux GPIB (login required).
This also has the subversion repository location for the GPIB module which is https://svn.ini.uzh.ch/repos/it/TRUNK/projects/gpib_ini/ (login required).
The routines for actually using the GPIB instruments (ib and get_scope) live in a different subversion repository here: https://svn.ini.uzh.ch/repos/hw/class/sw/ (login required)
Hardware problems (e.g. chips, the pot boxes, scopes)
K236
The source measure unit uses the center lead of the triax as the sensitive node, and the center lead of the separate coax as the voltage reference source node. If you want to use the middle shell of the triax to do “guarding”, read about it in Joe Keithley's “Low Level Measurements” book.
Function Generator HP 33120A
Always keep in mind that the offset value (DC level) that you set is not the value that you get as output (the output voltage). This is crucial in burst mode (see experiments on the synaptic circuits) since you'll use the function generator as a single pulse generator in the Burst mode. In this case, as long as you don't hit the Single button, the constant level that is applied by the function generator is the lowest part of the square wave (it might be the highest part of the wave as long as you don't hit Single for the first time!). For example, a single pulse from 0 to 5 lasting for 10ms can be produced via the following settings: Burts mode, Square Wave mode, freq. 100Hz, amplitude 2.5 V, offset 1.25 V and hit Single. Conclusion: never trust displays, always check signals with a voltmeter or a scope.
Pot box checklist
to do…
History of the lab
add here lines like ” - changed K236 (serial no. 123) with K236 (serial no. 678) because of following misbehaviour – details here.
Experiments
Synapse (Lab 10 or so) behaves strangely with no input
Symptoms
I put the function generator in burst mode, but the circuit is behaving strangely. Even when I don't hit the Single button the circuit doesn't look OK…
Solution
Try to hook the input pin (37?) (chip 2005rev2) to GND and check the all the nodes stay where they are supposed to stay (e.g. Vsyn to Vdd). It might be that the offset of you function generator is not properly set. Check the HP 33120A section above in this case