Re: unrealistic settling time

From: Petter Strandh <petter.strandh_at_vok.lth.se> · Date: Thu, 07 Jul 2005 08:14:34 +0200

Are you shure the problem isn't electric. Usually there is a protecting
capacitance on the reading channels. What can happen is that you carry
charge between channels when you multiplex. If you short channel 0, what
happens then.

Unless I misinterpreted you, reading from an unconnected reading channel
has undefined results. It should be connected with something that has a
low impedance as possible. 

Hope that I am not misleading you at this early hour.

Regards
Petter Strandh

tor 2005-07-07 klockan 06.37 skrev Noah Watkins:
> Hi.
> 
> I have a Measurement Computing PCI-DAS6025 board, though the question 
> may not be specific to the board.
> 
> In the User's Guide provided by the company, a settling time of 25 
> microseconds was typical for reading two different channels 
> consecutively and acquiring good data in the worst case.
> 
> I am experiencing with the cb_pcidas64 driver a situation where delays 
> much longer have to be introduced. For example with nine volts applied 
> to channel 8, and nothing applied to channel 0:
> 
> Reading from channel 8: ~9V
> Reading from channel 0 to reset
> Sleep for 50 microseconds
> Reading from channel 0: ~7V
> 
> In fact, it isn't until i sleep for around 80 _milli_seconds that i get 
> readings that are reasonable. Given the sleep time limitation of 
> instruction INSN_WAIT, all of my sleeping is occurring at the user 
> level. This seems completely unrealistic.
> 
> Has anyone ever encountered this degree of settling requirement before? 
> Where might be a place to start investigating.
> 
> Thanks,
> Noah W
> 
> _______________________________________________
> comedi mailing list
> comedi_at_comedi.org
> https://cvs.comedi.org/cgi-bin/mailman/listinfo/comedi