Jon,
I had the same problem a while ago and find out that the signal from
thermocouples is a very weak signal so that for an accurate measurement a
lot signal conditioning has te be done. This conditioning is done in the
measurement device and (dependend of the device but..) in most cases you
cannot affect it.
When it is possible to control the signal conditioning you may get
improvements by shielding the thermocouple wire and get them not longer as
neccessary.
When channels are programmed to measure at a higher frequency the accuracy
decreases.When faster measurements are needed than 1 Hz you need other
sensors (NTC, PTC, IR, etc) which give a strong enough signal to read it
out. When choosing another sensor be sure that the heat-up time (tau) of the
sensor
is short enough. When you really wan to measure fast temperatures,
use then an IR sensor which takes only (parts of a ) microsecond to read it
out with high accuracy.
Hope this information will help you
Dr Jon Atkinson schreef in berichtnieuws
95ehgb$ece$1@usenet.bham.ac.uk...
> Hi
>
> We have recently purchased a SCB-68 junction block, and I have been
> experiencing a problem when it comes to setting up the CJC. I am reading
my
> thermocouples in differential mode, so I have set the jumpers accordingly.
>
> I have then set the cjc source for each virtual channel to be "built in".
>
> The problem I am finding is that my program is now running much much
slower
> than it did before. My program controls the temperature of a furnace, and
I
> have placed a "time since last" function inside the PID loop so I can
> monitor how quickly it is running. Prior to using the built in CJC of the
> SCB-68 the loop time was 0.025 secs. Now it is 1 sec.
>
>
> Nothing is wired
into channel 0, and no virtual channels are set-up as
> channel 0, so I don't think it can be a conflict.
>
> Any ideas?
>
> Jon Atkinson
>
>
