Experimental Physics and
Industrial Control System

I'm wondering if there is a way to set a delay time (say 10 seconds) for record processing based on whether an input is in a BAD state. If the record is still in a BAD state after the delay of 10 seconds, I'd like to process the record and change its value. If the value went BAD, but changed back to OKAY, then I'd like to ignore the alarm.

I'm using a calcout record to determine if a temperature is below -180 C. If the temperature increases above -180 C, the calcout record outputs a 0. The CALC condition is the following: TEMP<-180?1:0. This output is tied to a hardware relay channel which opens a relay when its value is 0 and closes a relay when its value is 1. The temperature signal is from an RTD temperature sensor controller which outputs an analog voltage proportional to temperature and is wired into an analog-to-digital converter. We suspect that there is sometimes noise on the analog signals, causing false readings, which causes the relay to open and systems to shut down immediately. A small delay check would help filter noise and false readings.

I'd like to make a delay condition where the temperature must be BAD (above -180 C) for a certain amount of time before opening the relay. If the temperature turns back to OKAY within the time period, then I'd like to ignore the condition, and have the relay not change state.

The analog voltage is digitized using the same I/O controller (United Electronic Industries, UEI, Model UEIPAC 1200R) that drives the relays. The UEIPAC contains an embedded Linux romdisk which is running EPICS 3.14.12.

I've tried using the calcout record's DLYA field, but it seems to only delay the processing of the record, and changes its value whether or not the INP (temperature) has changed back to an OKAY state.

Has anyone had a need for this type of delay conditional check? If so, is it possible/recommended to write the logic using EPICS records, or by using another method, such as adding a feature to the driver? What might be necessary to implement a check like this in EPICS?

Thanks!

Chris Whiting

CHESS, Cornell University