Reading specs


Thread Starter

Bob Peterson

Just curious. If you got a control description that described analog alarms in some detail, BUT did not state whether the alarm was activated
when the parameter was > the setpoint, or >= setpoint (for high alarms), would you either:

a) assume its PV > SP
b) assume its PV >= SP
c) ask for clarification (the right answer of course)

In the real world, most of us would probably pick either (a) or (b) and the customer would decide that our 300 alarms are all wrong because they
think it should be the other way, of course only telling us this after its complete and ready to ship (or better yet after its in the field and


Bob Peterson

US Filter, Rockford Number 815-877-3046 X576

Anthony Kerstens

I doubt that an end-user would notice the difference between > and >= on and alarm, so that thought probably wouldn't even cross my mind. Most people wouldn't notice an alarm until it's been annunicated, by which point it's likely > anyway.

However, I would ask what they want for an alarm hysterisis.

Anthony Kerstens P.Eng.
WhatzamatterWitYou? You can't figure out for yourself where every alarm point should be for every process that you ever encounter?!! What
are we paying you people for?
NOBODY wants to set the actual alarm points; not the general contractor, nor the engineers, nor the customers, nor the end-users. That's why
they write up those detailed descriptions of how alarms will be be dealt with--it's a smokescreen, to pretend that they have addressed the issue.
If they won't give you a concrete answer on where the alarms should be, then it's their own fault that they have nuisance alarms if you make
them all SP.

Bruce Durdle

Given that, with a digital system, the resolution of a single bit is much better than the typical accuracy of a transmitter, (12 bits = .025%) I'd point out (politely or otherwise, as appropriate) that there is no significant difference between the two situations.


Bob Peterson

They actually specified the alarm hysterious where they wanted it, or how to reset the alarms, even alarm delays where they wanted them. They just did not specify whether the alarm trips when it exceeds the sp or when it is at or above the SP. I chose to make the alarm come on when the PV > SP. They made me change them so they trip when PV >= SP.

Bob Peterson

Greg Goodman

In the real world - at least at the HMI software level - it's a function of how the specific software package is implemented. Most packages don't give you the option; you specify an alarm limit in a dialog box, and the software generates the alarm according to whatever rules Invensys or Intellution or whomever decided on when they wrote the package. (And to save Curt the trouble, I'll point out that using a proprietary package precludes the possibility of you changing how it works to suit your different requirements.)

If a package offers an alarm limit and a choice of test condition (say, between '>' or '>=' for a hi alarm and '<' and '<=' for a lo alarm), then there's both an opportunity to consider the issue and a prompt to ask the user what he prefers.

As some have noted, the granularity of most process data is such that the issue is moot. However, there may be data available at the HMI for which the alarm limit test condition makes a difference. (Reaching for an example, I'm imagining a 'consecutive failed polls' counter. Do I want to gen an alarm on the 3rd or 4th failure?)

lost-in-minutiae-ly yours,