Looking for most accurate way for flow totalization.

Hey,
just looking for some advice hoping maybe you guys can help me.
I want to be able to totalize from a flow meter. It appears to work but I am unsure as to how accurate it would be, and was seeking some criticism from some experts.

I currently take the scaled output of my flowmeter which is setup for GPM and divide it by 60 to get GPS i then have a TON counting every second. Everytime the timer .dn becomes true It will add the scaled output/60 to the totalized flow.

Does anyone have a better suggestion for this?


Thanks a bunch

What PLC brand, model? Usually, when you do this you need to be careful. At some point you may be adding a very small number to a large number and then you won't be totalizing any more. This is what occurs when you use floating point numbers.

micrologix 1400 series b
rslogix 500

Using the DN bit of a TON instruction that resets itself means that your one second pulse is not exactly once every second. It will always be more than one second between two successive pulses.

First, for the one PLC scan where the DN bit is true, the timer will not accumulate any time, so you lose one scan duration every cycle.
Second, by resetting the timer to zero, you lose any fractions of a second stored in the timer between scans. For example, if the timer's ACC value in the scan before the DN bit is true is 999 milliseconds and the PLC scan time is 10 milliseconds, in the scan where the DN becomes true, the ACC value is 1009 milliseconds. By resetting the timer you've lost those 9 milliseconds in addition to the 10 milliseconds you lose during the scan while the timer is reset.

A lot of totalisers have a built in total pulse that tends to be more accurate than what you would be able to setup via your method. Each time it pulses, that indicates a quantity. For example I recently setup some on a plant for 1m3 per pulse (for large flows) and 10 litres per pulse for smaller flows.
The trick is to ensure the pulses occur slow enough for the PLC to pickup. A one second duration pulse occuring every 5 seconds would be nice. If you are getting 60 gal per minute, then setting each pulse to 5 gal would be OK. If you have a short pulsewidth of say 200ms, and a short scan time of less than 50ms (for example), then you could set the pulse for each gallon.

Otherwise the method you proposed would be good, just find a better timer. I can't remember how it is done for the MLogix range, but I think they had an S bit 1 second pulse (50% duty cycle). Otherwise I think one of the system words had a fairly accurate free running timer that could be used.