Use a Control Logix 1756-IB16I module as opposed to a 1756-HSC counter frequency

[tlona]

I have am trying to determine if I can use a 1756-IB16I (or some dc input equivalent) to handle a 24vdc input pulse frequency of 80 pulses per second. The Application is Flow and I have seven (7) pulse meters that provide 1600 pulses per gallon at a flow rate of 3gpm. The HSC module only has two inputs and therefore need eight (4) 1756-HSC modules. Thank-you

 

[joseph_e2]

I have an old spec sheet that says the input delay time is 1ms max plus filter time, which can be 0ms, 1ms, or 2ms. You would then need to consider your module's RPI (Requested Packet Interval) and the PLC's scan time. At 80 pulses/sec, you'll have an edge about every 6ms, so (theoretically) a very fast timed interrupt can catch it if your RPI is small (like 1ms or so).


You could also look at the 1756-IB16IF, which is a fast input module. I don't have a datasheet on that one, but it should be easy to find on RA's site. You'll still need a very fast timed interrupt and RPI on the module, unless that one has some sort of HSC type functionality built in.

 

[jstolaruk]

If lowering your cost is primary, check out these: ADAM-6151EI-AE - 16-CH Isolated DI Ethernet/ IP Module
They may be fast enough.

 

[Contr_Conn]

How about 1756-LSC8xIB8I counter module with 8 channels?

 

[jstolaruk]

Nice, haven't seen that one before.

 

[Contr_Conn]

Nice, haven't seen that one before.

I used this low speed counter more than 10 years ago, not sure why people are not aware of it.
It’s way better than just a regular input for counting.

 

[ceilingwalker]

How about 1756-LSC8xIB8I counter module with 8 channels?

Wow! I have never heard of this before! Thanks for posting that!

 

[Sergei Troizky]

I have an old spec sheet that says the input delay time is 1ms max plus filter time, which can be 0ms, 1ms, or 2ms. You would then need to consider your module's RPI (Requested Packet Interval) and the PLC's scan time. At 80 pulses/sec, you'll have an edge about every 6ms, so (theoretically) a very fast timed interrupt can catch it if your RPI is small (like 1ms or so).

I wonder how 80Hz became 6ms "edge".

 

[Steve Bailey]

I wonder how 80Hz became 6ms "edge".

80 cycles per second = .0125 seconds per cycle. Each cycle has two 'edges', a rising one and a falling one. In a square wave these edges occur one half cycle apart, thus "an edge about every 6ms".

 

[Sergei Troizky]

I don't think the flow meters need quadrature counters, so 12ms period is what I had in mind.

 

[drbitboy]

I don't think the flow meters need quadrature counters, so 12ms period is what I had in mind.

The issue is not quadrature.

In order to count pulses, the PLC needs to detect both the 1 and the 0 of the pulse signal; if it sees two successive 1s but do not see the 0 that was between them, it will miss the second count.

So the PLC needs to have a scan cycle every 61.25ms at least, and preferably more often (Nyquist and all that).

 

[Steve Bailey]

I don't think the flow meters need quadrature counters, so 12ms period is what I had in mind.

You may choose to only count on the rising edge, but you still have to know that the falling edge occurred.

 

[tlona]

Excellent, thank-you for the insight and other recommended modules (I have not her of them either). The meters are Mcnaught (MX09)PD meters with 2-wire pulsers, no quadrature.

 

[cardosocea]

I think they stopped making them, but if you have that flowmeter with a reed switch, you'll want a counter card that can filter the incoming signal. Depending on the use you give them and no matter what manufacturer's say, the reed switch will eventually die.
In my case the problem was compounded by the fact that the McNaught flowmeter was also a restriction plate. But a properly sized flowmeter with a reed switch still had issues eventually.


Edit: The McNaught flowmeter itself was pretty accurante and well built. My comment is about the use of reed switches, not dissing on McNaught.

 

[tlona]

Thank-you for that insight on the reed switch. The meters are still manufactured as we have a recent quote. There are two options for the output: 2-wire magnetic sensor or NPN with 3.1kΩ pull-up resistor. Cost is a concern and so is chassis space. The 1756-LSC8xIB8I sounds like the best option as opposed to the 1756-IB16IF. But either should do the job.