High Speed PLC

Lancie1 said:

If it is a bottle line, will this 30-msecond-output-PLC have to do anything else other than knock off bottles? Those other functions might need a little attention also.

Click to expand...

This small "machine" will be dedicated to removing downed bottles only.

This small "machine" will be dedicated to removing downed bottles only.

Click to expand...

Probably one of the programmable control relays would work just dandy (Eaton EZ, Moellar EASY, Cutler Hammer Pico, Schneider Zelio, Omron ZEN). Only one task, so the program will run fast, only output needs to be a transistor instead of a relay. That should be okay.

For example, an Omron ZEN with DC inputs has a input switch-on time of 15 ms, and an output switch-on time of 15 ms, making a round-trip signal input-to-output time of about 30 ms.

The Micrologix 1100 and 1400s are actually very fast. If this program is only a few rungs long it will have sub millisecond scans.


Edit to add:
I just jumped online with an ML1400 that is running 21 subroutines and about 450 rungs of logic with three PID loops and is messaging with a ControlLogix to coordinate control and handling e-net comms with a historian. It is scanning in the 2 to 2.3 millisecond range.

Would prefer not AB actually!

Click to expand...

Automation Direct CLICK maybe?

Omron have several small shoe box PLCs that would eat the job for speed.

The question is what are the specs on the solenoid? You have to allow for the actuation time and be sure that the PLC output can supply the required current to run it. Fast usually means high current.
It's no trick to get the Omron CP series PLCs that Bob mentioned down to the 2-3ms range for a small program, but if the solenoid response is 20ms it's not going to help you.

Need fast and cheap but not PLC? Go for Arduino - it has scan of 2.5 microseconds.

I'm usually working with S7-1200 and simple program ( 20 Direct drives, 5 VFDs, few analog sensors ) - scan is stable at 15-20 miliseconds.

To make S7-1214 60 ms scan i had to put a lot of SCL loops ( Profibus transfer ) transfering multiple bytes between arrays and processing it.

I think that in Your case S7-1200 may be fastest and cheapest PLC.

The Siemens FM352-5 High Speed Boolean processor is a 12 pipe parallel FPGA processor with a cycle time of 1µs. It can be used standalone without a PLC and costs around 700€ (sorry don't have a price list at home). It has 8 DIs and 8 DOs + 4 DIs for an incremental encoder which can also be used as normal DIs.

It will definitely cover your speed requirements - whether it's cheaper than the other suggestions I can't really say.

Looking at the Keyence KV-16. Has a minimum scan time of 140 microseconds. Price is in the $400 - $500 range. Somewhat limited on program size, but if its being used for a dedicated function or two, it may be okay. I've personally never used these, but I've seen them used in applications needing higher speed inputs or outputs than a "full size" plc.

Tough to beat a Beckhoff

I programmed an Omron CP1L recently (around $400) for a reject system with 3 cameras, 5 FIFO stacks (30 deep) and about 100 rungs of code with a high speed counter running in quad mode with 47kHz max running frequency. Scan time max was 1.8ms average was 0.8 to 1.2ms. You also get 1ms resolution high speed timers. Very impressed with the results. Go for the Omron!

The PLC scan cycle is an important time factor, but think about this: in order for the PLC to know WHEN to kick off a bottle, it must have an input device. That device will require some minimum time (10 mseconds ?) to turn ON (after the input sensor is ON). Then the program will see this input, and turn on the output (after the PLC scan cycle and scan time has elapsed), which will require some minimum time (15 mseconds ?) to turn ON. After that, the kicker will require some minimum time to activate (5 mseconds ?).

If you add up these 4 times, each could be an average of 7 mseconds, and the 30 msecond minimum would be met. The PLC scan time is probably going to be small compared to the other 3 time factors.

Because the scan time is probably the smallest of the 4 time factors (regardless of the PLC brand and model), that makes you wonder why Exabyte just doesn't use the main bottle-line operating PLC (AB L35E) to kick off the bottle, instead of adding another PLC. There are signs of a Student Problem.

Although this discussion about scan time and plc brand is interesting the objective is to remove a downed bottle. In my experiance in the bottle and canning industry most downed bottles are removed by a mechanical method. You simply remove the bottom side rail in an area of conveyor that changes direction. Either that or through tilted offset section and let gravity takeover. No need for solenoids, PE's or PLC.

A small CPLD with a few passive components will acomplish everything you want with cheap components, allthough you need some electronics knowledge to wire this together.

On the other hand maybe an off the shelf Arduino and passive components will do if you know some C-ish programming.


edit: completely missed page two, adfox already suggested arduino