calculation of high speed PLC system

Hi!
i need to choose very high speed processing system PLC components about 60 KHZ-/17 microsecond from triger the digital input to get the analog reading.
can someone tell me how to calculate this time include scan time, delay of input cards ,and so on and if its possible to achieve that with PLC.
thanks
Tzvi

You need to refer to your PLC manual for the execution times of the various instructions you intend to use.

Can you please give better explanation about your application.
It will help for good advice.
BTW Tzvi ana Hanan are Israeli name any connection?

calculation of consistence and delay time between reading D/I to analog reading

Hi !
The PLC doing sample of analog value .this is done by getting a triger from digital input doing interrput in the scan exuecte the code meaning reading the analog value.
the point is cosistence.
the tolerance for each sample should not exceed 17 microsecond .
the question when the triger happend ,if its happend during the scan
or during update of output image table ,which is 2 differnt sample time .
i need some more information on scan time and suggestion

Arik i hope this give you more details ,and you right for the name

I doing something very similar with Omron CQM1H-CPU51
I read high speed counter evey 1 ms.
I using interval timer to create interrupt and to read the counter.
If you want to be in safe side in your subroutine first make I/O refrash theî read the value.
I hope I gave you some help

ìäéúøàåú

What make and model PLC? The AD DL06-DR has fast inputs availabe that can be used to catch short input pulses. The selected input can be setup via the program to do an intrupt based on these inputs.

Your not telling us about the PLC make & model, etc.. so I cant really answer your question in much detail. Check your PLC manual. If the PLC your using dosnt offer fast, or catch pulse inputs, you might have to find antother PLC for your application.

Mike

Fast input would not help here it extremely fast.
What you need Tzvi is High speed counter with interrupt option.
Omron CQM1H with CQM1H-CTB41 can hendel up to 500KHz.
You can create interrupt every pulse so it should work.
AD DL05/06 can use HO-CTRIO who can hendel up to 100Kz.I am not so familiar with the instruction set of the AD so I canot tell if it will work.
Refer to both manuael for mor details.
AD should be significant cheaper then Omron.

áäöìçä

Fast analogue??

Is there a brand of PLC with an analogue input that can do a conversion in 17 microseconds?

PLC's primary function is as a controller, not DAQ system.
I am not aware of PLC system with so fast analog inputs.
You are probably looking into PC based data acquisition
system. They cost but they can do what you ask for...

Finally some one realizes the problem!

panic mode said:

You are probably looking into PC based data acquisition
system. They cost but they can do what you ask for...

Click to expand...

You need an analog card with a DSP on it to process the data at the same rate it is acquired or a lot of memory will be required to store the data for batch processing.

No sweat. Just don't expect a PLC to do it.

Peter,

This illustrates just one aspect of my idea about having both "real-time" and "PLC-time" combined into the same system!

Imagine, just as I suggested, a long time ago, Steve remembers, that each Input bit actually consists of an Input-Word, or at least an Input-Byte.

The hardware associated with each Input-Bit indicates not only the current state of the Input but also the number of times that it has transitioned between scans.

So, to acquire the state of the Input is also to acquire the number of OFF-to-ON transitions between scans.

Pull out any Input card and look at the topography of the chips involved. There is a tremendous amount of "real-estate" available!

Now, consider the amount of "real-estate" taken up by an 8088. Each Input could easily accommodate a "local, dumb processor"!

This means that a "debouncer" can be "built-in".

Now, look at what is packed onto a Pentium. The Pentium has a "built-in cache". Something, somewhere, between an 8088 and a Pentium could EASILY provide State and programmer-validated COUNT INFO on a Discrete Input basis.

EVERY SINGLE INPUT CAN BE A REAL-TIME COUNTER as well as a simple State-Indicator.

You can then examine a particular Input based soley on "current-state" or you could examine in terms of "current-state plus accumulated transitions"!

Granted, this would require a wider bus... but buses are getting logically wider and physically smaller all the time!

TWENTY-FIVE years ago I was working with buses that were 64-bits wide! At the time, that was right behind CRAY Computing; they were using a 128-bit wide bus!

The Primary Customers of CRAY Computing are the CIA and NSA!

Where the hell do you think CRAY is now???

Geeeee...... What if.....? What if someone got their $hit together?


(171)

Peter

Why you think it would not work?
If your PLC allowed to create interrupt by high speed counter
the can hendel the speed. you can refrash your i/o and ead the value
you need big memory to do it for long time.but with smart memory manage it possible.

tzvihanan said that he wants to trigger a read of an analog input, on a seperate digital input triggered by a 17uS wide pulse. He never actually explained the application. He also never stated that he is going to read the input every 17uS. (but he did refer to 60khz, so that may be interperted as a sample every 17uS). For all we know so far, the analog value will be processed, and tossed out on each scan, and the samples are taken at one hour intervals. In other words; who said anything about data logging?? Based souly on this premis, Arik is correct in his response that a high speed counter input could be used to catch the 17uS wide pulse to trigger the read of the analog channel, and process the info. However by the time the analog sample is read, and loged, or processed, the 17uS will be long gone.

If (???) tzvihanan wants to read the analog value every 17uS and process the info within that time, then he's got a real problem that wont be solved by any PLC I know of. Also if he's data logging the analog samples he's not going to be able to use any PLC I now of either.

So my questions to tzvihanan are: What are you doing???? Why is your trigger so fast? How often are you taking samples?

Products capture or measure lithe pulses

Capture rapid signals - It my reply: www.mauell.com
or www.mauell.com.br->Menu->produtos seriados->remotas de controle
It PLC view. Product M30, PLC ME300 ME400 ME800, one possible capture 1ms digital. Other possible PLC measure widht pulse LG, it have special inputs and hadrware frequency clock, precision measure, find guy LG PLC and confirm. (Search toll PLCs.net not find LG, two letters ) Good luck

Why you think it would not work?

Elevmike gave the short answer of why a PLC can't do this.

I don't understand why people mention 1ms updates when tzvihanan clearly mentioned 60khz.

What do counters have to do with this? A counter that can count at 500khz makes no difference except as a possible trigger. To what end? A PLC can't even respond to an interrupt within 17 microseconds let alone processes the data in that time. There are many periods of time when the interrupts are turned off.

Something to think about. Why does a CPU need to turn off interrupts?

Now here is the point that NONE of you got. Do you know what anti aliasing is? Do you know how it is prevented? What kind of filtering is done on a PLC analog input card? Think about this. Any PLC analog input card that could capture data in the audio range would not filter out the normal noise that you guys except to have filtered out. I shake my head when I read of PLC analog projects where one is just assuming an analog card is an analog card is an analog card. There are or should be differences based on the application.

Terry, we use FPGAs do process analog and digital data in a manner similar to what you suggest. Debouncing is done by a small counter/timer that counts how many times an input is high or low before changing state. No big deal. It works and is taken for granted. I don't consider this to be a selling point. I don't think this feature would have saved Cray. However, if it would benefit the users I would like to know. It would be easy to use but how many people actually configure the I/O on a control logix?

Are users willing to configure each bit with a on or off delay, whether it can count or time pulses? I don't think so. In any case it would be too expensive. It can be done. We have talked about doing it since a motion controller doesn't have a lot of I/O but it would be more expensive.

We use hardware to take analog samples too. We use either DMA or FPGAs to acquire the data. Sampling data in interrupts IS NOT reliable and the quality of the data is poor. Sample jitter is just a bad as noise and PLCs ARE NOT able to eliminate it.