benbrad

Two ways of skinning a cat. First way, used with a low number of input pulses per rev. Have a high speed timed trigger (ie 20ms pulse) feeding a word increment Then for every pulse in; RPM = 60seconds / (incremented word * time of pulse), then reset the incrementing word. Second way, used with higher numbers of input pulses. Use a timed trigger (1second) and every time it turns on check how many pulses have been counted since the last time it turned on. Then multiply by number of triggers per minute (60 for 1sec pulse), then divide by the number of pulses per revolution. Reset the counter Hope this helps Ben

I'm current using an AVG instruction in an interrupt and this appears to work. I need to know the size of an ever decreasing roll so I sample 16 times a revolution the length of stock unwound. Each sample I multiply by 16 to give me the circumference. This was fine but I found occasionaly I had a spurious figure which knocked the whole equation out. So now I average it over the 16 cycles in which samples are being taken. Seeing that the AVG command performs every scan that passes it, I figured I'd take it out of regular scan. You could put this into a cyclic task that you could call each time you need it. Use a one shot connected to TKON to call it and on the last rung of this sampling/averaging task put a P_On contact connected to TKOF. This way you call the task once with the one shot and the moment that it has done what it has to you turn it off. Hope this helps Ben

We use a DYMO label printer, only a little one, nothing fancy. Used to use it for labelling trunking lids, we used some diabolical labelling system for cables. Then some bright spark decided to use the DYMO labels on the cables and put some 3:1 clear heatshink around them. Effective and hard wearing, although we've not had to relabel anything yet, this may be a problem.

I'd have to say that although electrical engineers would probably be better than mechanical engineers, electronics engineers would probably be better than both. I'd agree with TERdON though, someone who has studied control engineering would be the best bet. Even better if they've got experience of one of the other branches of engineering.

Check the resistance between the input of the device you're switching and the ground associated with it (inputs are optically isolated from plc power). The ideal value based on a 24v system with 1.5mA of leakage would be less than 16kohms (V/I=R). If you find it has a higher resistance e.g. 20kohms, then the item will see an input even when signalled to be off. In this instance the addition of a resistor across the load would be required. For the above example of a 20kohms input impedance, use Ohms law and work it backwards i.e. 1/Rtotal = 1/R1 + 1/R2 for resistors in parallel 1/resistance required - 1/20k = 1/required load resistor 1/15k - 1/20k = 1/0.016666 1/0.016666 = 60k This is very common in the electronics industry as most devices have high input impedances but also have similar leakage currents to that of plcs. The way to test outputs to find the amount of leakage current, would be to disconnect the load and connect an ammeter in its place, as if it were the load. Make certain that the output is off before you start as a dead short could cause damage to the output module (there should be in-built protection but......). Using a bench power supply with adjustable current limit for any experiments would be the best thing to do. I use a Thurlby Thander with 2 x 32v 2A outputs as I can then have isolated circuits or reconfigure for higher voltage or higher current applications. Disclaimer - DONT DO THIS UNLESS YOU KNOW WHAT YOU ARE DOING. I cannot be held responsible if you put 24v@massive amps through a simple transistor and fry the thing. It's your fault! Hope this helps, Ben

That's great, cheers anonymous, I'll do just that. Ben

Anyone have any idea as to when they're coming to the UK? Ben

I'm a systems engineer from London, UK I do 36 - 40 hrs a week as salaried staff, but because I fall into maintenance dept I get the opportunity to go on call some weekends. Any OT on a weekday is unpaid. When on call I receive half the AVERAGE hourly rate for 8hrs/day, Saturday and Sunday. For attendance I'm paid min. 4hrs at time+half for Saturday and on Sunday it's double. The hourly rate is based on an average salary of certain members of staff (all the low paid). This is pants as I'm on about £10k above this, so I'm at a bit of a lose. Flip side is that provided there are no breakdowns I get paid for staying at home. If I get to go abroad the pay is negotiated prior to departure, each job is worth a different amount to the company. 3 months in Italy commissioning I was on 2 x Sunday rate, which at the time was more than double salary. Then 8 days in Spain on flat money +£10/day. Fine until I found that the child care for that period was £15/day. Up to 6months sick is pretty cool too, well it would be if I was ever sick.

Thank you so much Nibroc for that snippet of info, you've just cured a headache I've had for a long while. Another way to tackle this prob would be every sample that you take, raise a bit (flag) for a short while. On the NS, register an alarm/event to trigger on this flag. The alarm/event history has a time stamp built into it. You can also log 3 words of data alongside the stamp and the scroll buttons are already built in. NSDesigner allows 1024 records to be saved. After this you can either stop logging or, if set in ring mode, you can wrap round and start again, overwriting the oldest records. The alarm/event history function logs single word integers though, not floating point but these could be converted in the PLC prior to raising the flag. Ben

I didn't have a delay because the table was registered on first scan and the INI occured when an external button was pressed, thus starting the compare. You don't have to use the INI command if you want the compare to happen immediately, just change the control word of the CTBL instruction. When performing target value comparisons, execute the CTBL instruction with C=0000 Hex to register a target value comparison table and start the comparison. To register a target value comparison table without starting the comparison, execute the CTBL instruction with C=0002 Hex. Then execute INI to start the comparison.

W1.5 is a setup error. W1.6 is an execution error. Seeing as w1.5 is on, w1.6 will always come on. There's a problem with the setup of the compare, not the running of it. Is your data table valid? The only way I can get it to error is by removing the number of compares value at D100(first word of table).

Try putting an error flag contact after each suspicious instruction and get it to set/keep a bit. This should hopefully point to the problem in the code. As for masks of HSC1, I can find no mention of of this being the case in any of the pdf's that I have. I never experienced masks on HSCs. I just put your code in my CPU21 and sure enough it doesn't work. I removed the P_On contact in rung10 and put a P_First_Cycle contact in its place. Then I remove the @ on the CTBL instruction, it now works. After toggling 2960.0, A275.8 turns on. I've attached it for you to see. counter_1_test__2_.cxp

Here's the manual in case you don't have it W340_E1_07.pdf

Looking at your code you don't appear to fully set up the compare table. I've included a little example for you to have a look at but you should also have a butchers at page 776 of W340-E1-07.pdf. HS_Count_Example.cxp

File Name: Omron PLC Pass Through To NS Touch Screen File Submitter: benbrad File Submitted: 17 May 2005 File Category: Tutorials and Quickstarts Omron PLC Pass Through To NS Touch Screen Idiots guide written by myself to aid my failing memory Click here to download this file