Monitoring Start/Stop of machine

I have been doing a scada system for a year+ as of right now and have come across a problem that I have bandaided but am sure there are better ways. But right now I jsut can not think of those better ways, so this is more of a what would you do if, type of question.

We have these gilotene(sp?) papercutters and at one point in time they were all suppose to, or at least I think they were suppose to, go into a hybernate mode after like 30sec. of inactivity. Well I am using this signal as the machine run signal to my monitoring box that collects the data from the machine. Seems my initial observation was incorrect and these machines do not go into a hybernate state.

The management wants to have a good indicator of machine run time vs machine idle time. So without this signal going on and off the data collection always things the machine is in a run state and provides inaccurate data.

So how would you determine if the machine was in a run state or an idle state accurately? I am just at wits end with these grouping of machines. Thanks and have a great day.

Frequently I look at machine current with machines with motors or other decent draw devices to determine on/off. Low current draw, off state. High current draw, on state. Low/high draw are obviously relative to a particular machine. I also use air flow switches if the machine is pneumatic. When stationary, the pressure is likely stable, with no flow. But at cycle turn-on, valves open and flow ensues which indicates a cycle start. If we know how long a cycle is, either by measurement or assumption, productivity can be calculated, or at least estimated.

Definately some good ideas here although none will work in this series of machines. Machines have same current draw in both idle and run states, there is no devices such as valves or such that turn on or off, no vacuum, hydrolics, air pressure or other current devices so that is out.

Here is the basic principle of the machine. It has one motor on it, once you turn the power on the motor starts running, this motor controls the gilotene blade, paper advance and all other aspects thru a proprietary PLC(??). The company that makes it is very secretive about their product. If you call them for any help at all it is more like nope we will send a tech your way. So there is limited information available to John Q Public.

I was starting to think about making the logic in the program that monitors things look at the material if there is no counts within a 10 second period then drop the data from the box to idle if there is material then it is in run. But some of the people around here would find way to thwart the system and gain free counts. It has to be something idiot proof.

Hope this helps.

robertkjonesjr said:

Frequently I look at machine current with machines with motors or other decent draw devices to determine on/off. Low current draw, off state. High current draw, on state. Low/high draw are obviously relative to a particular machine. I also use air flow switches if the machine is pneumatic. When stationary, the pressure is likely stable, with no flow. But at cycle turn-on, valves open and flow ensues which indicates a cycle start. If we know how long a cycle is, either by measurement or assumption, productivity can be calculated, or at least estimated.

Click to expand...

mrtweaver said:

Here is the basic principle of the machine. It has one motor on it, once you turn the power on the motor starts running, this motor controls the gilotene blade,

Click to expand...

Surely the current rises in the motor when It's cycling. Can you use a current monitoring relay like one of these, they have adjustable trip points?

http://www.ssac.com/catalog/ECS01A01.pdf

http://web1.automationdirect.com/ad...ange_AC_Current_Operated_Switches/ACS200-AA-F

http://www.crmagnetics.com/newprod/ProductView.asp?ProdName=CR4395

Do you have access to the "Knife at Up" limit switch? If so, then start a timer whenever the knife is up, reset the timer whenever it moves off the switch. Make the timer's preset long enough to cover the max amount of time it takes to manipulate the stack of paper into position or to setup the machine for a new size. The timer's "Done" state shows that the knife has remained in the up position longer than the preset and can be interpreted as meaning the cutter is not being run.

If you don't have access to the OEM's switch, can you mount your own?

I will again take a look at this type of item but as mentioned the current is very stable, it might go up an amp or two with the machine cycling, but other than that it stays relatively constant. Come monday I will again check the current draw just to make sure.

milldrone said:

Surely the current rises in the motor when It's cycling. Can you use a current monitoring relay like one of these, they have adjustable trip points?

http://www.ssac.com/catalog/ECS01A01.pdf

http://web1.automationdirect.com/ad...ange_AC_Current_Operated_Switches/ACS200-AA-F

http://www.crmagnetics.com/newprod/ProductView.asp?ProdName=CR4395

Click to expand...

Because of the proprietaryness of these machine types we did have an outside company come in and they would have gut the old controls and placed in some nice new PLC equipment with Servos and other items and guaranteed that the machine would run as good or better than before. But then things got real busy and the owner did not want to have any of his machines in a non runable state. So this idea went by the wayside. As for the buttons and controls they are mostly touch screen and are handled thru a ribbon cable that connect between the controls and the logic boards.

As for the knife up, yes this idea is somewhat feasable. I will talk to the engineering staff and the machine shop staff to see if we can proceed with this type of idea. It would have just been more simple if the OEM had made a run stop logic that was accessable from the real world. But I guess you can not have everything.

Steve Bailey said:

Do you have access to the "Knife at Up" limit switch? If so, then start a timer whenever the knife is up, reset the timer whenever it moves off the switch. Make the timer's preset long enough to cover the max amount of time it takes to manipulate the stack of paper into position or to setup the machine for a new size. The timer's "Done" state shows that the knife has remained in the up position longer than the preset and can be interpreted as meaning the cutter is not being run.

If you don't have access to the OEM's switch, can you mount your own?

Click to expand...

What are you already monitoring on this machine? Even if it's proprietary, you must at east be getting a part count. I would think that if you stop getting part counts for a duration of, say, fives times what the normal cycle tme is, then the machine is paused for whatever reason.

Well as I stated when I first started to integrate these in to our data collection system I had thought that they were auto idled. And this has not been the case. There is a small solonoid that is for a clutch release, if the machine autoidled then this clutch release would turn off. This 18V clutch release is paralleled with a 18V relay. The dry contacts of the relay are fed into the data collection system and this is the signal that I would normally use to differentiate between running and idle. But since this signal does not change then I must use some other method. What I prefer is to use a method that is idiot proof, one where the average operator could not find a way to thwart it. One where the overzealous mechanics could not destroy it. And I sure with most of the people being on here and doing this kind of thing for so many years could add to this list but you get the idea. I want it to be idiot proof and robust.

So far the prox on the knife seems to be about the best idea. It is under a guard and can not be easily tampered with.

My idea of using the photocell which is used for counting the pieces is not as good. because someone could just wipe their hand under the sensor to reset it then wipe their hands again and get free counts.

The whole idea was to make a better system than the 1986 technology that was in place. And since that system did allow people to just move their hands under the sensors and give free counts we wanted a much better way. My best thought was to have the run/idle signal from the machine montiored as well. If the machine is not running ignore the sensor. And for 80% of our machines this is a great idea. It is just these few that have to be done a different way.

I was hoing that there would be some magic wand I could use fix the problem and meet all my criteria but that is not always the case and as my wife tells me take off the rose colored glasses and see things for as they are. Oh well enough rambling. Have a great day.

S7Guy said:

What are you already monitoring on this machine? Even if it's proprietary, you must at east be getting a part count. I would think that if you stop getting part counts for a duration of, say, fives times what the normal cycle tme is, then the machine is paused for whatever reason.

Click to expand...