sporadic shut down

I've seen it done with relays. You need to use the type that you can push in by hand and have some kind of indication.
Connect a relay to each node in your chain via its own NO contact. On power up push all the relays in by hand; they'll latch via their own contact. If the chain breaks, the first relay in the series that has dropped out will indicate that the limit which opened was the one before that relay's node. - Alot of relays though for 30 limits!

There was also a device used often before the advent of plcs & SCADA called an alarm annunciator with a first out alarm facility. Don't know if anyone still makes them.

This is just a temporary hookup to find the fault, right?
Counters cost about $40.00 each. Get a few and connect across 4 switches or so. When it's open, you'll have 24AC, so you need a counter with AC input, or a diode.

Another method is meters with peak-hold, but they may eat 9 volt batteries if run for a week or so.

There are time stamping module's available for AB. It will get your resolution down to 1 ms. The link below is for 1771 and 1746, I also saw a 1756 version at the AB fair.

http://www.ab.com/db/encompass/bps_ext.prod_list_all_regs?x_cat_id=280&x_co_id=11021

A Poor Man's first-out is to place small fuses in parallel with each contact in question. This can be done with inexpensive fuse holders. Just make sure that the fuse will blow under the load of the device that's being held in by the circuit. We will use 125mA fuses in circuits that hold in starter coils...

When the unexpected shutdown occurs, just check to see which fuse is blown to determing the culprit.

-William Newton

you can also use a Pico w/ expansion for this...I used one on the corrugator splice when it would mysteriously splice on it's own. Wired the thing in parallel to the splice inputs (there were 9) and programmed the inputs to display which input fired. Had a manual clear display, using a jumper, so we could tell which one was causing it. It turned out to be the operator looking for downtime.

I have a kick *** little micro logix program for this. You just wire as manby inputs as you feel like to the micro's inputs from the e-stop string. Wire a momentary PB to the last input, and the program works by stopping any time the input word changes. and turning on the output corresponding to the input that first changed states. Press the button to re-arm it. It's so fast that I was able to detect the faulty device in a circuit where they had L1 connected to the reset PB first with the e-stop relay sealing contact in parallel, then they fed all the switches. It was impossible to tell by any other method.

When I go to work tonight, I'll see if I can find an old copy. You could probably do the same thing just based on this description.

Another method is to wire test wires to "tell-tale" relays.
. cr1 cr1
----] [----( )

Use relays with a manual operator and seal them in by pushing the manual operators. If power is interrupted even briefly, the relay will drop out and stay off. You have to relatch them to repeat the test.
Good Luck!

Newton W said:

A Poor Man's first-out is to place small fuses in parallel with each contact in question. This can be done with inexpensive fuse holders. Just make sure that the fuse will blow under the load of the device that's being held in by the circuit. We will use 125mA fuses in circuits that hold in starter coils...

When the unexpected shutdown occurs, just check to see which fuse is blown to determing the culprit.

-William Newton

Click to expand...

Here I go sounding the alarm again, but I just want everyone ELSE (I know there are many here that don't need this warning) to know that this is extremely dangerous in an e-stop circuit if you really don't take the time to do this right. If the fuse does NOT blow in an estop situation, then there could be the potential for someone dying. Depending on the relay that is being energized, the current to hold in the coil can be as low as 20 mA for an ice cube relay or 60 mA for a small AB 700 contactor. This would mean that the machine will NEVER estop.

Also be careful with the relays. Most of my circuits have the reset button and the seal in contact as the first in the chain. If any of the circuit opens, then the entire circuit opens and you will not be able to tell which relay dropped out first. If it is not this way and you don't want to wire 30 relays, then just divide the circuit in half and take a couple of times to narrow it down.

On a completely different note, does any use safety relays with monitoring? I have pretty much abandoned using ice cube relays or contactor for e-stops. Most of what I am using is the AB Minotaur series or the Banner. These usually take redundant contacts from the e-stop device and energize (or de-energize) safety relays (700-CS). A lot more work, but a lot safer.

Brucechase

Couldn't agree more about the fuses - even if they're sized right in the first place. How many times do you come across fuses which have been replaced with different values to those fitted originally?

The relay idea will only work if there's only one contact at once that's causing the problem.

We had a similar situation once: 20 to 30 E-stops in series. Machine would stop but input was still lit, had to open the panel and physically reset the MCR to start machine. We would have never been allowed to come up with a slick monitoring system like your looking for. I wish you luck, let us know how you did it.

How'd we fix our problem? Determined which stop buttons were used the most, measured the resistance across these button's contacts after removing from the system, replace the one with the highest resistance. Step two: order 20 replacement contacts and schedule replacement.

Brucechase,
Thanks for the disclaimer. I should add that this should only be done by a qualified person and the system should be extensively tested prior to putting into actual service.

With that, this method comes in handy when many of your shutdown devices are in different locations and seriesed in the field. This would not be a permament installation and should only be conducted by someone that thoroughly understands the risks involved...

-William

When I run across the "mysterious it just stops" syndrome, I look at the e-stops that seem to have been used routinely. Many times operators will use an e-stop for stopping rather than going to the actual STOP button. Watch where the operators primarily hover during operation.

E-stops can only be operated by "pounding" the mushroon head. In each case, I have found the switch is cracked or the spring is distorted from repeated poundings. Also, e-stop buttons, or rather the box they are mounted in, are good "stepping up points" for when operators need to climb on the machine. New contacts and sometimes a complete operator are necessary to repair the problem. Inspecting the individual buttons will take about the same amount of time as adding a recording mechanism.

In so far as stopping the operators from pounding the e-stop or stepping on them, even unionized operators, there is a 100% chance there is a company policy against willful distruction of company property and improper usage of e-stop circuits. Check the company rule book, you would be supprised how easy operators suddenly get the message. Most of these infractions are grounds for immediate dismissal. Even in unionized plants, there are serious repercussions. E-stops are SAFTEY devices and even the unions cannot fight willful safety violations.

GIT's safety chain seems to have a variety of contacts so it doesn't seem that E-stop buttons are necessarily a problem.

The easy solution for E-stop buttons is to use use mechanically latched buttons with a key release.

I was so busy cutting, welding, fixing rollers, chains, and air cylinders last night, that I never had a chance to look for that code.

Bruce is right about connecting anything to a safety string. I only make temporary connections and warn all operators. I usually divide the circuit into three or four pieces and narrow it down after each occurrence. Never leave it unattended. Disconnect it if no one is avalaible to check on it if something should go wrong.

The program was very simple, you could easily recreate it:

'Press the reset button to store a copy of the first 8 inputs.
'And clear the Finshed flag (B3/5)
XIC I:0.0/19 BST MVM I:0 h00FF N7:0 NXB OTU B3/5
'If not Finished and any of the first 8 inputs changes states ' then jump to ladder file 3
XIO B3/5 MEQ I:0 h00FF N7:0 JSR 3
END

SBR 3
'Set the output bits that correspond to each mis-match between
'the stored copy and the current copy of the input word
XOR I:0.0 N7:0 O:0.0
'Set the finished flag
OTL B3/5

Hope that helps, I thinks that's about how it went. I'm tired so I may have made a mistake ... triple check me!