TROUBLESHOOTING METHODS? Interview question

The common thread:

1. Maintain Safety
2. Maintain Safety
3. Maintain Safety
4. Maintain Safety
.
.
.

X. Get Production Running
Y. Trace to the Root of the Problem (a lot of good approaches mentioned in previous posts)
Z. Fix the Problem so it doesn't happen again.

Don't quit until X, Y, and Z are done. In some cases X, Y, and Z all happen at once. Often X, Y, and Z become apparent when the proper homework is done to Maintain Safety!


Marc

My first rule for troubleshooting is OBSERVE. Before I touch any switches, buttons, or related controls I try to observe what the present conditions are. Too often valuable information is lost because someone has reset something or adjusted something that could have given key information about the problem. I treat the situation kinda like a detective walking up to a crime scene. Don't mess with the evidence! An exception to this is if there is a safety hazard and a machine or process needs to be shut-down right away.

I watched some electricians trying to trouble-shoot a "dead" input once. It took them forever and they got NOWHERE. They were opening junction boxes all over the place and poking around inside. Here was the process that lead to the solution. I went through this with "my" electrician to fix the problem.

1. There was more than ONE input not working. Two whole cards on the PLC were not lighting up. There should have been some inputs on because each card contained advanced and retracted status bits. One or the other HAD to be on!

2. Because the inputs on each card were fused as a group, I identified the local cause of the problem to be two blown fuses. But, I still needed to know why they blew before replacing them. I found that both circuits were still shorted to ground.

3. Using the drawing set for the project, I found the junction boxes fed by these circuits. By disconncting the wires that went from the control panel to the J-boxes we found the conductors that had shorted out (one in each case).

4. As it turned out, the first circuit we checked was jumpered to the second circuit. Removing the jumper and replacing the fuse brought one input card back online.

5. The second circuit led us to a junction box that we had reconnected wires in on the PREVIOUS day. The short was in the exact same switch that we had been servicing (coincidentally this was the same input that everyone was now griping about). Replacing the switch fixed the problem.

Here's what happened: On the previous day, when we were told that the switch was not working, someone had already serviced it BEFORE us. No one told us about any blown fuses, they just disconnected the wires. We reconnected the wires (thinking someone was just monkeying with things) and the machine started running again. Eventually, the short blew another fuse. This time, some GENIUS decided to pull power off another circuit because they were either out of fuses or sick of blowing them up. That worked for a while, apparently.

AK

I agree with rta53. It's called "FRONT PANEL MILKING"

Mickey

Some fine crack team of trouble shooters you are.

Not even once was there mention of getting a large hammer!

Whenever it's a process - related issue, I instruct the operators to start it back up, let it fault and DONT TOUCH ANYTHING!!! No reset, no e-stop, just leave it dead. Then I can usually find where the problem is occurring.

TM

Read all this over and over

When its all said and done...."go to the source"! Think about it next time you troubleshoot.

Human factors

So what do ya'll do about the human factor problems? What I mean by human factor problems is: machine operators flagging photoeyes, pushing the wrong buttons, overriding a sticky point with manual intervention....

As firing people is very hard to do here, we spend a lot of time making equipemnt "stupid resistant".

LHCD

Here we call dumbing-down the equipment "programming for the Lowest Human Common Denominator"

What we find worse are the temps. Not only are they inexperienced in running the machines correctly, but they aren't motivated to learn, since they could be anywhere else tomorrow. Unless they have a strong work ethic of their own, they simply don't care.

TM

Tim,

Try programming for an endless variaty of unknown operators, that will be totally untrained to operate your equipment, and will be provided no documentation, that you will never speek to, and that come in all sizes, ages, shapes, languages, creeds, colors, races, IQ levels (.5-200), dispositions, etc..

Thats my job.

Great answers all, one other thing to always remember is
never take for granted that the eletrician, maintenance, operator
checked everything even if they said they did. Like others said
follow your daisy chain, And remember if it's been running, chance
are its a real world problem.

Michael