Generic fault finding methodology

Hi All,

I am putting together a "fault finding guide" for some remote sites that i support.

Effectively i have stated that there are four aspects to fault finding...

1. What is the fault?
2. What does that fault mean?
3. What causes the fault?
4. How do you solve the fault?

and i have listed resolutions to those as...

1. Note any fault codes, flashing lights, etc.
2. access the device manual for fault descriptions
3. investigate from the manual what causes the fault, or contact supplier
4. investigate from manual for solving or contact supplier.

i have found that there is a tendency for guys to just recycle power on faults, and continue until the fault re-occurs, cycle power again.

any additions to my list?

wie is die doos wat dit gebreek het?
who is the ******* that broke it?

chud said:

wie is die doos wat dit gebreek het?

Click to expand...

Brings back memories.....


1. What is the fault?

This is the single hardest question your technicians/operators will ever face. 9/10 will get it wrong.
Before you even ask that question you should try to make sure they have extensive knowledge of how the system is supposed to work.

Better question would be:

1. What should the system be doing? (the more details the better)

When asking that question, most technicians/operators figure out the problem rather fast. They hate it when that happens, it means they're gonna have to fix it and not me...

in order to even know if there is a fault, you must first know what the machine is supposed to do.

a written sequence of events is required on all machinery i have dealt with over the last 20+ years, usually in the manuals.

i have had only 3 calls at home. each time, i ask the mechanic what the machine is doing when the auto cycle starts and where it stops. this information is vital.

with that information, you can direct those in trouble where to start looking in the plc program or what sensors to look at.

regards,
james

First question:
Has anything changed on the equipment recently?

Second question
Has any repair work been done on the equipment recently?
(sometimes you can ask this question, be told "no" and find out a few hours later that a repair was in fact done)

Ask and answer those two questions and you'll immediately know exactly where to look for a problem in at least a third of the cases.

All good info. so far.
I would add to be careful to differentiate "problem" from "symptom". I always see examples of "The problem is the motor is overheating". If that was truly the problem, I can solve it by spraying it down with water! Ask "WHY" 3 to 5 times, then you will be close to the root cause. As stated, the best method is to have a "profound" knowledge of what the process is supposed to do.

$.02

Ask, "What should the machine be doing next".

The fault will be with the sensor that says it's ok to continue.

An operator who can tell you this is all right by me!

The operator who tells you it's "No good" deserves the sack!!!

Oh, and the mechanic who says "It's a plc problem" can go with him to.

PLCRookie said:

Ask "WHY" 3 to 5 times, then you will be close to the root cause.

Click to expand...

We use 5-why cards. Its pretty simple, a card with a five row by two column table on it. Left column if for why question, right column is for answer. Second row question usually comes from first row answer. Its just a tool and some people don't use the card but still use the method. 80% of the time a solution is found before the card is full.

For more complex problems that have everyone stumped we sometimes employ a KT problem analysis.


The methods aren't limited to just troubleshooting - you can use it on business decisions and process improvements, etc. I even use 5-why on my kids.