PDA

View Full Version : Amusing day at work.........



J Tiers
04-16-2011, 12:10 AM
I spent all day blowing things up! Can't tell you what, because it's a legal matter, but....... can say some things.

Job came in to figure out what happened to a piece of electrical equipment. The statement was that it "caused a fire"..... but it was apparent that because of the type damage, and the way it is constructed, that was unlikely. However, a suspicion isn't good enough, there has to be solid evidence of what probably happened..... facts that can be laid out without needing opinions etc.

So I had to figure out if that could have been caused by any internal failure..... Basically I had to think up failures, cause them, and see if I could duplicate the damage.

Had to keep the original piece of equipment exactly as it arrived, for evidence, but I had access to several similar ones that I could do whatever to. So first I tried various possible failures inside the equipment..... no dice, damage nothing the same as what was present.

So then I had to see what external things could have happened.

I finally came up with two ideas, both involving external faults that applied a higher voltage to a lower voltage wire.

So there I am in the lab, deliberately cross-connecting the power line to things it shouldn't be connected to, and turning on the equipment to see what happens!

I did end up getting a damage pattern that looked nearly identical to the actual one, so I figure I have it fairly well nailed down...... It's up to someone else now. As for the fire..... I figure it came first, and caused the power cross..... not the reverse, which I have a suspicion is what someone really wants to prove happened.

Lots of sparks, close-up pictures, circles, arrows, but unfortunately, in this era of digital things, no 8 x 10 glossy photos with a note on the back of each one...........

Black_Moons
04-16-2011, 01:49 AM
I fully understand people cross connecting something and it causing a fire

But how does a fire cause a cross connection??

The Artful Bodger
04-16-2011, 01:51 AM
I fully understand people cross connecting something and it causing a fire

But how does a fire cause a cross connection??

Burns through the insulation...?

becksmachine
04-16-2011, 01:55 AM
Lots of sparks, close-up pictures, circles, arrows, but unfortunately, in this era of digital things, no 8 x 10 glossy photos with a note on the back of each one...........

You can get any thing you want at .......! :D

Dave

darryl
04-16-2011, 03:19 AM
Alices restaurant- can't remember if that includes Alice-

Weston Bye
04-16-2011, 07:05 AM
...Excepting Alice.

Jerry, I get the opportunity where I work to do similar things, though usually not so spectacular.

Most recently (yesterday) I spent a good portion of the day dismantling a solenoid coil in search of a wire break and the root cause. This was difficult because we wind and terminate the bobbin, assemble the frame and pole pieces, then overmold the whole thing with thermoset (Bakelite) plastic. I had to machine away most of the overmold and metal parts, leaving the bobbin and a thin film of overmold over the coil. I then carefully machined away one of the flanges on the end of the bobbin, exposing the coil windings. From this opening, I was able to pull out most of the wire, like a slinky, intact, while carefully watching for breaks, though I didn't expect to find any there. Rather, where the overmold touched the wire, I had to carefully tease each turn of wire free of the overmolded plastic. All this was done under the microscope, as the wire was hair fine 44 gage.

Ultimately, I was able to find the actual break. Overmold material had forced its way between the magnet wire and flange of the bobbin and under the coil, pushing one turn of wire before it and eventually breaking the wire before freezing off with the two broken ends of the wire preserved, fossil fashion, in the thin semitransparent film of plastic. Once this was located, a simple microphotograph told the tale. Having the smoking gun, further simple detective work discovered a recent mold pressure adjustment made in the process that caused the problem.

By the way, all of this, except the simple detective work, was done in my home shop.

Boucher
04-16-2011, 08:30 AM
If I had a mulligan on life the field of Forensic Engineering would get a real hard look. There is actually a Professional Forensic Society. They were originally oriented toward analysis of Boiler related problems but evolved into a broader view.

flathead4
04-16-2011, 08:57 AM
But how does a fire cause a cross connection??

You've discovered the "smart fire." At least that's we've nick-named it in Nuclear industry. The NRC came out with regulations concerning train separation and redundancy following incidents at a few stations. We were forced look at all wiring that supplied safety related components and develope procedures for manually failing these components in the safe position. In some cases, redundant circuits were addded or cables rerouted to prevent re-energization due to "Fire-Induced Hot Shorts."

http://www.nrc.gov/reading-rm/doc-collections/gen-comm/info-notices/1999/in99017.html

Tom

polepenhollow
04-16-2011, 09:37 AM
You can get any thing you want at .......! :D

Dave

at Alices Restaurant
Exceptin' Alice

J Tiers
04-16-2011, 11:08 AM
I fully understand people cross connecting something and it causing a fire

But how does a fire cause a cross connection??


A fire that burns insulation and maybe causes nylon cable clamps to melt and drop the power wires onto unrelated wiring, or vice-versa.....when the insulation is melted or burned away.


I very much enjoy the "forensic" type work. Always a new thing to think about.

Liger Zero
04-16-2011, 03:13 PM
...Excepting Alice.

Once this was located, a simple microphotograph told the tale. Having the smoking gun, further simple detective work discovered a recent mold pressure adjustment made in the process that caused the problem.


Here... certain key process parameters are locked. The setup-operator I have is permitted to adjust certain things. If those don't solve the problem I am to be contacted IMMEDIATELY. Everything is documented under the threat of severe pain and or firing.

98% of the so-called experienced techs and setups I interview tell you to solve under-fill or "short shots" with more pressure.

That is wrong. That is how you solve an "under-pack."

It takes an expert (and I am such an expert) to tell if the part is not filled out (lack of material) or not packing the mold correctly (lack of pressure.

My technique for troubleshooting involves moving the fill/pack transfer setpoint and watching the graph on the screen.


Thing is companies do NOT want someone who can actually troubleshoot they want someone who can push buttons and make visually correct parts. I've had interviews where they name a defect.. I outlined my troubleshooting process... and they told me "in the time it took you to figure all that out you made a dozen bad parts."

On the other hand if you simply turn the pressure up... you've now got molded in stress or insert failure (like Weston mentioned) or dimensional inaccuracy. So I may have made a dozen bad parts but at least I am 250% sure that I have solved the problem without causing two more. :)

gary350
04-16-2011, 04:00 PM
I have seen electrical boxes explode like a bomb.

My first experience was when I was working on a 480 volt 250 amp electrical box. The electrical box had a 2" hole in the top where an electrical knock out needed to be put to plug the hole. I measured the hole and returned with a knockout plug. The maintenance paid a pair of plyers on the ledge above the box while I was gone. All I had to do was put the plug over the hole and tap it with a hammer and I was finished. 2 large air compressors were running only a few feet away and everything was vibrating. As I was about to place the plug over the hole I was something fall then BOOM there was a big explosion brighter than an arc welder. The electrical box blew apart and both air compressors stopped running. Later I found the handles of a pair of plyers welded across the 480 volt terminals of that box. The maintance mans plywers fell directly into that 2" hole I was about to plug up.

I had an electrican install something to an exist box he drilled a hole in the top with a hole saw the slug fell on top of a 60 amp motor starter directly across the 480 volt terminals. Later when he turned on the voltage it blew the box apart like a bomb. The metal door was bent into a dish shape and the box was bent all out of shape.

Once I had an old dirty contactor explode because it was old and dirty. Very old machine lots of dirt and dust collected inside until it conducted electricity the whole motor starter blew up. There was pieces all over the place about 60 pieces and was burned black as charcoal.

Stuff like that is a CSI investigation after it happens to figure out what caused it.

Weston Bye
04-16-2011, 04:53 PM
Here... certain key process parameters are locked........ I have solved the problem without causing two more. :)

Just so. In this case, the press runs two very similar but not identical products. The press was adjusted (by a manager) to solve a problem with the other product, and then later the molds were changed to run this product. Dead coils started showing up. Now the manager will have to establish two separate sets of parameters. I'm surprised that they have gone so many years using the same settings for both products. Something else has changed.

Liger Zero
04-16-2011, 04:59 PM
Just so. In this case, the press runs two very similar but not identical products. The press was adjusted (by a manager) to solve a problem with the other product, and then later the molds were changed to run this product. Dead coils started showing up. Now the manager will have to establish two separate sets of parameters. I'm surprised that they have gone so many years using the same settings for both products. Something else has changed.

Either the material is slightly different or there is a wear in the machine. Check the screw tip/NRV and the end of the barrel.

(slaps $5 on table)

Your Old Dog
04-16-2011, 05:44 PM
Interesting read. Now, STAY THE HELL AWAY FROM MY PLACE! :D

J Tiers
04-16-2011, 07:12 PM
YOD....... have you no sense of fun?



Thing is companies do NOT want someone who can actually troubleshoot they want someone who can push buttons and make visually correct parts. I've had interviews where they name a defect.. I outlined my troubleshooting process... and they told me "in the time it took you to figure all that out you made a dozen bad parts."


And how do they propose to stop making bad parts if they are not interested in WHY they are making bad parts?

Never mind, they don't really care, do they? Maybe if they were working in china they would want to do it right. In the USA it don't matter so long as it's cheap. Apparently.

Liger Zero
04-16-2011, 07:44 PM
At least at the places I've been exposed to this is the mentality:

Get it done as fast and cheaply as possible then bail out. When the **** hits the fan someone else will take the blame for it while I reap the rewards.

THIS is why they want people who simply repeat what they've been told rather than troubleshoot. THIS is why they use temp labor. THIS is why it's better to used meat-based production (people) rather than automation.

It's all about shifting the blame. That's why we have these stupid alphabet-soup programs like TQM, ISO, 5S and QS. The automakers are masters of this. It's never their fault it's always the supplier. Then the supplier doesn't correct the problem: They just pick someone on the production line and ruin them.

*shrug* Not how I personally do things. I've been there I've been through it and I got the stories to tell.

And I'll be god-damned if I put others through it.


NOW LETS GET BACK TO THE FORENSIC ENGINEERING STUFF! EVERYONE LIKES EXPLOSIONS!!! :D

J Tiers
04-20-2011, 09:52 PM
Hah.........

I finally got pictures of some of the rest of the unit that the piece I was investigating came out of......

What an anti-climax....... I was expecting fire and smoke damage, but all I saw were"mirrors"...... in this case shiny metal and a small hint of darkening of one wire. Apparently I already investigated about all the actual fire there ever was, as far as can be seen from the pictures...... Maybe they just don't tell the story.

That's not uncommon...... apparently as common as assuming the lathe bearings are bad when some issue comes up with surface finish etc...... I suspect that if what I have seen is "it", the insurance company will be saying "no deal"..... But I would think that there has to be more that hasn't been disclosed......

Maybe I get more investigatory fun later.

The Artful Bodger
04-20-2011, 10:17 PM
I noticed a round brown plastic thing screwed to the wall inside a shower cabinet in a hotel in Prague, Czech Republic. Curious, I took a closer look and noticed it was stamped "Tesla 415V".:eek:

BTW. Tesla, was until quite recently a major radio and electronics manufacturer in Czechoslovakia/Czech Republic.

ckelloug
04-22-2011, 05:19 AM
Hi Liger,

It's widely believed that TQM and things like 6 sigma is how come the Japanese beat us in the automotive industry. The basic idea behind these two is that if you track and correct root cause of every failure that occurs at a frequency of more than 1 in a million (or some number) then you will have no quality problems. GM and American companies failed to implement these philosophies when Deming started the ball rolling after WWII and look where they've gone since.

I've worked in enough strange companies to know that ISO 9000 and QS 9000 are in principle good frameworks for identifying how things were done so that you can identify the cause of a failure and rectify it. Unfortunately, these processes generate huge incompetent bureaucracies in large organized companies which tend to try to hide failures rather than account for them. In a smaller more disorganized company which still has people who have to take responsibility for things ISO can be a nice way to organize the effort of figuring out why something happened.

I got to do failure analysis of a sort too this week;)

Personally, I have been working in the new discipline of Poltico-Spheroidal Failure Analysis. Or in layman's terms, sending an engineer to determine who dropped the ball in a non-technical entirely political soap opera. I have been dispatched for a month of joyous field work to identify and correct communications failures which lead to a great deal of money being wasted and an objective need to determine who dropped the ball. My conclusion is that the only failure worse than corporate empire building at the expense of productivity is government empire building at the expense of productivity. Near as I can tell, somebody let one of the braindead chimpanzees out of the cage and after it stopped flinging poo it started making X's on pieces of paper which got interpreted as signatures on guidance documents. Don't get me started.

All the electrical failures on my project involved neurons that failed to fire. . .

--Cameron

2ManyHobbies
04-22-2011, 12:01 PM
I've worked in enough strange companies to know that ISO 9000 and QS 9000 are in principle good frameworks for identifying how things were done so that you can identify the cause of a failure and rectify it. Unfortunately, these processes generate huge incompetent bureaucracies in large organized companies which tend to try to hide failures rather than account for them. In a smaller more disorganized company which still has people who have to take responsibility for things ISO can be a nice way to organize the effort of figuring out why something happened.
This earns a $4,000,000 project to fix a $400,000 problem.

The problem overlooked in defect analysis is that somebody somewhere has to understand both the process and defect. A multi-year multi-million dollar project is an excellent demonstration that nobody involved understands either the process or the defect.

ckelloug
04-22-2011, 12:55 PM
Agreed, 2Many Hobbies. Any process or business model implemented by complete dumb-a$$es is bound to have a predictable conclusion independent of the diet of alphabet soup and tainted koolaid served in the cafeteria.

madwilliamflint
04-22-2011, 02:04 PM
You guys have the coolest. jobs. ever.

Makes me want to get out of risk analytics IT. Oh wait... I've always wanted to.