Golden Rules of Troubleshooting

Oh wow! This one is up there with Murphy's law!

And, unless you are the manufacturer of said device, much of the above will never be available.

Jon

I've stared at well written service manuals (these actually exist, but tend to be old) that just made no sense until the next day, or after a break. Not a fan of touching service manuals with diry hands either.

I helps to try to figure out what a device should do when it is working. Extracting this information is sometimes really hard when people are fixating on what's broken and keep talking in circles.

I got a call at the plant I worked at from an equipment operator. When I got there he said an indicator light would not come on. He had changed bulbs and even swappend a glowing one with the one that would not come on to show me it was not the bulb.

I asked him about the equipment as I did not know anything about it. He said he presses one button and a light comes on , then he presses the second button and another light comes on, but now the second light would not come on. As this was just in a control room and the equipment could be located anywhere in the plant I asked him several times about the equipment and all I could get out of him was he just pressse the buttons and the lights should come. He did not seem to know where the equipment was, just the indicator lights..

Gallons of coffee are available at your local supermarket.

I practice what I preach. If really want something to work, I buy 3 of them, and cannibalize one or two for parts to make one of them work. I also buy or download the manuals: Notice that there are three HP8620C sweep generators, and three SSI/Wavetek 3000b service monitors in the photo.

If the schematics once existed, they will eventually appear on someones web pile. The sellers of manuals seem to be very good at obtaining obscure manuals. I've posed schematics for some of the stuff I've helped design on my web pile. Much more difficult are schematics with voltages and waveforms. That used to be standard practice until about the 1970's, but has lost favor probably because it requires the schematics to be printed in color.

Where I worked we had a copy machine in the shop and I would usually copy the important pages of the service manual to take with me. Usually had them already copied in a book I kept so I could fine the important pages quick and took a copy of that with me.

The electrical blue prints for much of the wiring and some equipment was on a computer and we had a large plotter so could run off what we needed to take with us. Sometimes I would make notes on the copies and leave them in the electrical cabinet, or just write inside the cabinet with a marking pen.

I also do pre-emptive replacements on caps. Worse, if I find a cap that is bulging or fails an ESR test, I replace *ALL* the caps that are the same brand and value. Also, if I find a string of parallel caps, and one is bad, I automatically assume that all of them are either bad, or will soon fail. That may seem extreme, but I spend more time extracting and replacing motherboards boards than I do fixing them. It's easier and cheaper to replace everything that is suspicious, than to deal with returns, rework, complaining customers, reputation issues, etc. If the customer returns with the same problem, but from a different cap, I have to do the rework for free, which wipes all my profit from the initial repair. Do it right and do it all the first time.

However, it also helps to pay attention. Do you see a problem here?

A blue one with a headache.

The "bad-caps" is in the URL, so that's rather obvious. However, that's the result, not the problem. The cause is quite obvious once you see it. I posted the picture previously and Phil Allison caught the problem almost instantly. I didn't.

Either the (out of focus) black capacitors are installed backwards or the board is mis-labled with the "+" signs.

Jonesy

Yep, you got it. Electrolytic capacitors are identified by a wide stripe on the negative end. To make it easy to hand insert electrolytics, the PCB is usually silk screen with a wide white area around the corresponding negative wire. Usually, except for this find Dell motherboard, that marks the positive lead instead with a wide white area. To their credit, they added "+" marks, which I missed. Having replace plenty caps on boards where the negative terminal is marked, I didn't think to look for the "+" sign, and so installed it backwards. The computer actually ran for a day or so, before it started acting funny and blew the tops open. I initially thought I had a batch of bad caps, so I replaced them a 2nd time, inserting them backwards again. I knew I was in trouble when the caps got rather warm. In desperation, I posted the photo to this newsgroup and got an instant response. Only then did I notice that the PCB silk screen markings were backwards.

Like I said, it pays to pay attention when re-capping.

Do you mean where someone mixed 85C caps (that failed) with 105C caps, that look OK?

John :-#)#

Oh...that was too obvious. Missed it!

John :-#(#

ll the caps through an ESR meter test and just replacing >those that fail. Given how fragile some of those old PCB traces can >be, I'd only want to re place the caps that really way out of spec. "

Not so much that, but the ones that matter. For example you could draw a gr aph of the dissipation of a vertical deflection IC (or transistors for that matter)against the ESR of the boost cap. Usually 100 - 330 uF in value the se boost the Vcc to the ouptu stage during retrace. When the cap ESR rises the output circuit operates in linear mode longer to try and compensate. Us ually, its dissipation is quite a bit higher before you see anything on the screen like a compression at the top or a foldover. It over heats and burn s out and about 80 % of the time the tech finds it with a foldover after re placing the IC or transistors.

So there are CERTAIN ones you want to just replace or at least check. But t hings are out there with hundreds of caps and it might be impractical to ch ange them all.

And another thing, I would highly recommend getting the unit working proper ly before recapping. Then, recap in steps, rechecking proper operation in b etween every one.

Maybe some people are lucky, but I have been where Murphy's law was strictl y enforced. And Murphy himself told me "You make your own luck". (the story of Murphy is something I will post eventually but it belongs in another pl ace)

Are *all* *four* of those large caps installed backwards?!?

Yah. Amazing you didn't have a chassis full of oily linguine there.

I know where to find coffee, although I don't drink it.

Manuals for test gear are usually available. But, for a lot of consumer gear, automotive modules and industrial gear, they may NOT be! That can be a real headache.

Old CNC control gear used to have fabulous manuals, with a theory manual three times as big as the schematics.

But, on later gear, you were lucky to get drawings of the cables, some configuration charts, and that was ALL that was ever allowed outside the factory. So, you had to reverse engineer how a lot of stuff worked to figure out what was wrong with it.

A woking unit to compare to is REALLY a huge help, when available.

Jon

Anyone care to share their experience on the correct approach to troubleshooting? I'll kick off by suggesting:

1. Assume nothing.

1. If at first you don't succeed, destroy any evidence that you even tried!

2. Try Percussive maintenance first. If it works after you bang it, it has intermitant connections, or components.

Shaun

Anyone care to share their experience on the correct approach to troubleshooting? I'll kick off by suggesting:

1. Assume nothing.

Check all user accessible parts and power components, like transistors and Diodes. After the visual check of the components for burnt parts or over heated parts and a smell test; I would flip the board over and inspect all the solder joints usually with a magnifying lens and and check for cold or fractured solder joints. When I'd find one I usually remove the old solder with a solder sucker or wick and then re-solder the joint making sure the component lead gets lots of heat during the solder process and I use real solder. Parts or the board around components that got hot often had bad solder joints from heating and cooling.

Shaun - serious post.