Troubleshooting in 10

In the July 2008 issue of 24×7, I covered 10 tips that can help you change the way you look at electronic troubleshooting. To put a formal name to the process, it would be called electronic troubleshooting using a failure analysis/experience—driven rote item prescreen—but that would be putting a complicated name to what most of us do naturally with equipment we are familiar with. We learn that a particular piece of equipment has some common failure modes, and we usually check those before getting out the schematics. The efficient thing about this list is that it produces results no matter what electronics it is applied to, which makes it invaluable to biomedical technicians and clinical engineers.

To say the following 10 tips, and the previously printed 10 items, are all-encompassing and exhaustive would insult an audience that possesses exceptional diagnostic insight. Instead, it is helpful to view the individual items as headings of chapters. Each item represents an entire category of failures; some of the items overlap. Also, because the emphasis is on speed, low cost, and easy access, some of the items should not be attempted without a specialist handy.

1) Has someone worked on it recently? Failed updates?

This happens to the best of us. We, or someone else, works on a device or does a PM, and afterward it soon fails. Recheck all work, and look for swapped boards, connectors, or faulty replacement parts. Look for unnoticed damage caused previously. Sometimes, a computer update resets settings to default and the system has to be reconfigured. It is a good practice to set a restore point before updates.

2) Are all cable pins and sockets solid?

Cable connectors should be visually checked for broken, bent, or dirty pins and sockets. Each little socket and pin should be pushed with a Q-tip wooden stick to see if they are all firmly held in the connector body. This check is especially needed on connectors that are connected and disconnected often, such as C-arm umbilicals. A shot of contact cleaner often helps.

3) Are cards seated?

Turning a device on and off can produce thermal cycling inside, which expands and contracts boards and components. Boards can be wiggled off contacts and develop loose connections. Reseating the card (in a clean card rack) clears the problem. Memory chip cards are a frequent offender.

4) Are chips firmly seated in sockets? Bent pins?

Memory chips, EPROMs, application-specific chips (ASICs), and similar devices are often press-mounted into chip sockets. Thermal cycling will actually wiggle the leads (legs) of the chip loose in the socket and produce a resistive contact. Firmly pressing the chip back into the socket (usually with a satisfying little crunch noise) will often clear up maddening intermittents or prevent them from further developing. It is often recommended to completely remove the chip and remount it, but I disagree. Simple, firm finger pressure on the loose chip does the trick, avoids static damage, and the problem of bent or damaged lead pins (broken legs). If you are following up on someone else’s mistake, look for those broken or bent lead pins.

5) Are power supplies good?

Are all the voltages present? Check voltages under load and on the board, if possible. Do not adjust if the voltages are close to nominal. Be careful with power supply adjustments if there are analog calibrations derived from that supply. If the supply is adjusted, all the downstream analog calibrations will have to be rechecked.

6) Lost fasteners in the works?

I once spent 6 months chasing a spooky intermittent in a critical multimillion-dollar device. It turned out to be a lost fastener that was found shorting out traces on a control board. Loose fasteners were found to be such a problem with Navy electronics that they specified that all fasteners be captive—unable to be lost. If possible, pick up the device and shake it. If it rattles, find the cause and remove it. If the device is too big to pick up and rattle, inspect every nook and cranny with a strong light, a small brush, and maybe some compressed air.

7) Burnt components or heat damage?

In troubleshooting and investigation there is a principle called the “Law of Parsimony,” also called “Occam’s Razor,” named for William of Occam, who first propounded the concept. The law says simply that the most obvious cause of a problem is usually the source of the problem. An example of this is when a burnt component is found on a circuit board; that component is usually the problem, and replacing it will usually cure things.

Sometimes, the circuit board will have a darkened area under a component. Darkened or charred PC boards have had their insulating and dielectric properties compromised. Radical repair is necessary to affect a temporary fix until a replacement can be installed. The next step is to determine why the component or the board is overheated and correct that.

8) Are there any cold solder joints?

When an intermittent or problem PC board is found, and reseating the board and socketed chips does not produce a fix, resoldering the most suspect solder joints (or even every solder joint) has a good chance of curing the problem. Wave solder, hot air solder, and vapor deposition occasionally will pass factory quality assurance and burn-in, only to crystallize and fail under heat load and flex stress. High optical magnification and good solder tools are necessary to find and reflow cold solder joints

9) Are diodes good?

Diodes, especially in the power supply front ends of older equipment and high-current handling diodes, are vulnerable to electrical spikes penetrating the P-N junctions and causing leakage of AC into the DC side of the power supply. More spikes continue to degrade the component until it fails. A multimeter on diode setting or a Huntron tracker will find the bad diodes. Check every diode, and replace any bad ones.

10) Facility AC OK?

AC can stand for alternating current or air conditioning. Facility electrical AC and air conditioning can be of varying quality since many engineering departments have had to cut back on electricians to do vital preventive and corrective maintenance. AC power quality and bad grounds will cause unending spooky system problems. Invest in a good power quality monitor and the new ground impedance meters. You will be surprised at how much you will find. Surge suppressors are mostly ineffective except for a few. At Edward Hines VA, Chicago, we have had a lot of success with surge suppressors from Sine-Control International Corp, Montclair, Calif. Isolation transformers also help. After the power problem is corrected, it will take about 6 months for already-weakened components to fail out. Systems run very well on quality power.

Air conditioning is the technology that makes computers and complex medical systems possible. If temperatures cannot be held rock solid around 70°F, give or take a degree or two, overheating could damage the system. The higher the temperature, geometrically more damage occurs. Damage is rapidly accelerated at 80°F, and sensitive (and very expensive) things like digital radiography detectors will definitely fail.

Humidity also needs controlling. Winter indoor dryness (less than 25% relative humidity) dramatically increases the static electricity hazard. The best fix is to get the humidity increased to at least 35% to 40% RH. The next best thing is the use of static control products.

So there you have it. View the list as dynamic, or something that will change in some respects based on technological advances and manufacturer improvements. When you are pressed for a fast repair, these and the other 10 items will work for you about 80% of the time. With practice and test equipment you will get a fix at warp speed.

---

David Meador, CBET, is a biomedical equipment support specialist at Edward Hines Veterans Administration Medical Center, Chicago. For more information, contact .