HOW TO FIX SWITCHING POWER SUPPLIES



What Can Go Wrong
Controlling those high-frequency pulses, with their fast rise and fall times, is tougher and stresses the components a lot more than did low-speed linear circuits. The fast pulses with high-voltage peaks punch holes in transistors’ substrates, and the rapid charging and discharging of filter capacitors wears those out too.

Consequently, switchers fail significantly more often than do linear supplies. Nonetheless, very few products still use the old technology; switchers are everywhere, from computer supplies to little AC adapters and chargers for cameras and cell phones.


How to Fix One
Most switchers fail from bad electrolytic capacitors, blown rectifiers or a dead chopper transistor. Look at the capacitors first. Any bulges? Change them. Leakage? Change them. Anything at all unusual about their appearance? Change them!

Checking the rectifiers is easy enough if they’re separate diodes. When you have a bridge rectifier, with all four diodes in one package, each diode must be tested as if it were a separate part. Take a look at the bridge rectifier diagram.

With all power disconnected and the big electrolytic near the bridge discharged, desolder the bridge from the board and use your DMM’s diode function to test each diode in it. You should see around 0.7 volts drop at each diode in the forward direction and an open circuit in the reverse direction, as with any silicon diode.

If you find an open or a short in any of the diodes, replace the bridge. The chopper is the big transistor, probably heatsinked, on the primary side of the transformer. Some choppers are bipolar transistors, but most are power MOSFETs. If the fuse is blown, it’s a good bet the chopper has shorted out. The transistor can fail open, too, in which case the fuse might still be good.

The transistor may have shorted and then opened, and the fuse may or may not have survived the momentary overcurrent. It’s an old technician’s anecdote that transistors are there to protect fuses! Check the chopper using the out-of-circuit techniques.

If you have an isolation transformer, you can do some powered tests before pulling parts. Check the voltage across the big cap on the primary side of the supply, near the chopper. Remember that you can’t use circuit ground on this side. The negative terminal of the cap will be your reference point, where you’ll connect the meter’s black lead. You should see at least 300 volts. If it’s much less, suspect a bad bridge rectifier. If it’s zero, the fuse is probably blown, which could mean a bad bridge, a shorted cap or a bad chopper.

It’s best not to try to scope the chopper directly, as the voltages are very high. The safer approach is to scope the secondary side of the transformer, using normal circuit ground. Many switchers have multiple taps on the secondary winding.

Any of them will do, as long as it’s not the one connected to circuit ground. If the chopper is running, you’ll see pulses at a significantly lower voltage than what’s on the other side. They won’t be tiny, though. Expect anything from 10 to perhaps 60 volts from the baseline to the peak. No pulses? She ain’t running.

If the chopper is good but isn’t running, suspect the pulse-width modulator (PWM) chip or the regulation circuitry near the output. Open zener diodes on the secondary side can allow the output voltage to rise too high, activating protection circuitry and shutting down the PWM, or even tripping the crowbar, deliberately blowing the fuse. No pulses, no chopper, no operation.

If the supply is running but not putting out proper power, caps on the secondary side are the primary suspects. Scope them. If you see much of anything but DC on an electrolytic that has one lead going to ground, change it. Either its capacitance has declined, its ESR has risen, or both. If you change the cap but the waveform still looks noisy, look for a leaky diode feeding the cap.

Finally, remember that most switchers will shut down if output current demand exceeds their safe limits. Some may blow their fuses for the same reason. A short somewhere else in the machine may be pulling too much current and causing the supply to act like it’s broken.

No comments:

Post a Comment