Hello all!
Short version: Cleaning dust out of a printer power switch makes it work better.
Long version:
Last night, as I was preparing to render unto Caesar the things that are his, I hit the power switch on my HP LaserJet 4 and got... nothing. No LEDs flashing, nothing spinning up, no smoke, no sounds - just dead.
Power cycling, reseating the power cord, a different power cord, a different outlet - nothing. My hunch was something like a line fuse having blown, so I started to take it apart. I extracted the low- voltage power supply, which has the IEC input socket and the main power switch. It provides +5 V and +24 V (I think) to the printer logic, and switched line-voltage AC (I think) to the fuser and high-voltage power supply.
Taking the lid off, it has an input circuit breaker, not a fuse, and the breaker wasn't popped. There wasn't any obvious mayhem on the board. There was a small wad of dust lying on top of a couple of 22K resistors tied into the primary of the main transformer; removing said dust revealed a darkened circuit board in that area. None of the resistors in that area were open and the diodes in the area acted diode-y to a multimeter.
There was also small amount of oily goop on the board around the base of something like a 33 uF, 16 V capacitor at the low-voltage output. The capacitor can didn't look bulged, and I cleaned up the goop with alcohol, but since I didn't prove that it was broken, I decided not to fix that further.
I started doing continuity checks from the IEC inlet forwards. I didn't notice any difference with the main power switch open or closed, which I thought was odd. I couldn't get continuity from either line pin in the inlet to one of the power switch pins I could get at. I took the circuit board out of the bottom half of the metal case, which exposed all four pins on the line switch. I figured it was a DPST switch, but I couldn't get continuity where I should have been able to.
Time to break out the iron and solder sucker. All four pins unsoldered, and the switch carefully (like, carefully: the circuit board wants to bend if there is still a little extra solder on the pin) pried out of its hole in the heatsink... still no continuity between terminals that really should have it. This switch is mechanically weird; the actuator is a goofy shape and is operated by a pushrod from the front panel, so I figured it was unobtanium. As a short-term fix, I was prepared to add jumper wires where the switch used to be, and just let the line plug deal with the inrush, but I decided took a closer look at the switch. (Learn By Destroying, (tm) Jeff L.)
I popped the weird rocker out of the switch body with a tiny screwdriver. It came out with two attached springs. Two stamped metal contacts fell out... and so did a wad of dust.
The dust was similar to what was on the circuit board... not oily or greasy; not animal hair; just a slightly lumpy and coherent wad of dust. The switch body and rocker didn't look melted; each fixed contact had a small pit in the center of the button, and so did each movable contact. The switch didn't look like it had ever had grease in it.
Thinking I had nothing to lose, I swabbed off all the contacts with alcohol, figured out how it went back together, and reassembled the switch. The multimeter liked the cleaned switch much better; 0.1 or
0.2 ohms across either contact when closed. (Shorting the probe tips is 0.1 ohms.) I pondered rigging up a higher-current test of the switch, but then I decided just to put it back together.I got it all back together enough to run the printer. I didn't put all the printer covers back on - partly for easier smoke detection, and partly because if you completely reassemble something before testing your repair, your repair will not work.
Plug in, switch on... and action! Things seemed to be coming up as normal. When the display got to "WARMING UP", I figured the fuser lamp was probably on, and I hit the switch to see if it could open up a few amps: no drama, the printer just powered off. Hit it again: power comes back up. Off and on a couple of times: OK.
With a watchful eye for smoke, I started a print job. Ran fine. 150 or so pages later and it's still cranking along.
I'm not sure why the dust picked last night to settle in that spot. A couple of weeks ago, I *did* haul the printer to another location to do some printing there. When bringing it back, I slipped on some ice while carrying it; I managed to end up between the printer and the ground, but it would have been tilted at some odd angles. I brought it back home and set it up and I'm almost certain I turned it on when I did that.
Perhaps the IRS is getting really creative in encouraging people to e-file. I didn't *notice* any dust delivery drones lately, though.
I'm still not 100% sure if I trust the switch... it seems a little weird to do sub-component-level repair on something that has RICH, CHUNKY VOLTS going through it. I might order a (used) low voltage power supply for future replacement.
Anyway, that's the story. The firmware date codes on this thing are from late 1993. With a little luck, maybe it'll run for another 20 years. :) I wonder if that would be enough to outlast HP itself...
Matt Roberds