Old HP DeskJet/ScanJet Power Supplies Failing?

[null] asks: "Has anyone experienced problems with HP DeskJet printers (mostly 600 series) or ScanJet 4100C due to the power supply going bad? I've just got back from a friend whose scanner is currently toast as the power supply decided to put out something other than the 30VDC/400mA/14W it is spec'd for. I'm extremely suspicious because I have a DeskJet 660Cse with the same power supply brick that was working fine up until a month ago when it would go from printing fine to printing several sheets of garbage out of any given print job. I checked it with my tester and saw it giving out 37.5VDC (25% over spec). This link from HP PartSurfer lists the HP products using this power supply and has a picture. This power supply usually featuers a model number of C2175A molded on it. My friend's supply was listed as made by Lucent and mine is by NMB, causing me to wonder if the design itself is possibly defective. Google is not being very helpful in finding people with similar problems, and we are talking about pretty old hardware here that people would probably throw out rather than bother trying to order parts for. HP also has recalls on power supplies for DeskJet 800/900 series and some PhotoSmart 1x00 series printers. How about it, anyone else had problems?"

Comments

Um...

[mattster999]

Almost any power supply, unless very well regulated will show a voltage higher than its rated voltage if you test it with no load. That's not likely to be the problem.

You'll be seeing a lot more of this SOON

[Anonymous Coward]

Just submitted the following to articles...

Low-ESR Aluminum Electrolytic Failures Linked to Taiwanese Raw Material Problems
Nic Components reproduces this article from Passive Component Industry Magazine telling the tale of a defecting scientist who stole an electrolyte formula for low ESR capacitors and GOT IT WRONG! Natch that the stuff made it to market and the result could be millions of premature motherboard, power supply, and other consumer electronics failures.

Re:You'll be seeing a lot more of this SOON

[anubi]

Unfortunately, I am replying to a post of score 0. If I could mod the parent up, I would. There is a good chance this post won't ever be seen, but I will post my reply anyway for the benefit for the poster above me and the one who asked Slashdot.

This phenomena noted by AC is real. I have noted from much "autopsy" that failure of aluminum electrolytic capacitors is very common after only a few years, even if that, of service.

This is the failure mechanism: the aluminum electrolytics are highly stressed by the high pulse currents generated by switching power supplies. Within a few hours ( often in the thousands, but could be in the hundreds ), the ESR (Equivalent Series Resistance) of the capacitor begins to climb. When that occurs, the capacitor begins not only failing to accept the current pulse which was supposed to charge it, but begins to dissipate that energy, which was intended for the load, as heat. It gets hot. Fast. Then it pops open as its releases its innards of boiling electrolyte. ( Thats why you see those little grooves pressed into the top of the electrolytic capacitor cans...they are stress reliefs... they are supposed to rip open and vent the capacitor before it holds the pressure of boiling electrolyte in long enough to explode. Literally. )

I see this a lot in the main switching power supply in PC's. The tip off is watching the power supply with an analog oscilloscope while loading the power supply with things such as 12V 25watt lamps while it is powering the computer. Under running circumstances, this amount of power may only be consumed during a disk seek, but to make the phenomena continuously visible, you have to provide this extra load. If the power supply is having problems, you will see it drop voltage as its capacitors fail to supply the energy. Usually at a 60Hz rate, as only one of the two capacitors in the main voltage doubler is failing, 120HZ drops mean they are both failing.

These instantaneous and very-short lived power glitches do wonderful things to the software being executed. It makes the computer do the damndest things you ever did see. You would think you were hit by a really weird virus... and may spend months trying to track it down. I did. By the time I was closing in on this, I was seriously concerned that my debuggers themselves were infected, because weird jumps kept happening - for no reason at all as I could discern! The pattern had absolutely no rhyme or reason to it as far as I could see, and it was driving me absolutely nuts.

Being I've seen this now, I am quite quick to grab the oscilloscope and test lamps ( banks of #93 or other automotive bulbs connected to a disk-drive power connector so I can just plug 'em in to whatever spare disk-drive power connector I can find in the computer ) when I see unexplained stuff that looks too weird to be a virus.

(Note: Power-up with the bulbs connected. The step-function of the load, as well as cold-resistance of the lamps - make it very inadvisable to connect/remove lamp load while the power supply is energized. )