TL;DR – I made a spring and fixed one part but couldn’t get the printer working again. I’ve written this to capture what I learned.
Our printer broke recently. It is a Canon Pixma MG7550. It just beeped a lot and reported error 6000: send it off to Canon to be fixed. It is about 5 years old, so I was not going to waste any money doing this if I could fix it myself. And so begins many a story of lost evenings…
I followed this excellent guide to open the printer up and take a look at what might have gone wrong. It was immediately obvious that the ink pads were completely full. This wasn’t the cause of the error, but I started off by disassembling the printer to the point where I could get them out and wash them in the bath. The volume that these pads occupy within the printer is quite something, and they were completely full. It took an hour to rinse them out. There was a lot of ink in there, and it made very pretty “storms of Jupiter” patterns in the water. Considering how often the printer cleaned its heads (every time it was switched on) it probably lost more ink to cleaning than it used in printing. And we’ve gone through at least 5 sets of cartridges. What a waste.
In the process, I found an ink blackened segment of a small spring rolling around in the bottom of the case. Eventually I spotted the rest of the spring sitting in its place in the head cleaning mechanism. Without this return spring, the little wiper which passes back and forth under the print head didn’t retract, and caused the whole mechanism to jam. Every time the printer went through its self check at startup the motor went Hnnnnnnng for a few seconds against this jam then gave up. This was the cause of the vague error: there is no way the electronics can tell why a mechanical device is stuck.
The spring is small (about 22mm long, 3.45mm diameter, and made of 0.23mm wire), but extends to 3 or 4 times its length. I asked the Ixion email list for a “Springs-R-Us” where I could buy a replacement. Lee Spring was the answer, but sadly they didn’t have anything small enough. But another suggestion was to wind your own with piano wire. Piano wire is far too fat, but guitar strings are not. Thicker guitar strings are not suitable because they are wound, but thinner strings are “plain”: they are simply a single strand of wire. In fact a plain 0.009″ top E string was exactly the right diameter. And having one in the house, and it being old, gathering dust, and in need of replacing anyway, this was perfect.
This was my first attempt, but even so it was just about usable. It used the whole string! This would not have been a surprise if I had done the maths before hand. Anyway, the result was certainly good enough to fit and try. If I was to do it again I would cable tie the ball end of the string to the screwdriver handle and wind it in a drill with a heavy weight on the other end of the string to keep everything consistent. Forming the end loops was tricky – the original has two nice tight turns in each loop. Mine are a bit mangled by the pliers.
Here’s a video of it in action, and these photos show the extremes of travel:
With this mechanism fixed and the ink pads washed and dried, everything went back together. When it was all connected, but still with the cases off so I could see everything working, I plugged it in. The power up self test which involves a great deal of head cleaning and carriage movement all went OK: the spring worked. But the process repeatedly got to a certain point and, just after moving the head back and forth a lot and flashing all the LEDs in the cartridges, reported error 6502. Having grown up with a BBC micro, I found this numberwang particularly pleasing. Wading through a number of terrible machine generated clickbait printer fixit websites, the words ink cartridge, level, and LED appeared to be important. I spotted a little sensor in the metal frame at the front which the print head passes over, and the carriage has a window above this point for each cartridge. Each cartridge has a corresponding lens moulded in its base.
So I worked out that this must all be to do with detecting the ink levels, and having the case wide open with a very bright daylight bulb in my work lamp shining at it was not helping. Sure enough with the case mostly assembled and lower light levels this error went away.
But then the printer wouldn’t feed the paper properly. A sheet would come all the way out of the front (so it was picking up OK), then a second would come out then go back in, before the printer stopped and reported 1008: no paper in lower cassette. This was clearly not the case so I was stumped again. I found another useful article, this time about feed rollers, and made the suggested fix with parts of a washing up scourer-sponge-thing. But to no avail.
At this point, having lost 5 evenings to the problem I finally gave up. It hurt, but it was becoming clear that everything I did upset something else in a complex machine that isn’t really designed for easy maintenance. Thanks Canon for no manuals, no parts diagrams, no spare parts, a printer that puts most of its ink into the cleaning pads, and a “you can’t possibly fix this: you must return it to us” repair policy.