QY6-0067 diagnosis

[turbguy]

If you look microscopically at the LSI die that IS the nozzle plate itself, you will see a significant amount of micro circuitry contained on the LSI die. Since there are only 40 (+/-) contacts on the print head assembly (several of which just monitor print head temperature) to control several thousand nozzles, a rather simple multiplexing scheme can be arranged with on-die diodes alone that would permit firing selected nozzle heaters. But I suspect it's more complicated than just diodes, as I suspect some nozzles must fire simultaneously.

In any event, as I have stated elsewhere, the current densities on the micro-conductors on the die are astronomical. IMO, All it takes is for ONE conductor (or ONE semiconductor component) on the die to fail to take out banks of nozzles from service.

The best was to reduce the duty on print head circuitry? PRINT SLOWER (use "quite" or "night" mode in the driver).

 

[Technician]

That's good advice turbguy, I always print as slow as I can to reduce the stress & strain on all parts of the printer.
Thinking about it I do everything slow to reduce the stress on me, with luck I'll last longer too !

 

[sneezer2]

Thanks, Stratman. I would not say boldly, though. Gingerly maybe.

@PeterBJ Thanks for the references. That looks like a lot of good info. The mp610 service manual is good but does not
deal with that kind of material. I guess the engineers occasionally write something good and it may find its way into
whichever manual is current at the time. So reference back appears to be worth it.

 

[sneezer2]

@turbguy Yes, again,a little confirmation helps a lot. I had come to the conclusion that there just has to be a significant
amount of logic on the nozzle die. I've only looked at it from the outside and then only at 10X. But just stop and think.
There are 4608 nozzles on the mp610 so to drive all of them from a chip on the flexible circuit would require 4608
traces from there to the nozzle die. That would be hard to do. I can see through a part of the flexible that is a little
transparent, where there are only about ten traces. There must be more that I can't see elsewhere but probably no more than
about 30 or 40. This is partly why I asked about the number of nozzles not being addressed in this particular printhead, because
I think it may give a hint of how the logic "spreads out" from a few conductors to many. The -0067 and -0075 printheads
have 47 contacts on the back. There are two ribbon cables leading in with about a dozen conductors each. These 128
nozzles have something in common that caused all of them to stop together, perhaps one conductor on the ribbon cable.
That would imply four conductors for the pigment black and maybe four for each of the other cartridges. And they would
then have to be driven serially through each of the conductors. A very, very rough calculation but it may suggest an idea
of what should be found on the logic board. I would like very much to get a good idea of this, even if not strictly
accurate because I think it would help a lot in diagnosing printhead problems other than clogs and maybe save some
expense when making replacements.

I did see the idea about using quiet mode and am now following that. Don't remember if it was your post or not. If so,
then thanks.

 

[sneezer2]

Here's an interesting point. The ribbon cables in the ip4200 have 20 and 45 conductors respectively.
Just looking at the mp610, without getting it open and a close enough look to make an accurate count,
the ribbon cables appear to be much smaller. They look like about a dozen conductors each.
If so, this is a big change, possibly a near complete redesign and they will have had to move a
whole lot of logic from the main chassis onto the print head.
It also seems Canon themselves may have thought highly of this machine since they appear to
have kept it on the market for at least five years. That also is different. Most of the
"consumer" models seem to stay around not much more than half a year.

Just rambling.

 

[PeterBJ]

The QY6-0059 print head used in the iP4200 has 42 contact pads on the green circuit board. The QY6-0067 or QY6-0075 print heads used in the iP4500 and MP610 have 47 contact pads. The print head carriages also have contacts for the chips on cartridges and both iP4200 and MP610 use 5 chipped cartridges meaning 20 conductors to the chip contact block. Both printers also have optical readers for the timing strip and optical sensors for print head alignment on the underside of the print head carriage.

So I would expect five more conductors in the ribbon cables in the MP610, not fewer than in the iP4200.

Here is a thread with scans of the contacts on many different print heads and more info, and this post has a photo of the inner works of an iP4000 print head carriage.

 

[sneezer2]

@PeterBJ
I'm confused about this (too?). I don't have an ip4200 on hand to make any real comparison and my head is too big
to be stuck under the top cover for a really close look at these cables. Looks like some disassembly will be in order
eventually and I'm not sure when I will have the time. That said, here are some points to chew over:

1) I do have an ip4000 and had a look at the ribbon cables this morning (fat head again in the way!). They do look
very different from the ones in the mp610 and similar to what might be expected from the ip4200 specs. The one in
front is much wider than mp610 and probably does have 20 conductors. The one behind is also wide with narrower
traces and could, yes, carry 45 of them. An impressionistic look. I didn't count but very different from mp610.

2) I do not understand the apparent difference between 47 pads on the QY6-0067 (0075) and what certainly "looks"
like many fewer conductors. I expect to be wrong here in some fashion but don't yet know how. Are some pads
idle? are some doubled? Are there more conductors hiding somewhere? Maybe two layers to each ribbon cable?

3) My interpretation of the cartridge chip contacts differs from yours. I think this may call for only 3 conductors,
not 20. The outside two pads on the chips are GND and +3.3V. I recall this from somewhere but don't remember
where. Anyway, they are power and that is going to be present on the ribbon cable anyway. Cut
my call from 3 to only 1 or 2. The middle two are communication and this has to be serial. If these were all strung together
(i.e. +5 to +5, GND to GND, etc.) then serial interrogation from the main chassis could specify which cartridge it
is looking for. If, then, only that one would answer, main board could validate it then pass to interrogation of the
others.

Here is a photo of the mp610 main logic board (got it off an Ebay seller, so probably right but no guarantee)

Notice the sockets. Nothing like the CN201 and CN202 on the ip4200 board.
Photo on p.49 of the 4200 manual.

 

[PeterBJ]

I guess you are right regarding the number of conductors to the cartridge chips.

To get a better view inside the printer a small web cam can be used, here is a photo from inside an MP610 showing it has got three ribbon cables to the print head carriage, where the iP4200 only has got two. So maybe some of the missing conductors are here?

The iP4200 is a printer only model and the MP610 is an all-in-one with a scanner, so the logic board is likely larger and more complicated.

In service manuals for newer printers it is recommended that in case of a failed print head that both the print head and the logic board is replaced, so probably no official testing methods for logic boards exist?

 

[sneezer2]

Yes, I see that other cable now. It was hiding before. Can't see right off where it connects.
Maybe the shorter brown object in the upper right. I think the scanner connects separately
though.
Just took another look and the third cable seems to have maybe 25 or 30 traces on it.

I agree that there will probably be no official testing method for this. Bear in mind Canon
is a large company and issuance of an official method would likely entail some liability,
especially if carried out by an "unqualified" person. And maybe due to the design, no
definitive test could exist. That doesn't mean there's nothing that would improve
our chances. Besides that, I don't think Canon is going to stand behind you or me to
replace a logic board; probably not even a printhead, especially if we didn't buy a board
too. After all, we're not even supposed to have those manuals, or refill inks either!
Or keep using old printers when they have nice shiny new ones to sell us!

Thanks for your help. Glad you have a webcam. By the way, Milwaukee makes a slick
little fiber-optic inspection camera for about $100. I don't need one but saw one when
the inside of my furnace had to be looked at.

 

[The Hat]

@sneezer2, I reckon you not taking into account the software issue here, if the signals coming from the print head go slightly wrong, (Varied) then the EPROM chip on the logic board will without apologies present a fatal error, regardless.

That’s how Canon print heads have always operated, they also very seldom come back from a fatal error, when temperature is a major factor, and mess with it to long and the logic board does the very same thing..