CLI-8 Cartridges, How does the chip really monitor ink?

[panos]

maybe Canon didn't protect the printer's code like they protected the data transmission between the printer and the carts. How hard would it be for the right person to hack the printer's code so that it ignored the data from the carts?

It can be extremely hard. Canon can store the private key inside the printer ROM and use a public key inside the cartridge ROM or vice versa. They can transmit random messages encoded from the printer to the cartridge and expect valid replies. No valid replies means the cartridge is not an original. Nor the public neither the private key is being transmited over the wires, only encrypted data. Unless one finds a way inside the ROM of either printer chips or cartridge chip, the communications themselves are practically useless.

The only solution to the problem is a CIS or another printer... Or simply accept the fact of ink levels loss and regularly change your cartridges. Personally I wouldn't have a problem with the last option. My printer has printed hundreds of pages with failing colors and the printhead still prints like it did from day #1. In fact that the printer will no longer perform cleanings and deep cleanings because of empty cartirdge replacement is a bonus, I think.

 

[mikling]

It looks like the sensors are a separate subassembly. So first thing to do is to disconnect it and see if the printer continues to function with it disconnected. If it doesn't then a terminator with a similar impedance as the subassembly needs to be added.
The next step would be some type of checking system. The easiest way to do this would be to get a dead printer from somewhere that uses the prism system as we now need to cut into it. Anyhow since there is only one sensor system for all carts there is some sort of timing going on during the check. On the printer, I suspect it is done by simply turning the sensor on and off when the carriage is known to be at certain locations by the firmware as it steps the carriage motor. If we don't have access to that, we need our own signalling. The easiest way I can think of is to place some markers on the cartridges themselves but you will need to add another set of optical sensors which is not too hard. Just place a sticker with a broad stripe that changes reflectivity ( e.g. alumiunum tape) on the front so that when the cartridge is positioned over the sensor it turns on the sensor circuit and tests for ink levels. All of this can be done without any microcontrollers at all. Just a few ICs, an optical sensor and some aluminum tape. That way, anytime the head passes over the sensor, the ink levels are tested, if they are OK, the LED remains green and if they are it turns red. Once you see this, you check the cartridges.
Now it is possible that the printer does in fact turn the sensors on and off at the right time and the checking of the prism is done and but this is not reported afterwards because of the chip. If so then it is a cakewalk, since all we would need to do is to buffer the output of the sensor assembly to an external indicator system. If this were true............ Any bets?

 

[mikling]

panos, are you indicating that when the chips are defunct, the printer will no longer perform any head cleanings? If so I missed that one.

 

[pebe]

...... If so then it is a cakewalk, since all we would need to do is to buffer the output of the sensor assembly to an external indicator system. If this were true............ Any bets?

That's what I am suggesting. Provided the IR is 'live', then it would be sure to work. If I knew what the connections were to the sensor in the IR assembly, I could provide a simple circuit.

 

[panos]

mikling, no, but the printer does an automatic head cleaning when a new cartridge is installed after the old cartridge went "low". This automatic cleaning will probably not happen anymore (since the printer doesn't keep track of the ink level) and I really find it unnecessary.

 

[Tin Ho]

Ink level status display isn't really a big deal. In fact I almost never check it with my printer which have BCI-6 carts. The ink level almost never change and is always full all the time. The only time it will change is when an ink tank becomes empty. I don't really care if it works or not. What I do care is the IR detection of empty tanks.

If Canon does disable IR detection of empty tanks it is clear that the intention of doing so is for the purpose of leaving your print head unprotected against damage caused by empty ink tanks. My question would be if this practice is legal. This feature serves no purpose of increased value to customers. It's entirely for a purpose of discouraging you from refilling. It becomes a question if this violets anti-trust laws. Canon may be challenged in court why they added this feature (actually they reduced a feature by adding one) THAT SERVES NO BENEFIT FOR CUSTOMERS. This can be easily interpreted as a proof that it is not a beneficial feature they added. It's but a feature to discourage use of 3rd party competition. In a court if the grand jury agrees Canon will have to provide a solution for customers to re-enable the IR detection.

Even if this is not an anti-trust issue it violets consumer's right still. How can they set up the printer to destroy itself when it detects that you are using refilled ink tanks? It makes sense to disable ink level disply. They can argue that it is no longer accurate because of refilling. But to disable IR detection so that you will destroy the print head it is in sane. US government needs to act on this to protect consumer rights. Maybe we should begin to send letters to FTC?

 

[mikling]

Initially inkjet printers did not have this empty detect and no one complained to the courts. The fact that they added it in is a bonus and an added feature not affecting the functionality of printing.
Look at it another way, on some cars by the time the low oil light flashes it is too late, and it has caused permament damage to the motor though it will still work. The aspect that an oil pressure indicator could have prevented this is akin to the ir detect. I don't see consumers lining up to complain that the idiot light was insufficient. In fact very few cars currently have these anymore.
The printhead can survive a bit of running "empty" but extended runs will cause damage. I don't think Canon would lose this in court as they do have a right to offer limited features to encourage new cartridge purchases.
And no the printer will not destroy itself as much a BC2X machines will when they run out of ink. If the user sees bad print quality and continues to use it then whose fault is it?

 

[canonfodder]

I tried once again to detect the IR which might be coming from the LED part of the sensor. Once I thought I got a very short burst, but I could not get it to repeat that.

When you have the printer on, and open the lid, the print head is brought out to the park position. To get the print head to go through the cycle where it appears to be checking with the IR sensor, you have to momentarily lift a cartridge whose light is out, replace it and when it starts blinking, either close the lid or simulate closing the lid by pushing the switch lever. The head will seem to go home, will be lowered, and then move over the sensor, stopping at one or more positions with cartridge(s) over the sensor, then head back toward home, maybe stopping one more time over a sensor then park. It would seem that the LED must be turned on just when the processor knows a cartridge is in place to be checked. The fact that it is stopping over the sensor tells me that some action is going on, but it could just be going thrugh the motions that it always uses, whether the sensor is active or not.

I have tried to see if the sensor assembly perhaps unplugs. It does not. A tiny screw holds it in place, but behind the plastic molding which contains the LED and detector is a tiny circuit board which the 4 sensor pins are soldered into. The circuit board makes the sensor assembly captive in that it can not come out the front side where it sticks out, and there is not room to take it out the back of the space where the circuit board sits. The circuit board has 3 wires leading away from it. The wires are a purple with blue tracer, a black, and a blue. The black is common, connecting to one pin of the LED and one pin of the detector. It would seem to require some significant disassembly to be able to tack wires onto the LED and detector at the edge of the tiny circuit board. The wires disappear underneath the metal covering the area.

I picture tapping into the wires, determining the polarity of the LED power, and if finding it really is only powered part time, power it full time through a diode. Then monitor the detector through a high input impedance amplifier, providing whatever gain is needed to bring an "empty" signal up to a level suitable for driving a visible LED. If the signal only appears for very short time when the head is sweeping back and forth, a simple pulse stretching would make it last long enough to show well. I wonder if the processor will object to the detector signal showing up sometimes when it did not expect it. Any ideas? I suppose a trial will be the only sure way to find out.

I haven't even tried to see how to remove the outside cover of the iP4200 as yet. I don't think I want to turn it upside down with ink in it, but I don't really see why not. Any comments, please ? With the cover off, perhaps the sensor wires will be more available in a much easier location than down where the sensor is tucked in.

 

[mikling]

Looks like a key thing to get somewhere is a service manual. If that is available it would make hacking a lot simpler( less difficult really) as it would probably tell us when the sensors are actually in a position to be sensed. As to the speed of the flash it would be short, so a temp holding circuit might be in order.

Before jumping in though, it might be good to get a broken printer with an IR system and see what sensitivites are required and get some timings. Sounds like donations are the order of the day. I have someone offering a broken i950. Gotta drive there though. Maybe I will do it just to see what is happening to the IR system. I would imagine Canon retained similar routines.

 

[Nifty]

Initially inkjet printers did not have this empty detect and no one complained to the courts. The fact that they added it in is a bonus and an added feature not affecting the functionality of printing.

Very good point. I've become so accustomed to my level sensors on my new canons that I completely forgot that this hasn't always been the case. I also take for granted the clear plastic used to make these cartridges. Hopefully Canon won't read this, but if they really want to mess with us, keep chipping cartridges AND make them out of black plastic!

I can't find the site, but I remember years ago seeing a "system" that consisted of a 1 x 2 piece of clear plastic that you were suppose to glue to your Lexmark cartridge (after cutting a big hole in the back) so you could visually see how much ink was in the cartride... I can imagine it wasn't an easy process and probably didn't have a high success rate.