Hacking cartridge HP 84/85
-
- Posts: 116
- Joined: Sat Aug 25, 2018 5:26 pm
- Location: Poland
Re: Hacking cartridge HP 84/85
Last edited by MAsic12345 on Sun Sep 09, 2018 7:51 am, edited 1 time in total.
Re: Hacking cartridge HP 84/85
Because it is unlikely that throw-away cartridges have a processor on them that manages drop size in relation to head temperature and many other little nasty issues with thermal inkjet printing. Thermal inkjet printing is not a digital task at all. Why do you think that the signals are digital? Pure digital heads like the Xaar 128 discussed in another thread include a CPU and all the analog components and cost 250 Euros and more.MAsic12345 wrote: ↑Thu Sep 06, 2018 8:01 pmI would recommend that you look at all the signals with an analog scope as well
why do you think that this can be analog?
-
- Posts: 116
- Joined: Sat Aug 25, 2018 5:26 pm
- Location: Poland
Re: Hacking cartridge HP 84/85
Wonko how to watch the analog range?
- dragonator
- Site Admin
- Posts: 597
- Joined: Fri Aug 14, 2015 4:48 pm
- Location: The Nethelands
- Contact:
Re: Hacking cartridge HP 84/85
HP84 seems like a printhead, not a cartridge. On the front looks a lot like an ink input. If it is a printhead, it is a lot less disposable than lets say a HP45. You can get separate cartridges for the HP84.
I am also not sure everything required has to be on the printhead like a piezo needs. Bubblejet is a lot simpler than piezo. The official HP45 (c8855m) driver board only talks about getting consistent volume, but that is managed on a full head level and mostly comes from voltage. This can be managed externally. As far as I am aware HP45 (and I hope by extension HP84) should not have variable ink drop size.
Considering the HP84 seems like a carbon copy of the HP45, except for the pinout and the fact that it is a printhead instead of a cartridge, isn't it possible that there simply is some simple electronics in the printhead to lower the pin count? Like I have on my own controller, except inside the printhead. You can save a lot of the 52 contacts if you can send your data serial instead of a multiplexed grid.
I am also not sure everything required has to be on the printhead like a piezo needs. Bubblejet is a lot simpler than piezo. The official HP45 (c8855m) driver board only talks about getting consistent volume, but that is managed on a full head level and mostly comes from voltage. This can be managed externally. As far as I am aware HP45 (and I hope by extension HP84) should not have variable ink drop size.
Considering the HP84 seems like a carbon copy of the HP45, except for the pinout and the fact that it is a printhead instead of a cartridge, isn't it possible that there simply is some simple electronics in the printhead to lower the pin count? Like I have on my own controller, except inside the printhead. You can save a lot of the 52 contacts if you can send your data serial instead of a multiplexed grid.
-
- Posts: 116
- Joined: Sat Aug 25, 2018 5:26 pm
- Location: Poland
Re: Hacking cartridge HP 84/85
@dragonator, you are right. It is just a head with a matching tank. That is promising. I would still highly recommend to use, find, borrow and analogue scope and at least look at each signal once. If they end up all being digital, then that is a nice confirmation. If not, it may save hours and days of reverse engineering because a fundamental assumption was wrong.
- dragonator
- Site Admin
- Posts: 597
- Joined: Fri Aug 14, 2015 4:48 pm
- Location: The Nethelands
- Contact:
Re: Hacking cartridge HP 84/85
@Masic12345
While I myself do not believe it to be likely that the data pins will be analog, there should be pins that are. There are several components on the HP45 that give back an analog signal, and it might be helpful to identify if those pins are on the HP84 and which pins they are.
Coming back to your last scope, which is really interesting. I suspect I know how data is sent. Here is what I think right now. 5 Is a clock pin, at around 3Mhz by what I can see. It also seems to be on the very limit of what your analyzer can pick up. 9 and 13 seem always offset by the clock signal, and they are random enough to be data. I measured the length of 19. It seems incredibly close to 1.8us, the pulse time of the HP45. This I think it the trigger for the printhead. High is trigger. It should also be possible that each low advances the address, but that I do not know for sure. The problem of what resets the address would then need to be solved. The only mystery then is 14, 15 and 16. Another mystery is why 9 cuts out so much earlier. Does it have fewer nozzles that need to be on? (Is this the top or bottom of your image?).
Considering timing, 84us for a single burst gives 3.8us per pulse. I do however suspect it started later, because each cycle of 19 only seems to be 2.5us'ish long, and there should be 22 of them. 2 data lines and a 3MHz clock give 15 bits of data per 2.5us, an almost perfect match with the 14 nozzles on each address.
Can you get some data in this resolution, where you have several of the 5us all low sections are included? I think this is the end of a single burst (all nozzles on all addresses fired) and then we would have a few full bursts to analyze.
While I myself do not believe it to be likely that the data pins will be analog, there should be pins that are. There are several components on the HP45 that give back an analog signal, and it might be helpful to identify if those pins are on the HP84 and which pins they are.
Coming back to your last scope, which is really interesting. I suspect I know how data is sent. Here is what I think right now. 5 Is a clock pin, at around 3Mhz by what I can see. It also seems to be on the very limit of what your analyzer can pick up. 9 and 13 seem always offset by the clock signal, and they are random enough to be data. I measured the length of 19. It seems incredibly close to 1.8us, the pulse time of the HP45. This I think it the trigger for the printhead. High is trigger. It should also be possible that each low advances the address, but that I do not know for sure. The problem of what resets the address would then need to be solved. The only mystery then is 14, 15 and 16. Another mystery is why 9 cuts out so much earlier. Does it have fewer nozzles that need to be on? (Is this the top or bottom of your image?).
Considering timing, 84us for a single burst gives 3.8us per pulse. I do however suspect it started later, because each cycle of 19 only seems to be 2.5us'ish long, and there should be 22 of them. 2 data lines and a 3MHz clock give 15 bits of data per 2.5us, an almost perfect match with the 14 nozzles on each address.
Can you get some data in this resolution, where you have several of the 5us all low sections are included? I think this is the end of a single burst (all nozzles on all addresses fired) and then we would have a few full bursts to analyze.
-
- Posts: 116
- Joined: Sat Aug 25, 2018 5:26 pm
- Location: Poland
Re: Hacking cartridge HP 84/85
Here I did an analog oscillograph
impulses due to the fact that the carriage runs
just then you can see the signal
Under video numbered print head contacts are compatible with the color of the signal
you can slow down the video and see more
https://www.youtube.com/watch?v=EUITrfJoskQ&t=1s
https://www.youtube.com/watch?v=SLQ6tskNWlY
https://www.youtube.com/watch?v=RKsuzli7uYg
https://www.youtube.com/watch?v=CyaZ62IoMq4
impulses due to the fact that the carriage runs
just then you can see the signal
Under video numbered print head contacts are compatible with the color of the signal
you can slow down the video and see more
https://www.youtube.com/watch?v=EUITrfJoskQ&t=1s
https://www.youtube.com/watch?v=SLQ6tskNWlY
https://www.youtube.com/watch?v=RKsuzli7uYg
https://www.youtube.com/watch?v=CyaZ62IoMq4
-
- Posts: 116
- Joined: Sat Aug 25, 2018 5:26 pm
- Location: Poland
Re: Hacking cartridge HP 84/85
and sampled 24 megahertz
- Attachments
-
- 2.01.png (58.26 KiB) Viewed 17064 times
-
- 2.02.png (62.78 KiB) Viewed 17064 times
-
- 2.03.png (64.68 KiB) Viewed 17064 times
-
- 2.04.png (59.98 KiB) Viewed 17064 times
-
- Posts: 116
- Joined: Sat Aug 25, 2018 5:26 pm
- Location: Poland
Re: Hacking cartridge HP 84/85
then what printed
- Attachments
-
- проба полосы тонкие.png (162.27 KiB) Viewed 17073 times