Some notes on HP printer hacking
Posted by
Graham Stabler
on 2007-05-31 06:01:02 UTC
I thought I would post a few notes on my printer hacking exploits:
http://www.youtube.com/watch?v=a14zELKPw8M
First thing to note is that the paper feed encoder is actually
analogue, to get around this I added a dual comparator referenced at
2.5v to each phase, this converts the analogue signal to a digital one.
The paper encoder has 1200 lines, this gives 4800 pulses per
revolution after quadrature. The paper rollers are just over 16mm in
diameter, probably 16mm when compressed.
In order to sync the stepper with the paper feed and ignore the
general head cleaning movements I am trying a variety of methods, at
the moment my software does the following.
1. Wait for encoder frequency to rise above a given level
2. Move gantry until paper pick up sensor (a slotted opto) detects
paper pick up. It does this when the gantry moves it off a piece of
card that is blocking the beam.
3. Wait for paper feed roller to reverse.
4. Sync with paper feed, ratio of 12 quadrature pulses to one step
pulse when half stepping.
5. Wait until paper eject (gantry moves slotted opto over another
piece of card)
6. sync until speed drops below some level
7. Stop sync
What I then do is reverse the direction of the process so that if I
then print a blank page the gantry returns to the start. By
interleaving blank pages and pages with images you get fast and
automatic multi image printing which is needed for 3D printing.
Alternatively if the return speed needs to be controlled accurately
the prints could be delayed and the gantry returned by the uP.
Roger's software has this delay feature already!
One further issue is that if two prints are done one after another the
starting position of the print is different to if they are spaced
apart in time, I think this is because of the way the cleaning process
effects the paper pick up roller. I'm investigating.
If anyone wants to get into the propeller programming side or wants
the code they are welcome to it, I've not posted it anywhere yet
because its pretty messy at the moment.
Graham
http://www.youtube.com/watch?v=a14zELKPw8M
First thing to note is that the paper feed encoder is actually
analogue, to get around this I added a dual comparator referenced at
2.5v to each phase, this converts the analogue signal to a digital one.
The paper encoder has 1200 lines, this gives 4800 pulses per
revolution after quadrature. The paper rollers are just over 16mm in
diameter, probably 16mm when compressed.
In order to sync the stepper with the paper feed and ignore the
general head cleaning movements I am trying a variety of methods, at
the moment my software does the following.
1. Wait for encoder frequency to rise above a given level
2. Move gantry until paper pick up sensor (a slotted opto) detects
paper pick up. It does this when the gantry moves it off a piece of
card that is blocking the beam.
3. Wait for paper feed roller to reverse.
4. Sync with paper feed, ratio of 12 quadrature pulses to one step
pulse when half stepping.
5. Wait until paper eject (gantry moves slotted opto over another
piece of card)
6. sync until speed drops below some level
7. Stop sync
What I then do is reverse the direction of the process so that if I
then print a blank page the gantry returns to the start. By
interleaving blank pages and pages with images you get fast and
automatic multi image printing which is needed for 3D printing.
Alternatively if the return speed needs to be controlled accurately
the prints could be delayed and the gantry returned by the uP.
Roger's software has this delay feature already!
One further issue is that if two prints are done one after another the
starting position of the print is different to if they are spaced
apart in time, I think this is because of the way the cleaning process
effects the paper pick up roller. I'm investigating.
If anyone wants to get into the propeller programming side or wants
the code they are welcome to it, I've not posted it anywhere yet
because its pretty messy at the moment.
Graham