Hacking the HP45
Posted: Mon Nov 09, 2015 5:14 pm
(TL;DR: I am going to attempt to hack the HP45 for use on the next 3DP printer. I know some stuff, and need to figure out some more. Anyone is welcome to join this valiant quest.)
I have been putting this off for far too long, getting distracted with props, and clocks, and games. I want to continue with powder printing, and I cannot design a new printer with the C6602. It is too coarse. What I need is a better printhead. Main contender: The HP45 (HP51645A). What can it do, what does the internet know about it, what do I know about it, and what do I still need to figure out about it.
The HP45 is a fairly old printhead, but it is still in use and it boasts a decent list of features, even for today. 600DPI, 300 Nozzles capable of firing at 12kHz (20kHz max). It costs roughly €30 for an original, and €15 for an aftermarket. It does have a ciss compatible system, though they are hard to get. 52 contacts on the back provide connection with the printhead. They can be refilled without too much trouble by drilling a hole and filling it. It has a hollow reservoir without sponge, held at a slight vacuum with a spring. It can be emptied using a centrifuge.
Lucky for us, HP, when making the patents for their new, fancy inkjet thingamabob couldn't decide on what to make. A datasheet or a Patent. They decided both? I provided links to 2 patents, where the second one is the most interesting. It has EVERYTHING. Pin-outs, timing diagrams, even nozzle locations. I cannot guarantee that this is exactly the HP45, but this is huge. Even ink viscosity and surface tension is mentioned.
There are 52 contacts on the back of the printhead. It has 22 Addresses (connected to the gates) and 14 Primitives (12V source). It cascades through the 22 adresses, triggering the primitives where a nozzle needs to be fired. Closing the gates while the primitives are still high will break the printhead. There are also 14 common pins on the there, but it is not known if these can be used to sink (I have been told no). The remaining 2 contacts are a so called 10X resistor and a Thermal sense (temperature sensor). How exactly these work is not really known to me.
http://www.google.com/patents/US5635968
http://www.google.com/patents/US5946012
http://spritesmods.com/?art=inker&page=2
There are also a few things that I know myself. I have someone I know, who has worked a lot with the HP45, and he has shared a lot of incredibly useful information here. The printhead, as shown in the patents, has a multiplexed array of heating resistors. The ground is common. The source and the gate make the grid for multiplexing. The voltage is 12V, both for the source and the gate. The gate requires roughly 300mA.
There is 2 parts to timing. First the gate is opened. Then there is a 2us pulse on the source to trigger the nozzle. The nozzles need to heat up and stay warm, so the nozzles need to be triggered every now and then to keep it warm. It takes 3 pulses of 2us to trigger a nozzle. Simply triggering the printhead for 6us will break the nozzle. The once every now and then pulse can also be replaced with a 200ns pulse every cycle. Theoretical max speed is 33" per second, but even at 100us per cascade and 3 cascades per trigger, it gives a respectable 5.5" per second.
I made a full detail CAD of the HP45 for future use. I hope everything is properly positioned, but a few things were incredibly hard to measure by hand. You can download the cad model for personal use here.
http://ytec3d.com/wp-content/uploads/2015/11/HP45.zip
All above is theoretical. I first need a good way to make contact with the printhead. For this I will design a set of breakout boards and a 3D printed holder. If people want to join the hacking, I can also send you a set once I have it.
With a way of connecting the printhead, I need to start tests. First I need to be able to trigger a single nozzle. The fact that both the gate and the primitive are sourced does not help, but I have a few things I can try. Once I can trigger single nozzles, I need to:
-Trigger a full address
-Trigger a full cascade
-Trigger everything in the right order
-be able to test the condition of the nozzles (measurement is explained later)
-Make the temperature sensor work
-Figure out what the 10x resistor does
-Design a controller for it (Help is really appreciated here)
-Design a shiny new printer around it
-Post everything on a set of comprehensible Ytec pages so others can benefit too
Lots to do as you see.
Now in all honesty, I will still probably not do this in one go, considering that Fallout 4 is about to happen, and I like Fallout and making props, but I at the very least wanted to make a start. More posts once stuff happens.
I have been putting this off for far too long, getting distracted with props, and clocks, and games. I want to continue with powder printing, and I cannot design a new printer with the C6602. It is too coarse. What I need is a better printhead. Main contender: The HP45 (HP51645A). What can it do, what does the internet know about it, what do I know about it, and what do I still need to figure out about it.
The HP45 is a fairly old printhead, but it is still in use and it boasts a decent list of features, even for today. 600DPI, 300 Nozzles capable of firing at 12kHz (20kHz max). It costs roughly €30 for an original, and €15 for an aftermarket. It does have a ciss compatible system, though they are hard to get. 52 contacts on the back provide connection with the printhead. They can be refilled without too much trouble by drilling a hole and filling it. It has a hollow reservoir without sponge, held at a slight vacuum with a spring. It can be emptied using a centrifuge.
Lucky for us, HP, when making the patents for their new, fancy inkjet thingamabob couldn't decide on what to make. A datasheet or a Patent. They decided both? I provided links to 2 patents, where the second one is the most interesting. It has EVERYTHING. Pin-outs, timing diagrams, even nozzle locations. I cannot guarantee that this is exactly the HP45, but this is huge. Even ink viscosity and surface tension is mentioned.
There are 52 contacts on the back of the printhead. It has 22 Addresses (connected to the gates) and 14 Primitives (12V source). It cascades through the 22 adresses, triggering the primitives where a nozzle needs to be fired. Closing the gates while the primitives are still high will break the printhead. There are also 14 common pins on the there, but it is not known if these can be used to sink (I have been told no). The remaining 2 contacts are a so called 10X resistor and a Thermal sense (temperature sensor). How exactly these work is not really known to me.
http://www.google.com/patents/US5635968
http://www.google.com/patents/US5946012
http://spritesmods.com/?art=inker&page=2
There are also a few things that I know myself. I have someone I know, who has worked a lot with the HP45, and he has shared a lot of incredibly useful information here. The printhead, as shown in the patents, has a multiplexed array of heating resistors. The ground is common. The source and the gate make the grid for multiplexing. The voltage is 12V, both for the source and the gate. The gate requires roughly 300mA.
There is 2 parts to timing. First the gate is opened. Then there is a 2us pulse on the source to trigger the nozzle. The nozzles need to heat up and stay warm, so the nozzles need to be triggered every now and then to keep it warm. It takes 3 pulses of 2us to trigger a nozzle. Simply triggering the printhead for 6us will break the nozzle. The once every now and then pulse can also be replaced with a 200ns pulse every cycle. Theoretical max speed is 33" per second, but even at 100us per cascade and 3 cascades per trigger, it gives a respectable 5.5" per second.
I made a full detail CAD of the HP45 for future use. I hope everything is properly positioned, but a few things were incredibly hard to measure by hand. You can download the cad model for personal use here.
http://ytec3d.com/wp-content/uploads/2015/11/HP45.zip
All above is theoretical. I first need a good way to make contact with the printhead. For this I will design a set of breakout boards and a 3D printed holder. If people want to join the hacking, I can also send you a set once I have it.
With a way of connecting the printhead, I need to start tests. First I need to be able to trigger a single nozzle. The fact that both the gate and the primitive are sourced does not help, but I have a few things I can try. Once I can trigger single nozzles, I need to:
-Trigger a full address
-Trigger a full cascade
-Trigger everything in the right order
-be able to test the condition of the nozzles (measurement is explained later)
-Make the temperature sensor work
-Figure out what the 10x resistor does
-Design a controller for it (Help is really appreciated here)
-Design a shiny new printer around it
-Post everything on a set of comprehensible Ytec pages so others can benefit too
Lots to do as you see.
Now in all honesty, I will still probably not do this in one go, considering that Fallout 4 is about to happen, and I like Fallout and making props, but I at the very least wanted to make a start. More posts once stuff happens.