Hacking the HP45

[david]

that's great! also thanks for the information regarding the firing order and the nozzle offsets. It also makes the CD4017 scanning very challenging if you want the head able to do dual direction printing, i.e. reverse the scanning order.

[david]

also does the address line have to be 12V, If 5V can open the gate, the driving would be a lot easier.

[dragonator]

To my knowledge, the address needs to be 12V. I am not entirely sure if I have ever tested 5V. The C8855m documentation speaks of 10.8-13.2V, so I think 5V will not work. Weirdly enough, the primitives are stated to run at a minimum of 5V (this is depending on a lot of other factors though, so keep it variable up to 13.2V).

At lower resolutions the lack of scanning order should not be a problem, but if you want the highest resolutions, a better solution needs to be found. The problem though is the 12V, and the requirement of pull down resistors. Most 12V components are designed to drive, not for logic.

[david]

The 10.8~13.2V for the address voltage is a bit annoying, which means it may need a separate power while the primitives can go from 5V to 13.2V. The minimum 10.8V for the address also implies lower than this voltage the gate may not be fully open and excessive voltage drop over the gate FET can burn it.

I think for my first version I'll use CD4017 for a single direction scan. For a reverse direction scan the host computer can re-scramble the bit map to compensate the firing order.

[dragonator]

I have the whole HP45 controller on a breadboard again. This is for 2 reasons. 1: I still have a bug in the HP45 V1.01 controller (no more than about 6 primitives can fire reliably at any time) that needs fixing. 2: I want to have a breadboard circuit for the Standalone controller (Hence the LCD screen, voltage regulators and joystick).

I have already found a firmware flaw. The testing function only tested 20 nozzles. It never toggled through the primitives. This is why my new heads always showed 276 working nozzles. I have patched it so it now works. I will later rewrite it so it is more solid again. New heads now show 300 nozzles. If anyone is using my firmware, say so and I will share the new firmware.

I hope to find the reason for the lack of nozzles in the next week or 2.

Some goals for the standalone controller.

• Single voltage input (probably 19V and/or 24V)
• On board teensy 3.6
• Variable primitive voltage from 5V-13.5V controllable through the microcontroller
• Temperature feedback from the printhead
• on board temperature sensor (DS18B20)
• Optically isolated inputs for encoders and trigger signals
• SPI and serial interface. CAN bus is left free.
• Extra on-board SD card for file input (Teensy SD is pointing down so the usb is facing up)
• Optional user interface with LCD and joystick

The way it looks now the controller will consist of 3 boards of max 35mmx90mm. A controller board has the microcontroller, user interface, busses and inputs. A voltage regulator board will take in the power and convert it to all the required power levels. A Driver board will do all the controlling of the HP45. All of these boards can be changed separately later. I have finally stepped away from Fritzing and am currently learning KiCAD. This is what future project will be designed in.

HP45 standalone breadboard WIP1.jpg (393.89 KiB) Viewed 12337 times

(PS, the teensy 3.6 breakout board is really neat)

[david]

That's cool! Also, meeting my Smart Printhead Driver 45 (SPD45) ...

I'll try to get the PCB done this week.

[dragonator]

What am I looking at here. It is a new PCB in an old C8855m carrier if I see the design right. What are you using to drive the primitives and addresses. What type of microcontroller is on board. How are you going to test the nozzles. I am quite curious on how this is going to be built and how this works.

[david]

This board has similar MCU as you are using (ARM cortex). It's on the other side of the PCB. The trouble was with all the parts on this tiny board (22mmx92mm) the board was barely routed, even I had to break a few of my good practice design rules to get this PCB done. I should be able to get the PCB back from the factory next week and provide more test result.

PS the old HP45 carrier was a quite good design, I'll resuse it as much as possible. However, the original pen PCB has been a bit old. It's time to revise that part. Also, my design did not always follow the official instructions. For example, I'm using the high side current sensing to diagnose the nozzle resistors and to monitor the working of the head at real-time.

[david]

For the driver I'll give TLC59213 a go, if there is problem, after all this chip was not designed for the printhead job, I'll use the L6452.

[dragonator]

I finally got the HP45 connectors. They should give a lot more design freedom on the PCB, being surface mount. I am still working on all the 3D models and PCB footprints right now, but once I have finished those I will share. They are available to anyone who needs one. This is an early batch, and I only have ten, so it is one per person and €50 including shipping worldwide. Just PM me or use the contact form on the main site to let me know. Later batches should be a lot cheaper, but I will need quite some time before I have reached that point.

HP45 connector 1.jpg (222.74 KiB) Viewed 12545 times

HP45 connector 2.jpg (336.8 KiB) Viewed 12545 times

HP45 connector 3.jpg (204.55 KiB) Viewed 12545 times

I have the L6452 ready to test here, soldered on a breakout. (sidenote, The 100 pin ecopack is a disaster to solder.) If either the primitive problem is caused by the TLC's or if I have the breadboard free, I will try to get the L6452 to work. For now though I have the TLC driver circuit on the breadboard and I need it to stay there while designing the standalone.

I have given myself at most 35mmx100mm (and preferable 35mmx90mm) for my own controller, but because I am using a full teensy at the moment and will continue doing this for the foreseeable future, I will need more than 1 PCB. At this moment I have 3 planned, though just the printhead drive PCB should be able to work o nit's own. I have rearranged the parts. I now have the driver board connecting to the printhead. This has a 12V and 5-13V regulator on board. The controller board has the teensy and some interface components, and the Interface board has an LCD, buttons and the other I/O's

How do you plan to do real time current sensing? I have not found a component with the response time required to measure a nozzle firing.