Hacking the HP45

[dragonator]

There are 16 primitives and 2x13 addresses. Multiple primitives can be on at the same time in a trigger. Addresses are different. Only one is on at a time, always in the same order. the first 13 addresses are used from the first head, then it is switched to the second head and goes from 1-9. The chip can switch between heads within a timing cycle.

[Philipp]

That's great.
I would be interested how you plan the architecture of your control. I'm always a friend of modular systems because they allow more degrees of freedom.
Single print head, multi colour, cascaded print heads ....

I would also be interested in how you want to synchronize the position of the print head and the drop triggering, you use an encoder?
I believe that the Encoder solution much more versatile than when you synchronize the step motor position with the print head.

Most industrial manufacturers used an encoder signal, and at the begin of each swatch a PD signal (product detect). So always know the print head driver where he is and when to print. Also, the system can not miscount.

Best wishes
Philipp

[dragonator]

STUFF!

IMG_20160531_182114.jpg (347.21 KiB) Viewed 16684 times

Mouser really sends their chips in neat packing and quickly too. I can now start with the MIC5891. I got SMD and DIP so I could do the breadboard testing. Not today though, today is hackerspace day. Tomorrow I hope to connect the MIC5891 and do some testing. Maybe I can even attach the HP45 to it, but no promises.

L6452 will need to wait until I have more breadboard wires (ordered) and the smd adapter boards.

Also, I now understand why people mostly use positive LCD screens and not negative ones. They are rather dim. Not complaining though, I maintain my point that negative LCD screens are way prettier than positive.

IMG_20160531_181842.jpg (334.1 KiB) Viewed 16684 times

[ezrec]

Nozzle condition test:

I have never seen this feature before on a printer.

You may want to consider filing a patent for the concept.

If not that, you can at least use the data from a nozzle condition test to do auto-correction for dropouts when printing.

[dragonator]

The nozzle condition test is so common that the HP45 printers have it (and will prevent printing if nozzles are broken) and that the L6452 printhead driver has a constant current source to check the condition of nozzles. On my own tests I used a constant voltage and measured how much power passed but the results are the same.

Do you mean drop-outs of full nozzles (a nozzle breaking or being broken) or do you mean a nozzle not depositing ink for one or a few pulses. The first is indeed something I want to implement in the long run, the second is impossible with a nozzle condition test. For that, some other things may be possible, though tests are needed for that first.

[ezrec]

I know some HPs have an optical nozzle spit check, but didn't know they also had an electrical one.

Regardless - yes, I meant only checking for permanent failures at the start of the print run, and adaptively modify the coverage planning to make sure all dots are covered after the full left-right cycle.

[dragonator]

I do not know if I can do the recalculations required on the fly. I can probably do them when I am exporting the data from the SD card to the SRAM buffer (at the expense of extra time). For the longer term future, I am sure I can figure out a good way to implement broken nozzles into the firmware. It is not at the top of my list though, getting a working printer is. After that, I can optimize and add features.

I got to test a few things today. One is a simple 3.3V to 5V level shifter with a single NPN transistor and 2 resistors. It works really well (1-2us switch time), so I am happy. I will add it to the PCB. The LCD runs at 3.3V, but I knew that was possible.

The best thing were the MIC5891's. They behave just as I expected from the datasheet. An incredibly big thank you to Wonko for suggesting them. I apologize for the inertia when you suggested them. For now they run LED's at 5V, but they shift the data really nicely. I have the output enable tied to the ground, so I am not even turning on and off. I am simply shifting the data through. On the PCB I will add the optional option to connect the output enable to a control pin. I still need to test the speed and 12V, but they are controlled the way I expected them to.

I was going to add a video, but Youtube was laggy and it is literally 16 leds running in a sequence. Just try to imagine.

[dragonator]

This will be more or less the final design. I will test it this weekend. It requires both a 5V and a 3.3V line, but it handles 3.3V logic (as well as 5V logic). I still need to add the component labels and do a final routing check, but other than that I am done. I will verify the design this weekend on my breadboard. If I find no surprises I will order somewhere next week.

HP45 Controller V1.png (188.4 KiB) Viewed 16639 times

HP45 Controller V1 schematic.png (292.77 KiB) Viewed 16639 times

[Philipp]

Hello Dragonator,
that's great. I wonder if everything works as it should. I am not ready with my project yet. It will take a few weeks until I can do the printhead installing for the first time.
I am very excited about the L6452, if this function, the board is considerably narrower. Then the cascaded and parallel arrangements of more printheads is much easier.
I am impressed by your progress.

I have seen in the office printers that they evaluate the linear encoder on the printhead board. Could that have benefits for us?

Philipp

[dragonator]

The L6452 is not on this board, this one is still the older design with TLC59213 and MIC5891 chips. I have not yet had the opportunity to test the L6452 chips, and will not be able to for a few more weeks. This one will be the first controller. It has more or less the same features. The final width is 25mm. The printhead is only 19mm, but you also need to clamp it in place and position it, so with all mechanisms the PCB should be as narrow as the hardware.

I plan (for now) on positioning by stepper motor. This is still fairly accurate (in the 20th of a millimeter range) and accurate enough for initial tests. I am aware of the encoder tapes in commercial printers. I have one at home. While I like them and probably going to test with a DC motor and an encoder tape, I have not yet found a good place to buy them. One of my requirements for using a part is that it can be bought as an off the shelf part. If it needs to be extracted from another machine, it will not become a part for my machines (or at least the ones I share). I need others to be able to replicate my results.