Hacking the HP45

[dragonator]

1. Switching register states should have been visible. I specifically have had the scope on the fragile address signals, and they remained stable (probably partly because they run 12V logic).

I do have at least 100uF on the primitive driver side, so that will be a 0.12V drop. Not noticeable on my scope.

3. I always thought that that was thermally, not current limited. I will see if triggering 5 pins on the same TLC will give me the same issue as triggering a full bus. That might be a good cause of this problem.

I might post a few pictures of the signals later today, just for reference.

The PCB looks nice. I can understand most components, but what are the D4 and such components doing there? I hope that it will work. I am quite curious to see it work.

PS. This is the Rigol of the hackerspace. It is default 50MHz, and can be hacked to 100MHz. I think that they hacked it, though my logic does not require it.

[david]

The D4~D8 are ESD protections for the CD4017.

[dragonator]

And as promised earlier, the scope results for a 1 pulse burst and a 3 pulse burst. The top 2 lines are 2 random primitives, the bottom line is the address. One photo is in 2us per division, the other in 5us per division. Otherwise all settings are the same.

HP45 20170508 1 pulse burst.jpg (237.9 KiB) Viewed 12135 times

HP45 20170508 3 pulse burst.jpg (243.79 KiB) Viewed 12135 times

[david]

the rising and falling edges seem to be slow. Have you applied the 20MHz BW filter? Those signals are too clean in the first picture, I always got over-shot and rings on my digital signals, maybe just my bad.

also, what do you mean by 1 pulse burst and 3 pules burst? in pic1 there's more than 1 pulse and in pic2 I don't know which 3 pules to pick.

[dragonator]

I don't know for sure if that filter was on. I might look tonight what that scope defaults to when I am at the hackerspace. Attached below is a photo of the signal when testing all nozzles. The last address was broken, so what you see at the end is the primitive driver trying to shut down but not having enough load to drain the pin. It takes quite long (millisecond range) for the pin to go completely low. I later tried with a new head and the spike at the end disappeared.

HP45 20170509 Full burst.jpg (253.43 KiB) Viewed 11527 times

The other reason why I have little noise in the first picture is because everything that is causing noise is also the thing that I am measuring. In the 3 pulse burst you do see more noise from every other primitive firing when the measured one is not. I am measuring on the driving pins, not the logic signals. I might actually do that later, just for reference.

What you are looking for in the images is what is going on when an address is high. Everything that is above the purple line going high is of interest. In the 1 pulse burst you see that the address goes high, then all primitives trigger, and then the address goes low again. In the 3 pulse burst you see the address go high, and there are 2 pulses shown, but there is space (and noise) for 3. That is because the primitives are firing in groups of 5 at a time.

[dragonator]

I somehow managed to get all the HP45 pins connected to their drivers and all other wires routed to the header. I now have the TLC59213 circuit, address circuit, test circuit and power regulation (minus the fuse) routed on board. I do still lack the variable voltage. I have little space left at the moment, but that is just because a lot of traces are in the way. At some point I had to start snaking through the PCB, making me wonder if I could do it better if I made another attempt at routing the board. At this point I give it 50/50 whether or not I can still fit the variable regulator.

@david, you make it look easy.

HP45 Standalone V0.01 Controller WIP3.png (99.52 KiB) Viewed 11485 times

HP45 Standalone V0.01 Controller WIP4.png (90.5 KiB) Viewed 11485 times

HP45 Standalone V0.01 Controller WIP5.png (95.15 KiB) Viewed 11485 times

[davidk]

Hi, can you, please, tell more about variable voltage regulator? Why do you need one? Thanks!

[dragonator]

I am following the original documentation for the C8855. This states (along with heatup cycles and timing) that it has a variable primitive voltage. It also has a calibration cycle where it deposits 1000 drops of ink per nozzle, waits a bit, and measures the current printhead temperature. An example graph shows a valley at the voltage where the ideal current print voltage is. Enough to create the bubble, but not too much. This voltage probably shifts depending on air pressure and the temperature of the printhead. The voltage is somewhere between 8 and 13V.

I suppose you can run the head simply from 12V. I have for a while. It will just run hotter and has less life this way. I just wanted to make the next printhead controller in such a way that it more closely resembles the original and test function including variable printhead voltage.

[david]

An updated version, the test result will follow soon ...

[davidk]

Hi, what are the chips do you use on this pcb?