Hacking the HP45

[david]

there is also a USB to the printhead interface.

[dragonator]

Am I mistaken if I say that that is the L6452? It looks like the 100 pin ecopac that it is housed in. Curious to see how it works.

I managed to get everything on a single PCB. I finished the driver board, containing the voltage regulators (with fuse this time), check printhead circuitry, primitive circuitry and address circuitry. The variable voltage regulator has both a potmeter and a digital potmeter, so it can be used without the SPI, simply set to a fixed voltage. I also have additional CS pins available so I can make a replacement board that has an SPI I/O expander to drive the primitives. The board has grown from 35mmx90mm to 40mmx90mm to give me a bit more room to work with. Anyone placing the board side by side will have to wait for me to design the 2 and 5 printhead versions a lot further down the line. The header has enough space for a boxed variety to allow for direct microcontroller access through a ribboncable.

I still need to modify some of the resistor values on the variable voltage regulator because it is still set to the 12V settings without the potmeter. Labels and markings are still missing on the board. I also need to drag around some of the traces and might reroute some traces to remove vias, but everything I need is there right now.

I will probably start with the controller layer and interface layer this weekend. Once those are decent, I will order the PCB's and components to assemble this board. That will also be my first attempt at using the HP45 connector.

HP45 Standalone V0.01 Controller WIP6.png (71.46 KiB) Viewed 12056 times

HP45 Standalone V0.01 Controller WIP7.png (85.34 KiB) Viewed 12056 times

HP45 Standalone V0.01 Controller WIP8.png (80.11 KiB) Viewed 12056 times

[davidk]

Interesting, I like the design. Do you think now you'll be able to fire all 14 primitives at once? I see you use the same TLC59213 which can supply 200mA per output when all output source power.
Another question, why do you need Q3 and the comparator?

LE: Ok, reading page 8 again gave me the answer about mosfet; it's for nozzle test.

I decided that it would best forme to design the head driver with an ATMEGA64 that controlls 5x TLC59213 for addresses and primitives. Printing data comes trough SPI This way I can develop the firmware for the head waveforms then focus for the application firmware on Arduino or a different AVR. The head driver would be addressed in the way like Xaar128 does. I wouod apreciate if someone could attach the head footprint for kicad so I won't do it from scratch. I'll come back with the schematic when ready.

[dragonator]

Attached is what I think should be importable into Kicad, including a rough drawing of the connector. This is by the way not the pogo pin pads, it is the connector I received a few weeks ago. I did some final checks with a file I got later, it is at most 0.05mm out, so for now I am not going to alter the model.

Also for Kicad the 3D model to attach to the footprint. Scale and position is X,Y,Z scale: 0,3937; Y offset 0,655; X rotation: -90. If you have any issues let me know and I will try to fix. More news later, but I am busy right now.

[dragonator]

I did a few tests. From what I can see, if at any time more than 3-4 nozzles are pulsing on on a single TLC59213, the head will glitch. I tried a bunch of patterns and any time I tried to fire all nozzles in 2 pulses in any configuration, nothing. Most patterns going in 3 and 4 pulses worked. I then added a 3rd TLC59213 and got a few patterns that did allow for firing in 2 pulses.

I am still not entirely sure if that is the whole story, but for now that is what I have found. My conclusion is that if you want to fire every nozzle at once, you will need at least 4 TLC's, maybe even 5 (any test I did with 4 did not result in success). With a third TLC59213, you can get away with 2 pulses, but if this is worth the extra driver I do not know.

I do get the feeling that there is more at play here than I see. The pattern I get when I try to fire all nozzles at once stays the same whether I fire it with 2, 3 or 4 TLC drivers. I do however not have any clue why, and so I will consider adding the 3rd TLC driver on my PCB (I have space) and leave it at that.

[davidk]

I was thinking of parallelling two TLC lines for each primitive. This way we have 400mA for each primitive.

[dragonator]

I will try that on the breadboard if I do not forget about that.

Last post of the day. Controller and Driver board are now both 90% finished. Driver has 3 TLC drivers instead of 2, plus some minor tweaks. Controller has SPI, Serial, USB (on the teensy) and opto isolated inputs and outputs for encoders, trigger signals and feedback signals.

HP45 Standalone V0.01 Driver WIP1.png (64.61 KiB) Viewed 12003 times

HP45 Standalone V0.01 Driver WIP2.png (58.87 KiB) Viewed 12003 times

HP45 Standalone V0.01 Controller WIP1.png (70.6 KiB) Viewed 12003 times

HP45 Standalone V0.01 Controller WIP2.png (87.36 KiB) Viewed 12003 times

[david]

That's what I've got so far from ST regarding the L6452. So I'm wondering if I should just use discrete transistors to drive those nozzles.

SOLUTION PROPOSED BY SUPPORTER - 5/5/2017 18:07:51 :
---------------------------------------------------------------------------------
Hello,

What is your production forecast in thousands and what is your company's project name? What is your company email? Which distributor do you normally work through? What region is your request associated with?

Regards,
Support Team


Description:
Hi, I'm trying to use the L6452 to drive a printhead. However, in the datasheet there is not enough information regarding how to connect the boost voltage, i.e. pin 87,88 and 89 as well as related specification, such as min boost voltage and max boost voltage. Is there any more detailed user manual available for L6452? any instruction on how to use these three pins or even a reference schematic would be very helpful. Many thanks,

[dragonator]

That is a rather disappointing answer. I would just tell them that a: this is not for thousands of units and b: that his answer is not relevant for the question asked. The documentation is lacking and we want to know the details. We would still benefit from knowing this.

I finished the last layer today. The interface layer. It hardly has anything on it, so it was not that much of a challenge. I will now do final checks, test tiny blocks on the breadboard and then order all layers.

HP45 Standalone V0.01 Interface WIP1.png (40.06 KiB) Viewed 11961 times

(PS, answering if on the scope results a few pages back the 20MHz BW filter was on. 95% chance no)

[david]

can any one help to confirm what's the chip number on the original HP45 pen drive (C8855m)? I got two possible versions: 1821-5613 and 1825-0057, also the 1826-4353 seems capable of same job.