[TODO, there's a lot that could be written about my past-endeavors with Printze!]

Today I decided to revisit Printze. My "super dooper secret idea" was proof-of-concept-implemented months ago, and Printze has been mostly sitting unused since.

The other day I picked up a Silhouette-brand vinyl cutter at the thrift store for $10... putting a pen where the blade goes makes it a plotter, a bit like the ol' HP7475A [except, without the ability to change pens]. Also managed to find a linux/inkscape plugin to get 'er running. Yes!

But... Printze's super-secret-project is quite similar, actually, so I wanted to revisit that...

So, I ran my old code... and... well, I found/fixed some minor bugs [mostly in comments]... and... it still runs!

But... here's something completely unexpected:

[Shitty contrast/macro-mode compensation, sorry]

What we have here is two stars printed, literally, one dot at a time.

It prints the way we draw 'em... It starts printing a line from the tip of the star, down and to the left on the paper... then up/right, then horizontally-left, then down/right, then up/left back to the tip.

Notice how the tip doesn't meet up!

First I thought it a software glitch... then I figured... well, most printers are only expected to go vertically /down/ the page, and here I'm having it feed the paper /backwards/... so, yahknow, shouldn't /expect/ that ability at all, nevermind its being accurate. But, actually, I've been accounting for that possibility throughout this project, so as not to be too let-down if it turned out not to be possible... [similarly, with printing dots right-to-left]

But, looking closely [which almost requires a magnifying glass... hmm, wonder if i'd get better close-ups through one? Duh!]... Looking closely, it appears pretty clear that the only mangled-motion direction is *down* the page! Exactly the direction it's designed for!

In fact, the last line-segment, which is drawn backwards in /both/ axes [up the page and to the left], is dang-near *perfect* every time! My camera can't pick it up, but you can literally see each individual dot making up each line; can see where the vertical axis apparently *completely* misses a step, but only when going the normal direction, down the page.

That "perfect" line is actually *far* more precise /and/ accurate than I'd've even expected of normal printing motions. I'm literally moving the printing position 1/120th of an inch each time. That may not sound like asking a lot, but consider that there's no encoders, just open-loop steppers, and the printer prints each dot *while moving* the carriage to the right. 

That means it has to get a running-start before the droplet, and has to decelerate after. So, every drop consists of at least a reversal, to the left of the dot, and a forward/right motion passing it. Further still, every so-often the printer inserts, on its own, a head-wipe, which brings the carriage all the way to the left, and back again for the next dot. Further still still, the two motors serve various functions *besides* x and y motion, such as the cleaning pump and head-wipers, through the use of clutches and unidirectional/ratchet-clutches. So, those motors have quite a bit of weirdly-varying loads on them. Nevermind, still, the paper rollers' grip on the paper, etc! 

I mean, it's almost shocking how precise and accurate the positioning of the dots on that one line [printed backwards! On both axes!]... i'd've been surprised if the /normal/ printing-direction turned out that good!

BUT: the /normal/ printing-direction [vertical, anyhow] is completely fubar. At first printout the error looked repeatable. About one missed step every ten. So I tried again, and again it looked about the same, except the missed steps had /shifted/. So, then, I thought maybe some sort of math-error, or maybe caused by the unidirectional clutch ratchetting... internal to the printer [e.g. maybe it took 5 extra steps to load the paper the second time...]...

So, then a third run... and that one was a /complete/ mess. The line is more of a curve!

So... now... I dunno. Am thinkink maybe it's related to power somehow. I had to replace /all/ the big power caps, which had leaked electrolyte, corroding quite a few vias and traces. I also had to replace a transistor in the boost-converter, and used a poor match [specwise] but, oddly /that/ power source drives the droplet-heaters, not the steppers...

Printze is rated 23W 9.5V. I installed a 12V-9.5V buck-converter... my 12V source is rated 2A [24W]... I suppose any of these could be factors... and, that's a *lot* of motion, back-n-forth one droplet at a time... maybe taking more power than normal.

Then there's the unidirectional clutches, etc. which may require more power in one direction than the other...

So... my original goal for the day has been put on hold; guess I've gotta open 'er up. This seems to be a recurring theme for various projects, of late, and not something I've had the energy for.

But, it is quite encouraging to see that near-perfect line of dots in the backwards direction I wasn't even sure was possible!