Inkjet printer direct to PCB (printer hack/conversion)?

[HankMcSpank]

Just wondering if anyone here has done one on here?

I have an old HP printer lying around which uses pigment ink...I'm tempted to have a dabble. (basically have a lead screw which pulls the media mounted on a flatbed underneath the print head as it does its thing)

In my naive reckoning, I think it'd only need a paper feed 'servo' to stepper converter (a counter) to get the bed feed rate going at the correct rate.

[mark_s]

Hank,
Have a look at this project.

http://techref.massmind.org/techref/pcb/etch/c84-st.htm

I found one these printers in a second hand store $10, but have not
had time to give it try. There are also some videos on youtube.

Motherlode of info

http://www.cnczone.com/forums/showthread.php?t=30951

Mark

[HankMcSpank]

Hank,
Have a look at this project.

http://techref.massmind.org/techref/pcb/etch/c84-st.htm

I found one these printers in a second hand store $10, but have not
had time to give it try. There are also some videos on youtube.

Motherlode of info

http://www.cnczone.com/forums/showthread.php?t=30951

Mark

Hi Mark,

Thanks for the links....I've gots me a 'epson' donor (I coaxed my work into scrapping it - it's sat collecting dust in the storeroom for ages - they also had some unopened Durabright ink cartridges in the stationary cupbaord :-) ).

A perfectly functionable Epson C66, about to be mullered...




Sidecheeks ripped off (literally), after which the case shroud popped off easily enough...




Not sure how much time I'll get to proceed (other things are sumping my time), but just wanted to case the job & get a general vibe for the task at hand

[Ioannis]

Please do post your results.

It is something that bothers lot of people...

Ioannis

[HankMcSpank]

Please do post your results.

It is something that bothers lot of people...

Ioannis

I will do...but I'm one of those guys that needs the project to "talk to me" - no I don't mean I'm going insane, but now I've stripped it down, there appears to be several options & there's no clear winner at the moment...so rather than rush into it, over the coming days/weeks eventually one option will percolate to the top.

Quick option - have this printer print onto pcbs only (that would just mean raising the print head a little - easy)

Less quick option - convert this to a flat bed printer, where I can feed all manner of stuff below the print head. This would need a significantly different approach (much more clearance need below)...the roller 'feed' would need to be binned - a leadscrew would likely need t be the way etc etc.

the initial probem facing many conversions is that there's an optical 'paper detect' sensor - the printer feeds the paper in, but then backs it out a little until the leading edge of the paper is detected - it the proceeds to print .....now that's a great idea when printing on paper, but for pcbs...all the to'ing & fro'ing is not welcome.

so at this early stage, I can feel a pic being brought into the equation to spoof the optical sensor! (something that surprisingly nobody seems to have done - they mainly go low tech with a mechanical approach to spoofing the optical sensor)

[Ioannis]

Don't worry about the time this will take. I'll be subcribed to this thread for ever (sort of speak...)

Now about the sensor, why that bothers you?

If it works, leave it. I suppose it contributes to the initial position of the paper/pcb, right?

Ioannis

[mackrackit]

If you have some sort of tray to place the PCBs on then the sensor could be used as a locator as it is designed to. Then when the etching is finished the PCBs could run again for the silks.
Maybe...

[HankMcSpank]

Now about the sensor, why that bothers you?

If it works, leave it. I suppose it contributes to the initial position of the paper/pcb, right?

Ioannis

At this early stage, I'm not even sure why it bothers me!

I know it's certainly appears to be a hassle for a lot of diy 'direct to pcb' printer conversions.....it's my understanding that it's only role is to find the leading edge of the paper - but the end goal here is near perfect repeatability & 'registration' - not sure how achievable that will be with a system that use friction rollers & an optical edge finder.

....but like I say, it'll become clearer in head the best strategy for me to pursue over the coming days.

[Ioannis]

I do not like the friction too. And this technology is not suposed to have good repeatability. Every time the rollers move the paper/pcb/whatever, they are losing a little of their diameter. So it is sure that this cannot be very accurate.

Another idea might be to use the steppers that move the rollers, to move a table having the pcb on it. Difficult as there are no datas about the motors and the gear involved.

Ioannis

[HankMcSpank]

I do not like the friction too. And this technology is not suposed to have good repeatability. Every time the rollers move the paper/pcb/whatever, they are losing a little of their diameter. So it is sure that this cannot be very accurate.

Another idea might be to use the steppers that move the rollers, to move a table having the pcb on it. Difficult as there are no datas about the motors and the gear involved.

Ioannis

Well fortunately this particular printer (Epson C66) appears to use a stepper for the paper feed motor, so in principle it should be just a matter of feeding those wires to either the original stepper but relocated to the front/middle of the printer driving a leadscrew (or use my own stepper - there's seems a little small!). Then just a matter of dividing the pulses to achieve the right feed rate.

I believe someone has done this (albeit with an HP printer) as I saw a youtube video relating to it.

The hour or so of dabbling to date....shows I'm in diffs already. Apparently the optical sensor has to be relocated very precisely to get the right timing - else the printer just gives a paper feed error. Hey ho!

[HankMcSpank]

Ok, making some progress.

I've decided on the easy option - don't dick with the head height & simply feed 0.8mm copper into the printer! (no outlandish hacking needed)...

This is 0.8mm double sided copper....



So the procedure is...

1. Buy a secondhand epson printer that uses durabrite ink (this is pigmented ink & means your 'possible contender' printer is capable of doing direct to copper printing)

2. Remove just about all the paper feed mechanisms *except* for the friction roller which pulls the copper through)

3. Relocate the optical sensor to a more appropriate place (this was an earlier test using paper to establish how to spoof the sensor)...



(that's the optical sensor to the right of picture with the blu-tack holding it in place)

4. Tape a little bit of card to the copper (to spoof the optical sensor that it has found the paper edge)

An then reference this video for the actual feed sequence...




About 2 hours work in total (less now that I've done one & therefore know the shortcuts!)

[Ioannis]

OK. Well done so far.

What about the printed PCB? Is the ink solid and dry on the copper?

Have you tried to etch it?

Ioannis

[mark_s]

Hank,

Great progress! You might inspire me to rip my Epson apart soon.
My big concern was the ink. Some of the guys were having trouble while others
seemed to be getting ok results. Need to research some more

I was thinking of taking a couple of ball bearing drawer slides and mounting a tray on top to hold the pcb. This would keep the alignment perfect and lower the friction.

Keep us updated

Regards
Mark

[HankMcSpank]

OK. Well done so far.

What about the printed PCB? Is the ink solid and dry on the copper?

Have you tried to etch it?

Ioannis

This (Epson) ink smudges (even after an hour) - but I knew this would be the case. Alas you have to bake the board to 'fix' the ink.

I've not baked nor tried to etch - this is just early experiments (clearly I need a better method to feed the copper in) - also it would appear a couple of the nozzles are blocked (I'm seeing the odd horizontal streak/line where no ink is on the pcopper)

Hank,

I was thinking of taking a couple of ball bearing drawer slides and mounting a tray on top to hold the pcb. This would keep the alignment perfect and lower the friction.

Mark

I'd say that's got to be a good idea - I'm thinking about a 'bed' behind the printer which the pcb rests on - but with adjustable cheeks, to align/guide the pcb where I want it (I can't be done with the faff as seen here - ...not knocking his achievment, as that video helped me a lot, but IMHO that's what the net is all about...being inspired by others aaccomplishments & chievments to hopefully take it to the next level).

So I'd say 'go for it'....it's surprisingly easy (if you're happy printing onto 0.8mm copper clad that is!)

[mark_s]

Hank,

I agree, advance it beyond those before, ditch the cardboard and bubble gum. But if you can't get the ink to dry in a consistant manner, there is no use advancing the mechanism.

You could print out some nice capacitive switches and keypads with the flexiable pcb material.
I might try this first with an unmodified printer. This way I can get a feel for the ink.


P.S. If we could determine if the same ink used in "Sharpie" pens would function in an epson print head. This ink dries fast and is etch resistant.

[HankMcSpank]

Hank,

I agree, advance it beyond those before, ditch the cardboard and bubble gum. But if you can't get the ink to dry in a consistant manner, there is no use advancing the mechanism.

You could print out some nice capacitive switches and keypads with the flexiable pcb material.
I might try this first with an unmodified printer. This way I can get a feel for the ink.


P.S. If we could determine if the same ink used in "Sharpie" pens would function in an epson print head. This ink dries fast and is etch resistant.

Now I'm no expert on printer ink, but it seems there are a two types...

1. dye ink (which is what most inkjets use) - water soluble - no use to us for etch resist.

2. pigment based ink (as used by the imkjet modders) - prints out wet, so needs to be baked to dry it. (& this is where inkjet pcbs turn into art form - the process of pre-prepping the board & drying the ink.



Solvent inks evaporate (likely to be what sharpie pens are)....ideal for pcbs, except it dissolves the glue in the epson print heads! this is why pigment based inks are the only option.

[Ioannis]

You can use AZAX or similar Glass cleaner for the blocked nozzles. It does a great job. Just remember after to do a head cleaning with the printer itself.

Ioannis

[daniel5]

Yeah I've used that, it does do a good job.

[kellyseye]

I've been looking at thermal printing mechanisms, particularly those that use waxed-transfer-to-paper i.e. older fax machines. The resolution isn't great but quite sufficient for most general pcb work. The only stumbling block is that the element used to provide the thermal transfer process won't accept a true flat paper (or pcb) pass through due to the 'bump' housing the heaters (presumably).
If the thermal print head manufacturers could make one that kept these 'lumps' out of the way then we're set to go!

Unless someone has come across a particular model that already has a 'flat' route through?

[happytriger2000]

Here is my Epson S22/T22 mod.

http://www.rhinofablab.com/profiles/...ct-pcb-printer

[mark_s]

Nice build!

Thanks