0

Ask an Engineer 5/14/2011
Moderators: adafruit_support_bill, adafruit

Please be positive and constructive with your questions and comments.

Ask an Engineer 5/14/2011

by george_graves on Wed May 18, 2011 10:03 pm

- It's Jeri!
- Q&A

http://vimeo.com/23885161
New blog - www.digitalunderpants.com

george_graves
 
Posts: 342
Joined: Thu Feb 26, 2009 2:10 am
Location: Seattle, WA

Re: Ask an Engineer 5/14/2011

by abrooks on Wed May 18, 2011 11:47 pm

Hey, great interview with Jeri Ellsworth!

With regard to the discussion about masking PCBs for etching, has anyone ever tried modifying an inkjet printer to print a mask for exposure directly on the copper? That seems like it wouldn't have the focus/edge blurring issues that you can get with laserjet printed masks and would be fairly high resolution. I'd be interested to know if anyone has done this and if they were able to get decent results.

-A.

abrooks
 
Posts: 5
Joined: Tue May 17, 2011 11:23 am

Re: Ask an Engineer 5/14/2011

by george_graves on Mon May 23, 2011 5:59 am

IIRC there was a post on Hack-a-day of someone trying it. It sounded like the ink they were using goes on super thin - and requires a long time to fully dry. Even then you'll need to etch quickly.

It would be interesting if someone could find a replacement ink/head. I think that would make it better.
New blog - www.digitalunderpants.com

george_graves
 
Posts: 342
Joined: Thu Feb 26, 2009 2:10 am
Location: Seattle, WA

Re: Ask an Engineer 5/14/2011

by abrooks on Mon May 23, 2011 9:24 am

Thanks for the reply.

I wonder if there's some multi-stage approach that uses the properties of the ink to build a layer (such as dusted toner) on the wet ink (though that sounds imprecise). I had the thought of filling an ink cartridge with etchant and printing the etchant instead of a mask (i.e. no mask — you'd need to make sure the etchant is applied thinly enough to not run and would need to clean it thoroughly).

Of course, inkjet printing etchant has Other Problems. You'd definitely want to avoid thermal bubble printers. Canon and HP printers heat the ink to the vaporization point to create bubbles which burst, creating micro-dropplets. This seems like a bad thing to do to etchant. Piezo/ultrasonic printers such as Epson printers (All? Some? Are there other companies?) might work well though I'd be concerned about etchant vapors which would be both an inhalation and a fire hazard. You'd want a sufficient ventilation setup.

-A.

abrooks
 
Posts: 5
Joined: Tue May 17, 2011 11:23 am

Re: Ask an Engineer 5/14/2011

by schuldes on Mon May 30, 2011 9:48 am

Inkjet printing of echants will destroy the printheads itself.

We did sucessfully printed etch masks with inkjet directly onto copper as well as printing the soldermasks; both with converted epson-printers and a special water-based ink.
Attachments
InketPCB6_ArduinoMega_20x.jpg
inkjet etch mask - traces
InketPCB6_ArduinoMega_20x.jpg (348.44 KiB) Viewed 2809 times
PCBVergrösserung2.jpg
inkjet etchmask - traces
PCBVergrösserung2.jpg (334.42 KiB) Viewed 2809 times
InketPCB8_ArduinoMega_macro.jpg
inkjet etchmask - traces
InketPCB8_ArduinoMega_macro.jpg (341.5 KiB) Viewed 2809 times
schuldes
 
Posts: 3
Joined: Mon May 30, 2011 9:36 am

Re: Ask an Engineer 5/14/2011

by schuldes on Mon May 30, 2011 9:53 am

... and here some photos of pcbs with soldermasks. The PCB was only intended as a test, I know that vias etc. is missing.
Attachments
PCBkomlett2.jpg
inkjet pcb with soldermask
PCBkomlett2.jpg (897.31 KiB) Viewed 2809 times
PCBkomlett3.jpg
inkjet pcb with soldermask
PCBkomlett3.jpg (923.53 KiB) Viewed 2809 times
schuldes
 
Posts: 3
Joined: Mon May 30, 2011 9:36 am

Re: Ask an Engineer 5/14/2011

by abrooks on Mon May 30, 2011 3:51 pm

Wow, those results look very good. Did you run into any particular problems? Is your process detailed anywhere? I'm sure other people would be interested in your results. Thanks for sharing!

-A.

abrooks
 
Posts: 5
Joined: Tue May 17, 2011 11:23 am

Re: Ask an Engineer 5/14/2011

by george_graves on Tue May 31, 2011 5:14 pm

Can I ask how you make a solder mask? With what? And how did you mask out the pads?
New blog - www.digitalunderpants.com

george_graves
 
Posts: 342
Joined: Thu Feb 26, 2009 2:10 am
Location: Seattle, WA

Re: Ask an Engineer 5/14/2011

by schuldes on Thu Jun 02, 2011 8:23 am

We are converting Epson printers to flatbed version on a commercial basis. These printers are usually used for printing of dials, frontplates etc. As we use a specially developed ink which is extremly resitant, we decided to start to make some test to use it as a etch mask (for PCBs as well as for others things) and also as a soldermask.
Doing one-sided PCBs works very well and reliable so far; the smallest traces we have etched so far was 2mil wide, but you have to work very careful to get that.
I would say that we can get 5mil /5mil (measured by a microscope) quite reliable at the moment.

The first customer who bought a system started with it to do flexible PVBs, which works quite easily, and is now in development of doing pcbs out of conductive fabrics; with first results which look quite good.

Double-sided PCBs are still a problem with the quite cheap printers we build, as we a looking for fine tolerances within a few hundred millimeters between the two sides. Thats hard to get with cheap equipment, as the mechanics have to be very accurate for that (which raises the price).

The ink itself is a waterbased 2K-ink. Thats means, that first a preatretment has to be applied, which reacts with the ink so that you will get extremly precise results on non-absorbant materials, and you can cure the ink (from a starting point of 100 Celsius - for PCBs we usually use 130 Celsius for 15 minutes) to very high resitance. If cured totally, the ink is resitant even to acetone.
Its also quite resitant against heat - yesterday I made a test with hot air of 650 Celsius. That destroyed the FR2-pcb (delaminated in the material itself), but the ink still looked okay. Up to now, I couldn`t see any problems during soldering; I guess you would destroy the componets much faster than the ink by applying too much heat.

We keep on working to improve the process and also the ink itself; our goal is to get even more precise traces with epson-printheads.
schuldes
 
Posts: 3
Joined: Mon May 30, 2011 9:36 am

Re: Ask an Engineer 5/14/2011

by abrooks on Thu Jun 02, 2011 2:53 pm

Thanks for sharing those details!

You say that the converted printers are available on a commercial basis... Is there a site where they are sold?

It seems that a pair of square positioning cutouts (one for each side, probably on breakaway tabs for later removal) could be used for precise positioning (or circular holes with two, fixed edge guides, though I'd think there would be more play and less accuracy there), better allowing for dual side alignment.

This sounds like a very interesting project!

abrooks
 
Posts: 5
Joined: Tue May 17, 2011 11:23 am

Please be positive and constructive with your questions and comments.