The problem with scratch and etch is the same as milling. Users must have a toolpath calculation program to
generate the outline toolpaths. In practical terms, that's expensive for robust software that can do small trace
pitches.
Outline paths also have the problem of "islands" of varying sizes which can cause a lot of problems
for high speed boards. I don't think you could practically route a differential trace pair with isolations. There would
be an enormous amount of rubouts required to clear all the copper between the traces but keep the traces at the
proper spacing and width for the correct impedence. One tiny hair of copper running along between your traces
can really mess up the signalling. I wouldn't even try an LVDS bus with isolation.
Sure, Eagle can do outline paths but not everybody uses Eagle. T-Tech's Isolator Pro does excellent toolpaths
but is expensive!. KellyCam generates pretty good toolpaths but sucks as a CNC controller. I always used Isolator
Pro for toolpaths when I did mechanical etching.
I played around with "scratch and etch" too for a while. I also plotted directly to copper using modified drafting pens
filled with acrylic floor polish. The smallest pen I have found locally is a .18 mm tip technical pen. That produces a trace
of .0070866 wide. So it's a 7 mil trace. I successfully plotted and etched some 48 pin TSOP to 48 pin DIP adapter
boards so I could read some flash roms on my programmer. But it was SLOW! Also filled areas like pads
had concentric circles in them where the carbide pen tip scratched into the just layed down path of resist.
However, using Inkjets produces solid fills with no voids. No special software is needed either. All you need is
something that can print your gerber files to a standard printer.
Modern inkjet printers are fast and are accurate enough to do some very small trace widths and spacings.
You could also lay down solder mask and the component identifiers using the same modified printer.
While I agree that the scratch and etch idea is great, and has some benefits over direct resist application, the biggest
downside I see is the problem of the toolpath generation for the average user. Also, the tool must be tightly controlled
to produce the necessary isolation path widths. For me the perfect width would be a 5 mil tool. But will a tool that is
.005 wide cut a clean swipe through the bluing without splintered edges? If your tool has a very sharp tip, then you'll
need multiple passes to get a 5 to 10 mill isolation path. I never got a clean enough path through the resist. "furry"
or jagged edges on traces also cause lots of noise problems on high speed circuits.
The market for these products are not just hobbiests. There are a lot of us guys out there that make their living in the
electronics industry that would LOVE to have something to quickly make a board without having to send it out!
Lots of corporations would buy them for the EE labs!
I just wish someone would run with it and work out all the bugs so I could just buy one. I want to make boards not
tools to make boards.
later,
Larry E.
----- Original Message -----
From: crankorgan
To: Homebrew_PCBs@yahoogroups.com
Sent: Saturday, September 14, 2002 6:41 AM
Subject: [Homebrew_PCBs] Re: Inkjet printing PC boards
Larry,
I sell plans for a PCBmill (Brute) and I just finished a
machine that will mill or draw simple boards. (Morph)
Months ago I brought up the idea of Scratch and Etch. This
idea is not dead. First I developed a super simple CNC plotter. Now
I have a machine strong enough to drag a scribe through the blueing.
Several people tried Scratch and Etch With good results. Coating the
whole board and then scribing and isolation between pads for etching
seems easier.
Crankorgan John
--- In Homebrew_PCBs@y..., "Larry Edington" <ledington@a...> wrote:
> This is a copy of a reply I just posted in sci.electronics.design.
I thought I would
> post it here as well for you folks to pursue if you are interested
in it. Maybe
> Crankorgan John could go work up some plans for it. It's promising
technology
> but I like working with boards more than belts and gears.
>
>
> > >Has anyone every tried putting Etch resist Ink in a BubbleJet
or InkJet
> > >printer and printing a PCB design directy onto the copper side
of a PCB.
>
> As a matter of fact I have. I thought about turning this into a
commercial product, but
> I just don't have the time. Too many other projects to do. I even
went so far as to
> turn an inkjet into a flatbed plotter. The problem there is
converting the stepper motor driven
> motion of the platten roller into linear motion with the stepper
using belts or gears or
> electronic gear ratio reduction for the steps. All three methods
work but I like the
> belt and gears method best.
>
> I tried modifying printers to feed a board through as is, but that
was just too much trouble.
> Some would feed thin stock just fine but others wouldn't. The flat
bed approach solves that.
> Plus the flat bed approach with some simple guide pins lets you do
double sided boards with
> accurate side to side registration.
>
> For ink, I found two things that work well standing up to the
etchant.
>
> Acrylic Floor Polish.
> Shellac
>
> Floor polish ( sometimes called wax ) ( I used the Future brand )
works well and won't gum
> up the printheads as easily as Shellac does.
>
> Floor polish is removed with Ammonia.
>
> Shellac with Denatured Alcohol.
>
> I used three different types of printers. A Lexmark where the head
is built into the cartridge,
> an HP where the printhead is separate from the cartridge but still
uses thermal "bubble" jet
> technology and an Epson which uses piezoelectric technology.
>
> The winner was the Lexmark. Easy cartridge to refill and cheap
enough to replace.
>
> The problem in all this is that to use straight "ink" that is very
water ( and then etchant ) proof
> is, it will dry on your printhead orifices. The solution to that is
a valve. One path to ink, one path
> to a flushing fluid. Before you print the board, run a cycle
of "ink" to clear out all your flush fluid.
> Once you have finished printing, run a cycle of flushing fluid to
purge all your "ink".
>
> Ammonia diluted with water works great on the acrylic polish ink.
You can even put a little regular
> inkjet ink into it to color it so it's easier to see on the PC
board blank.
>
> Once you get the "ink" on the board and have etched it, you can
clean it off with just a quick spray of
> Windex or other ammonia based cleaner.
>
> I have no idea as to the benefits or problems with acrylics as
conformal coatings but I plan to try it
> some day on a finished board. Just dip it in Future polish and see
what happens!
>
> Another thought I had was a PVA ( poly vinyl alcohol ) based ink
but the PVA I had on hand didn't
> thin out well to run in an inkjet. The acrylic ink worked well
enough so I didn't really pursue PVA much.
>
> Now go have fun......
>
> later,
> Larry Edington.
>
>
> [Non-text portions of this message have been removed]
Yahoo! Groups Sponsor
ADVERTISEMENT
Be sure to visit the group home and check for new Bookmarks and files:
http://groups.yahoo.com/group/Homebrew_PCBs
To unsubscribe from this group, send an email to:
Homebrew_PCBs-unsubscribe@yahoogroups.com
Your use of Yahoo! Groups is subject to the Yahoo! Terms of Service.
[Non-text portions of this message have been removed]