Homebrew PCBs

Hi All,

This is a continuation of the discussion on the forum CNC_PCB-Design 
with Stefan on EDM & Etch Resist.

Stefan suggested tin as an etch resist, and I followed this up. Here 
are my tests to date.

Built a simple EDM power supply, 2 x 30V trafos with 70 Volt 6800uf 
reservoir cap, and dubious 2 x 1uF discharge caps for ablation. Used 
Don Lancaster's trick of light bulbs as ballast resistors, but only 
had 2 x 100 watt (240V), (yes I'm in the UK).

Obtained strong spark, but slow spark repetition rate.

Coated a section of copper clad pcb with tin/lead solder using 
soldering iron. Etched a line through the tin/lead solder by spark 
erosion by hand. Electrode was 1mm copper wire (all this came out of 
my junk box). Dielectric was diesel fuel (works well but smells).

Continued ablating the tin/lead until the copper showed. Copper is 
readily visible through the tin. Etched in Ferric Chloride solution. 
Result is complete etching of the EDM exposed copper.

This demonstrates to me that isolation milling by EDM is possible, 
and that pcb production by this method is a real possibility.

Spark eroding by hand was difficult, I could generate a repetitive 
spark for about 2-3 seconds, but when moving the electrode the spark 
ceased either due to over current (short circuit) or under current
(open circuit).

Next trial is to mount an electrode control on the cnc and try for an 
isolation mill on a pcb. I need to adjust the RC time constant for 
high spark rate for this application, since the objective of this 
excercise is to ablate the tin relatively quickly, and then etch the 
copper.

Will keep you posted.

Roger

Thanks for posting your results.
I didn't know ferric chloride can be used in conjunction with tin, i  
thought only sulphuric/peroxy etchants can be used for that....


your method is surely a possibility, if direct EDM of copper doesn't work.
However, it also is a large effort in comparision, and I'm not sure there  
aren't easier ways to apply resist for  etching.

ST


On Wed, 24 Aug 2005 00:05:31 +0200, ralucas4277 <ralucas4277@...>  
wrote:

> Hi All,
> This is a continuation of the discussion on the forum CNC_PCB-Design
> with Stefan on EDM & Etch Resist.
> Stefan suggested tin as an etch resist, and I followed this up. Here
> are my tests to date.
> Built a simple EDM power supply, 2 x 30V trafos with 70 Volt 6800uf
> reservoir cap, and dubious 2 x 1uF discharge caps for ablation. Used
> Don Lancaster's trick of light bulbs as ballast resistors, but only
> had 2 x 100 watt (240V), (yes I'm in the UK).
> Obtained strong spark, but slow spark repetition rate.
> Coated a section of copper clad pcb with tin/lead solder using
> soldering iron. Etched a line through the tin/lead solder by spark
> erosion by hand. Electrode was 1mm copper wire (all this came out of
> my junk box). Dielectric was diesel fuel (works well but smells).
> Continued ablating the tin/lead until the copper showed. Copper is
> readily visible through the tin. Etched in Ferric Chloride solution.
> Result is complete etching of the EDM exposed copper.
> This demonstrates to me that isolation milling by EDM is possible,
> and that pcb production by this method is a real possibility.
> Spark eroding by hand was difficult, I could generate a repetitive
> spark for about 2-3 seconds, but when moving the electrode the spark
> ceased either due to over current (short circuit) or under current
> (open circuit).
> Next trial is to mount an electrode control on the cnc and try for an
> isolation mill on a pcb. I need to adjust the RC time constant for
> high spark rate for this application, since the objective of this
> excercise is to ablate the tin relatively quickly, and then etch the
> copper.
> Will keep you posted.
> Roger

--- In Homebrew_PCBs@yahoogroups.com, "Stefan Trethan" 
<stefan_trethan@g...> wrote:
> Thanks for posting your results.
> I didn't know ferric chloride can be used in conjunction with tin, 
i  
> thought only sulphuric/peroxy etchants can be used for that....
> 
> 
> your method is surely a possibility, if direct EDM of copper 
doesn't work.
> However, it also is a large effort in comparision, and I'm not sure 
there  
> aren't easier ways to apply resist for  etching.
> 
> ST
> 

Stefan,

Following my last post, I did some more experimentation with the 
simple power supply, a single sided pcb and tap water dielectric. I 
produced single shot sparks and examined the craters in the copper 
layer under a microscope. Some had shiny bottoms and some had dull 
bottoms. Shining a strong thin light beam onto the rear of the pcb, 
it became obvious that the dark bottomed holes had in fact been 
blasted to the laminate, ie a clear hole in the copper. I also 
received Bob Langlois's book, and read all the posts I could find 
from Carl et al regarding his PCB-EDM. I now understand his approach, 
and the Garden of EDM PWM power supply, and the more I think about it 
the more I like this approach. The essence of success seems to be in 
the controllable power supply. Seeing the small holes convinced me. 

I am therefore going to dump the tin resist approach and concentrate 
on a machine similar to Carl's philosophy.

First is to build the Garden od EDM power supply and do some tests on 
hole size and quality with varying parameters.

If anyone has any suggestions on how to replace the IGBTs with 
MOSFETs in that design, I would appreciate any comments.

Roger

On Wed, 24 Aug 2005 14:16:51 +0200, ralucas4277 <ralucas4277@...>  
wrote:

> I am therefore going to dump the tin resist approach and concentrate
> on a machine similar to Carl's philosophy.
> First is to build the Garden od EDM power supply and do some tests on
> hole size and quality with varying parameters.
> If anyone has any suggestions on how to replace the IGBTs with
> MOSFETs in that design, I would appreciate any comments.
> Roger


I would not dump it, but first try the method that would result in a great  
simplification of the overall process.
If it doesn't work you can still go back.


I think you are right, a defined amount of energy, and thus a defined  
amount of abraded material might be the key to etching down to the  
laminate. What i see as a potential problem is how to figure out the ideal  
process to position the electrode. One approach could be to treat each dot  
as individual sinker EDM process, but the trouble is when you are finished  
there is solid material after the copper.
Ideally, one would be able to blast each dot with a single spark, but what  
if you were just advancing and the gap was so large you only achieved  
maybe half a proper spark. What do you do? stop and try more sparks?  
advance?

I think you might be able to use high power FETs in a very similar  
fashion, as they have similar gate characteristics.

ST

--- In Homebrew_PCBs@yahoogroups.com, "Stefan Trethan" 
<stefan_trethan@g...> wrote:

> 
> I would not dump it, but first try the method that would result in 
a great  
> simplification of the overall process.
> If it doesn't work you can still go back.
 
You are right, in a sense, one process is an extension of the other. 
Raw copper clad pcb to circuit board in one go looks very attractive.

 
> I think you are right, a defined amount of energy, and thus a 
defined  
> amount of abraded material might be the key to etching down to the  
> laminate. What i see as a potential problem is how to figure out 
the ideal  
> process to position the electrode. One approach could be to treat 
each dot  
> as individual sinker EDM process, but the trouble is when you are 
finished  
> there is solid material after the copper.
> Ideally, one would be able to blast each dot with a single spark, 
but what  
> if you were just advancing and the gap was so large you only 
achieved  
> maybe half a proper spark. What do you do? stop and try more 
sparks?  
> advance?

I understand that Carl does one spark per 0.001in step with his 
machine. I am assuming that the spark is only allowed to occur 
between steps, not during continuous motion. My option would be to 
use the Garden of EDM control to allow as many multiple sparks as 
were required at each step to ensure ablation, perhaps coupled with a 
rotating electrode. Of course, the step does not have to equal the 
electrode tip diameter, half tip diameter steps would ensure that the 
electrode always covered some copper, for example.

Perhaps Carl could give some intial parameters on his spark control 
settings, voltage level, gate duration etc that he has found to work, 
to save duplicating experimental work?

My first trial will be jog steps on the cnc with hand firing a 
selected number of shots in order to judge the quality of the mill.

> I think you might be able to use high power FETs in a very similar  
> fashion, as they have similar gate characteristics.

Yes, I can't see why not, after all we are not talking power station 
currents.

Roger

ralucas4277 wrote:

>Coated a section of copper clad pcb with tin/lead solder using 
>soldering iron. Etched a line through the tin/lead solder by spark 
>erosion by hand. Electrode was 1mm copper wire (all this came out of 
>my junk box). Dielectric was diesel fuel (works well but smells).
>
>Continued ablating the tin/lead until the copper showed. Copper is 
>readily visible through the tin. Etched in Ferric Chloride solution. 
>Result is complete etching of the EDM exposed copper.
>
>This demonstrates to me that isolation milling by EDM is possible, 
>and that pcb production by this method is a real possibility.
>
>  
>

  Following this idea, I'd be strongly inclined to forget coating and 
removing, reverse the polarity, and work on point plating the tin to the 
copper..

  Used to use a 12V supply, clip one side to a thin piece of steel, the 
other side onto a pencil lead from a mechanical pencil, and draw fairly 
permanent lines on the steel with the charged lead.  In fact may have to 
try this myself today, just skip the tin and draw lines on copper with 
pencil lead, should be a good enough resist and I'd imagine the copper 
would work ok in place of the steel..

  I should've thought to try this before, I used to do this all the time 
writing things on small scrap steel pieces just because it's fun to see 
the sparking and results..

Alan

--- In Homebrew_PCBs@yahoogroups.com, Alan King <alan@n...> wrote:

> 
>   Following this idea, I'd be strongly inclined to forget coating 
and 
> removing, reverse the polarity, and work on point plating the tin 
to the 
> copper..
> 
>   Used to use a 12V supply, clip one side to a thin piece of steel, 
the 
> other side onto a pencil lead from a mechanical pencil, and draw 
fairly 
> permanent lines on the steel with the charged lead.  In fact may 
have to 
> try this myself today, just skip the tin and draw lines on copper 
with 
> pencil lead, should be a good enough resist and I'd imagine the 
copper 
> would work ok in place of the steel..
> 
>   I should've thought to try this before, I used to do this all the 
time 
> writing things on small scrap steel pieces just because it's fun to 
see 
> the sparking and results..
> 
> Alan


Wouldn't point tinning take forever by electrolysis?

Would the graphite be dense enough to act as an etch resist? Would be 
interested in your test results.

Roger

ralucas4277 wrote:

>
>Wouldn't point tinning take forever by electrolysis?
>  
>
  How long does EDM take?  :)   Neither is really electrolysis.  It's 
more in the line of electrical discharge with vapor deposition when used 
the other way around...  Just works straight with dry materials clipped 
to a power source for steel and pencil lead..

>Would the graphite be dense enough to act as an etch resist? Would be 
>interested in your test results.
>
>  
>

  Results on steel are blacker than black, it looks like you've burned 
the steel.  Depending on many factors it could take several passes to 
write my name on it very well, but it was to fill in any gaps.  Anywhere 
you had the spark you had a transfer that was almost impossible to 
remove short of grinding it off, it looks like you've burned the steel 
to a cinder where you've sparked it..  I've got to find my lead though, 
all I have on hand is the waxy kind.  Works much better with the 
'graphite' feel leads, the kind that feel cold and gritty and leave 
black on your hands from handling them, mine came from old type 
mechanical pencils.  The waxy kind of leads don't work nearly as well, 
probably the binder turning to vapor soaks up the heat.  Plus it takes a 
good battery or supply, needs some amps behind it..

  I'll find my stuff and try it out, even if it doesn't work on copper 
or isn't very practical, it works great on steel and it would be neat 
just to see it work and write my name or similar in some steel to 
etch..  Haven't done it in a long time, this is something I played with 
a lot when I was 12-15 or so..  Sort of like making a solder bridge and 
plugging it into 120v, more a gimmick than something I ever thought of 
as useful.  Might just work though, but also likely to take a lot of 
other work to get it to do something like a board with it..

  As I recall it was quite fun to do and see instant results on steel 
from barely touching it with a wired lead, I highly recommend trying it 
out to everyone whether or not it works for any other purpose..

Alan

--- In Homebrew_PCBs@yahoogroups.com, Alan King <alan@n...> wrote:
> ralucas4277 wrote:
> 
> >
> >Wouldn't point tinning take forever by electrolysis?
> >  
> >
>   How long does EDM take?  :)   Neither is really electrolysis.  It's 
> more in the line of electrical discharge with vapor deposition when
used 
> the other way around...  Just works straight with dry materials clipped 
> to a power source for steel and pencil lead..

I don't suppose a high intensity, short range electrolysis process
to remove the copper would work, would it?  No, probably not, since
it would require the areas of Copper to be eroded to be connected
to a common plane.

> >Would the graphite be dense enough to act as an etch resist? Would be 
> >interested in your test results.
> >
> >  
> >
> 
>   Results on steel are blacker than black, it looks like you've burned 
> the steel.  Depending on many factors it could take several passes to 
> write my name on it very well, but it was to fill in any gaps. 
Anywhere 
> you had the spark you had a transfer that was almost impossible to 
> remove short of grinding it off, it looks like you've burned the steel 
> to a cinder where you've sparked it..  I've got to find my lead though, 
> all I have on hand is the waxy kind.  Works much better with the 
> 'graphite' feel leads, the kind that feel cold and gritty and leave 
> black on your hands from handling them, mine came from old type 
> mechanical pencils.  The waxy kind of leads don't work nearly as well, 
> probably the binder turning to vapor soaks up the heat.  Plus it
takes a 
> good battery or supply, needs some amps behind it..
> 
>   I'll find my stuff and try it out, even if it doesn't work on copper 
> or isn't very practical, it works great on steel and it would be neat 
> just to see it work and write my name or similar in some steel to 
> etch..  Haven't done it in a long time, this is something I played with 
> a lot when I was 12-15 or so..  Sort of like making a solder bridge and 
> plugging it into 120v, more a gimmick than something I ever thought of 
> as useful.  Might just work though, but also likely to take a lot of 
> other work to get it to do something like a board with it..
> 
>   As I recall it was quite fun to do and see instant results on steel 
> from barely touching it with a wired lead, I highly recommend trying it 
> out to everyone whether or not it works for any other purpose..

I have to wonder if the Carbon from the graphite lead may be 
alloying with the steel to form a new material.  After all, it 
sounds sort of like micro-power welding, or, actually, more like
micro-power cutting.

> Alan

Dave