Homebrew PCBs

Hello Pete:

I have not used the T-Tech machine, but I do have some experience with 
the LPKF milling machine.

With the software that that comes with that machine, you could specify 
the bit size, as well as the isolation width and overlap area.  With 
this, you could generate your isolation paths and actually see them with 
the software.  I am not sure if your software will let you do that, but 
I would look in that direction as well. 

Also, the LPKF machine had a small microscope that was used to set the 
depth of the V grove made by the  milling bit.  By adjusting the 
mechanical setting of the depth limiter.  I also had an 8 mil bit, but I 
was also able to  set this for 4 and 2 mils as well.  Of course, this 
would use the very tip of the mill bit, and they would not last long. I 
know that LPKF does sell a small  mill bit good to about 2 mils.  But I 
have always used the 8 mil bit.  I am not sure if this is how you are 
setting your depth.  Edmund Scientific would have this in stock, lined 
with 1 mil increments.  I would scibe a line outside of the work area 
and set the depth limit, sometimes at both ends of the PCB, to check to 
see if they were the same. 

I agree that making two passes will remove too much copper.  with only 
11 mil room, I would only make one pass, as that would only leave you 
with 1.5 mil of head room on either side.  Of course this is also 
starting to push the limits of the machine as well.  You may try having 
an overlap of say 60- 70% if the software will let you do that.  It 
seems like that would give you about a 10.4 - 11.2  mil isolation area 
with an 8 mil bit.  The LPKF machine  had a 60,000 RPM three phase 
motor, so it might have been able to cut a finer line and maybe a bit 
faster as well.  I have not done a 100 pin QFP yet, but it seems like 
did place 5 traces between 100 mil centers before.  Wish I would have 
kept a sample, but they all went to the customer.

Sad to say, milling a PCB is not like sending a document to the 
printer.  There is a bit of an art form involved as well.  At the 
geometries that you are trying to achieve, little things like board 
warpage may cause the the copper removal process to be uneven. I can 
remember setting the depth gage and coming back minutes later, and there 
was still copper left in areas on the other side of the board.  The LPKF 
software would let you highlight the areas of the board, and you could 
mill those again after adjusting the depth limiter.  I used to use the 
microscope to check the board as well before removing it from the 
machine, as it would sure be hard to try and line it up again to remove 
some errant copper.  I use to use an exacto knife to remove little 
pieces of copper that may have been left behind.

My company is looking at possible buying the new S62? LPKF machine.  I 
guess it changes the tools by itself and other nice things. I guess this 
would be harder to fudge, as you loose some of the control of being able 
to use the tool that you would like to.  I am not sure if the machines 
are worth the going price of 20K or so.  Thats about the price of the 
this new machine.  It seems like we paid that much for are old one a 
while back.  I am not sure if we have done $20K worth of boards on it 
yet.  As I remember the price of the mills, end mills, and drills were 
expensive.  Of course that was straight from LPKF.  You may be able to 
find them cheaper at a second source.

Ok, it's 2:30 in the moring and I am getting tired, that can only mean 
one thing....coffee.

Good luck.

Mark









Pete wrote:

>Hi, Group!  I teach at a university which recently bought a T-Tech 
>Quick Circuit 5000 for student project fabrication.  I'm trying to 
>mill a TQFP-100 footprint, which has 0.22 mm (nominal) wide pads 
>(approx. 8.7 mils) on 0.5 mm centers.  This means removing a 0.28 mm 
>(11 mil) wide space between pads, but IsoPRO tells me it needs to 
>use a 6-mil tool  The reason seems to be that it cuts twice between 
>pads, and I wonder if that's really necessary.
>
>I've made one attempt so far, using a T1 pointed tool which has a 60-
>degree point and which T-Tech says will mill 8-12 mil paths.  I set 
>it for a depth of about 7 mils, which should have made the cut width 
>about 5.4 mils on a 1-oz copper board, if I'm not mistaken.  I set 
>the depth by carefully lowering the tool until it just touches the 
>copper surface, then counting clicks.  The manual says each click 
>lowers the tool by 0.4 mil, but it seems as it may actualy lower it 
>a bit more than that.
>
>The results of my first cut were not acceptable.  Most of the pads 
>appeared to be narrower than 8.7 mils, and pretty ragged.  some 10-
>mill traces also looked pretty chewed up.  
>
>I'm going to try  the following:
>
>1)  run the spindle at top speed (24,000 rpm)
>
>2)  Slow down the rate of head movement
>
>3)  use 1/2 oz copper instead of 1 oz.
>
>4) use a smaller pointed tool, the T-4 which is supposed to be good 
>down to 4 mils.
>
>What is your opinion of these three steps?  Am I going in the right 
>direction?  Is there a way to keep the machine from cutting twice 
>between pads, and would that be a good idea?
>
>Thanks,
>
>Pete Goodmann
>Assistant Professor,
>Electrical and Computer Engineering Technology
>Indiana-Purdue University Fort Wayne (IPFW)
>
>
>
>
>
>
>
>
>
>
>  
>