Homebrew PCBs

Hi all.

I have been looking into making 4 layer boards. I have most stages 
sorted but the big stumbling block is the bonding of the layers 
together. It seems a special press is required that has enough pressure, 
exhausts the gases away and generates the correct temperatures. I am 
surprised its so difficult. I have been told second hand presses cost 
\ufffd4000 to \ufffd5000. This is totally out of budget for me. Has anyone any 
experience/thoughts/comments on bonding boards together?

Thanks

Trev

I use double-sided tape when I do four-layer boards.  They don't end
up as stiff as "real" boards, but it's good enough for me.

Trevor White wrote:

> 
> I have been looking into making 4 layer boards.  Has anyone any 
> experience/thoughts/comments on bonding boards together?
> 

It depends, are you using true prepreg laminates or common fr4?

In this list some people accomplished it or at least in a homebrew way.
My first thoughts go to Markus Zingg who used common fr4 glued with 5min
epoxy pressed with vises with PTH
http://tech.groups.yahoo.com/group/Homebrew_PCBs/message/3266

And Dj Deloire who uses innerlayers for gnd and vcc and uses a bit of
solder in bigger holes where gnd or vcc is needed in the 2 layer
circuit. Using thin fr4 bonded with thin double sided tape
http://tech.groups.yahoo.com/group/Homebrew_PCBs/message/19442

Dear Trevor White 

  with regards to your question about 4 layer boards (which are easily 4 times more expensive to have made,in my experience, than double sided ones....), are you a hobbyist or professional? Because if you are a professional, the price (almost !) does/should not matter.....

  If you are Hobbyist, then you should give some thought to redesigning the board for double layer, even if its a bit bigger because of that.... Fault finding on 4 layer PCBs is not for the faint hearted either and repairs can be a nightmare.....

  If you are developing something privately, for maybe later selling, may I still suggest that until the design is working perfectly, that you stay with double layer, get some experience with that and when you really go for "IT", do the 4 layer design. You may get some new problems but YOU WILL KNOW EXACTLY WHY!!! Which is like money in the bank!!!

  I would be most interested in learning why (if) you feel that a 4 layer is the only way to go in your particular case......

  (I have to admit that my personal (not business I hasten to add, we used multilayer pcbs all the time there) PCB level is far less complicated than yours, I design using my PCB software at first double sided (my old software makes a mess/gives up when telling it to design single sided!) and then try and redesign by hand to fit a single sided PCB - it appears that I like wire links a lot!!! Or better said, they do not bother me at all......I can usually manage to get probably 90% of my designs (thats probably between 4 and 8 boards a year!!) onto single sided - not that I am anticipating you changing a 4 level design to ever fit single sided PCB, that was NOT my meaning!! Double at least I would still guess.....)

  Stay well and regards from

  Andy

[Non-text portions of this message have been removed]

Thanks for the responses. My reasons for asking are a mix really. I am 
an engineer by trade but also I am just really being curious. I do not 
have a specific need for a 4 layer board right now. I have laid out 4 
layer boards in the past though and for some reason always fancied 
making them myself. I have never tried yet.

I like to make things smaller and smaller so 4 layer boards enable that. 
For one company I worked at a few years ago I did the cad work for a 
board that was 12mm x 10mm and needed to be 4 layers because of all the 
tracks.

Last year I quite fancied making a board based on the AVR32 and this 
looked like it would have to be 4 layers at least. Getting a couple of 
prototypes made would cost both time and money and well I did not have a 
real need for the board. I just wanted to make one. So that project got 
shelved.

So I think I would probably use the skills in my work occassionally but 
i am not sure if I was truely doing it for professional reasons it would 
make financial sense. I certainly do not have cash to invest to prove 
that. I think I just want to make more and more complex, small boards.

Trev






Andrew Mathison wrote:

>
> Dear Trevor White
>
> with regards to your question about 4 layer boards (which are easily 4 
> times more expensive to have made,in my experience, than double sided 
> ones....), are you a hobbyist or professional? Because if you are a 
> professional, the price (almost !) does/should not matter.....
>
> If you are Hobbyist, then you should give some thought to redesigning 
> the board for double layer, even if its a bit bigger because of 
> that.... Fault finding on 4 layer PCBs is not for the faint hearted 
> either and repairs can be a nightmare.....
>
> If you are developing something privately, for maybe later selling, 
> may I still suggest that until the design is working perfectly, that 
> you stay with double layer, get some experience with that and when you 
> really go for "IT", do the 4 layer design. You may get some new 
> problems but YOU WILL KNOW EXACTLY WHY!!! Which is like money in the 
> bank!!!
>
> I would be most interested in learning why (if) you feel that a 4 
> layer is the only way to go in your particular case......
>
> (I have to admit that my personal (not business I hasten to add, we 
> used multilayer pcbs all the time there) PCB level is far less 
> complicated than yours, I design using my PCB software at first double 
> sided (my old software makes a mess/gives up when telling it to design 
> single sided!) and then try and redesign by hand to fit a single sided 
> PCB - it appears that I like wire links a lot!!! Or better said, they 
> do not bother me at all......I can usually manage to get probably 90% 
> of my designs (thats probably between 4 and 8 boards a year!!) onto 
> single sided - not that I am anticipating you changing a 4 level 
> design to ever fit single sided PCB, that was NOT my meaning!! Double 
> at least I would still guess.....)
>
> Stay well and regards from
>
> Andy
>
> [Non-text portions of this message have been removed]
>
>

--- In Homebrew_PCBs@yahoogroups.com, Trevor White <trevor.white100@...> wrote:
>
> Hi all.
> 
> I have been looking into making 4 layer boards. I have most stages 
> sorted but the big stumbling block is the bonding of the layers 
> together. It seems a special press is required that has enough pressure, 
> exhausts the gases away and generates the correct temperatures. I am 
> surprised its so difficult. I have been told second hand presses cost 
> £4000 to £5000. This is totally out of budget for me. Has anyone any 
> experience/thoughts/comments on bonding boards together?
> 
> Thanks
> 
> Trev
>

I've only done it once and I glued 2 double sided boards with epoxy, but added a sheet of fine fibreglass tissue between them also

worked, but was fiddly

alignment was assured by drilling outer holes, soldering a component lead in and snipping off (leaving about 0.5mm protruding)

glued up between 2 steel blocks in the vice

On Mon, 14 Dec 2009 13:25:21 +0000, you wrote:

>Thanks for the responses. My reasons for asking are a mix really. I am 
>an engineer by trade but also I am just really being curious. I do not 
>have a specific need for a 4 layer board right now. I have laid out 4 
>layer boards in the past though and for some reason always fancied 
>making them myself. I have never tried yet.
>
>I like to make things smaller and smaller so 4 layer boards enable that. 
>For one company I worked at a few years ago I did the cad work for a 
>board that was 12mm x 10mm and needed to be 4 layers because of all the 
>tracks.
>
>Last year I quite fancied making a board based on the AVR32 and this 
>looked like it would have to be 4 layers at least. Getting a couple of 
>prototypes made would cost both time and money and well I did not have a 
>real need for the board. I just wanted to make one. So that project got 
>shelved.

I have a board design with a MEGA2560 and an epson graphics driver
chip, a 512K*8 RAM and assorted other parts.  Does well on a double
sided board.  4 layer would be nicer, but not needed by the design.  A
bit of careful layout has produced only one jumper on the board.

Harvey

>
>So I think I would probably use the skills in my work occassionally but 
>i am not sure if I was truely doing it for professional reasons it would 
>make financial sense. I certainly do not have cash to invest to prove 
>that. I think I just want to make more and more complex, small boards.
>
>Trev
>
>
>
>
>
>
>Andrew Mathison wrote:
>>
>> Dear Trevor White
>>
>> with regards to your question about 4 layer boards (which are easily 4 
>> times more expensive to have made,in my experience, than double sided 
>> ones....), are you a hobbyist or professional? Because if you are a 
>> professional, the price (almost !) does/should not matter.....
>>
>> If you are Hobbyist, then you should give some thought to redesigning 
>> the board for double layer, even if its a bit bigger because of 
>> that.... Fault finding on 4 layer PCBs is not for the faint hearted 
>> either and repairs can be a nightmare.....
>>
>> If you are developing something privately, for maybe later selling, 
>> may I still suggest that until the design is working perfectly, that 
>> you stay with double layer, get some experience with that and when you 
>> really go for "IT", do the 4 layer design. You may get some new 
>> problems but YOU WILL KNOW EXACTLY WHY!!! Which is like money in the 
>> bank!!!
>>
>> I would be most interested in learning why (if) you feel that a 4 
>> layer is the only way to go in your particular case......
>>
>> (I have to admit that my personal (not business I hasten to add, we 
>> used multilayer pcbs all the time there) PCB level is far less 
>> complicated than yours, I design using my PCB software at first double 
>> sided (my old software makes a mess/gives up when telling it to design 
>> single sided!) and then try and redesign by hand to fit a single sided 
>> PCB - it appears that I like wire links a lot!!! Or better said, they 
>> do not bother me at all......I can usually manage to get probably 90% 
>> of my designs (thats probably between 4 and 8 boards a year!!) onto 
>> single sided - not that I am anticipating you changing a 4 level 
>> design to ever fit single sided PCB, that was NOT my meaning!! Double 
>> at least I would still guess.....)
>>
>> Stay well and regards from
>>
>> Andy
>>
>> [Non-text portions of this message have been removed]
>>
>> 
>
>
>
>------------------------------------
>
>Be sure to visit the group home and check for new Links, Files, and Photos:
>http://groups.yahoo.com/group/Homebrew_PCBsYahoo! Groups Links
>
>
>

Well the epoxy idea seems to have legs. I am going to order some and 
give it a try. I was feeling quite down last week when I found out the 
prices of a press and being told that you really need one to be able to 
do 4 layers. Reading over the old postings it really does sound like it 
can be done at home.

Brilliant, brilliant group. Thanks for the help guys.

Trev

On Mon, 14 Dec 2009 16:31:14 +0000, you wrote:

>Well the epoxy idea seems to have legs. I am going to order some and 
>give it a try. I was feeling quite down last week when I found out the 
>prices of a press and being told that you really need one to be able to 
>do 4 layers. Reading over the old postings it really does sound like it 
>can be done at home.

I haven't tried 4 layers, I've done two layer boards.  It's often
quite nice to make each layer independently, then match them.  Lots
less stress when one half (always the second one) of a two layer board
turns out badly.

Some small suggestions:

1) don't use 5 minute epoxy, use 20 minute epoxy.  You want the
additional working time, and the overall setup time is the same.

2) try a dispenser first, see how well it works.  You'll probably want
more depending on how many boards you use.

3) I use both guide holes and one or two drilled vias.  I try to match
them all.  Layout and trying to make the boards exactly match on the
printer (left right, for me, always at the top) seems to be critical. 
4) find yourself two pieces of cheap marble tile, and use them for a
press.  Use waxed paper or the equivalent so that the epoxy does not
glue the marble together.

5) I've tried superglue, but I also manage to use the vias and such to
also hold the boards together.  Epoxy seems to be better.

6) be absolutely sure that the epoxy is not contaminated by scraps of
already cured epoxy.  It will cure in very little time, giving you
very little working time.

Harvey

>
>Brilliant, brilliant group. Thanks for the help guys.
>
>Trev
>
>
>------------------------------------
>
>Be sure to visit the group home and check for new Links, Files, and Photos:
>http://groups.yahoo.com/group/Homebrew_PCBsYahoo! Groups Links
>
>
>

I've 3M spray adhesive, the I use several registration holes with 
needles poked through.
As the sides come together you align them so the needles are are 
perpendicular ( as best can
be done ) before allowing the sides to make contact. -Lee Studley


__________ Information from ESET Smart Security, version of virus signature database 4686 (20091214) __________

The message was checked by ESET Smart Security.

http://www.eset.com

Making four layer boards sounds interesting but what do you do bout the
plated through holes. I find even soldering double sided boards with no
plated holes a bit of a pain.

 

Sage

 



[Non-text portions of this message have been removed]

I have just set up a plating system so i have that covered.

Trev

----- Original message -----
> Making four layer boards sounds interesting but what do you do bout the
> plated through holes. I find even soldering double sided boards with no
> plated holes a bit of a pain.
>
> 
>
> Sage
>
> 
>
>
>
> [Non-text portions of this message have been removed]
>



[Non-text portions of this message have been removed]

Once everything is lined up, put it in a strong trash bag and suck out the air with a vacuum.  This may give you a few PSI of pressure, which, on a 12 inch tile, is like piling a couple of hundred pounds of books on top, plus it tends to suck the bubbles out of the material. 

--- In Homebrew_PCBs@yahoogroups.com, Harvey White <madyn@...> wrote:

> 4) find yourself two pieces of cheap marble tile, and use them for a
> press.  Use waxed paper or the equivalent so that the epoxy does not
> glue the marble together.
>

"Dave Sage" <davesage12@...> writes:
> Making four layer boards sounds interesting but what do you do bout
> the plated through holes. I find even soldering double sided boards
> with no plated holes a bit of a pain.

It's a pain, but this is what I do...

Your stackup is: 32 mil DS clad for power/ground plane "core".  8 mil
SS clad for the two outer signal layers.

Inner layers are for power and ground ONLY.  Wherever you have a
through-via, no problem, solder as usual.  For power/gnd vias, you do
one of two things:

* To connect to the OTHER side's plane, just drill out the whole pad
  on the far side (you aren't using it anyway), exposing the inner
  plane, solder as usual.

* To connect to the SAME side plane, you do NOT drill (or even etch a
  hole) on the plane, just leave it solid.  Put an oversized pad on
  the outer layer and drill it big enough to get a fine-point
  soldering iron in there, and solder a wire from the outer layer to
  the plane.

* When both outer sides connect, you do the oversized copper and hole
  on the plane's side, then re-drill a smaller hole through the board
  (off-center helps).  When you solder the wire through, you should be
  able to get the tip inside the hole to get the inner layer.

Yes, this means you drill a bunch of holes in the outer clad before
adhering the layers together.  I put the sticky tape on the outer
layers before drilling them, so that I drill through the tape also,
then peel the second layer off and stick it to the core.

Obviously, I avoid four-layer boards when I can :-)

Trevor White wrote:
>  
> 
> Hi all.
> 
> I have been looking into making 4 layer boards. I have most stages
> sorted but the big stumbling block is the bonding of the layers
> together. It seems a special press is required that has enough pressure,
> exhausts the gases away and generates the correct temperatures. I am
> surprised its so difficult. I have been told second hand presses cost
> \ufffd4000 to \ufffd5000. This is totally out of budget for me. Has anyone any
> experience/thoughts /comments on bonding boards together?
> 
> Thanks
> 
> Trev

Hi, I've done a number of 4 layers, its very laborious (about 5 hours 
was my record time to make a 4 layer from printing the artwork to the 
final solder mask bake)
  As for lamination, you can use room temperature curing liquid epoxy 
and a layer of fibre glass cloth for separation of inner layers. My 
method cheats, and uses two DS 0.8mm boards bonded together, so inner 
layers are about 0.2mm apart.

4 layer steps (over simplified)
------------------
* cut two pieces of 0.8mm double sided board
* protect outer layers from etchant with self adhesive book covering.
* drill 4 or more registration holes
* print and etch inner layers (clean, photoresist, expose, develop & etch)
* place one board on aluminium plate
* place sheet of class cloth,
* pour epoxy
* place 2nd PCB, align with pins on holes
* place 2nd aluminium plate
* place some bricks
* place on top of electric hot plate and heat PCB to 90C to 110C, 30 
minutes.
* cool, trim sides, remove self adhesive film.
* process like an ordinary plated through PCB.



To improve copper/epoxy bond you can make it black oxide by wiping it 
with strong bleach solution with 10% NaOH for a few seconds. this 
improves bond a lot.

Not that I have the remotest plan to make a 4-layer board, but with the
epoxy sandwich method, how do you get any electrical connection between
either side of the fibre glass cloth?

PG

On Wed, Dec 16, 2009 at 3:55 PM, Adam Seychell <a_seychell@...>wrote:

>
>
> Trevor White wrote:
> >
> >
> > Hi all.
> >
> > I have been looking into making 4 layer boards. I have most stages
> > sorted but the big stumbling block is the bonding of the layers
> > together. It seems a special press is required that has enough pressure,
> > exhausts the gases away and generates the correct temperatures. I am
> > surprised its so difficult. I have been told second hand presses cost
> > £4000 to £5000. This is totally out of budget for me. Has anyone any
> > experience/thoughts /comments on bonding boards together?
> >
> > Thanks
> >
> > Trev
>
> Hi, I've done a number of 4 layers, its very laborious (about 5 hours
> was my record time to make a 4 layer from printing the artwork to the
> final solder mask bake)
> As for lamination, you can use room temperature curing liquid epoxy
> and a layer of fibre glass cloth for separation of inner layers. My
> method cheats, and uses two DS 0.8mm boards bonded together, so inner
> layers are about 0.2mm apart.
>
> 4 layer steps (over simplified)
> ------------------
> * cut two pieces of 0.8mm double sided board
> * protect outer layers from etchant with self adhesive book covering.
> * drill 4 or more registration holes
> * print and etch inner layers (clean, photoresist, expose, develop & etch)
> * place one board on aluminium plate
> * place sheet of class cloth,
> * pour epoxy
> * place 2nd PCB, align with pins on holes
> * place 2nd aluminium plate
> * place some bricks
> * place on top of electric hot plate and heat PCB to 90C to 110C, 30
> minutes.
> * cool, trim sides, remove self adhesive film.
> * process like an ordinary plated through PCB.
>
> To improve copper/epoxy bond you can make it black oxide by wiping it
> with strong bleach solution with 10% NaOH for a few seconds. this
> improves bond a lot.
>
>  
>


[Non-text portions of this message have been removed]

The big boys use some kind of through-hole plating... solder should work

-----Original Message-----
Date: Tuesday, December 15, 2009 11:26 pm
To: Homebrew_PCBs@yahoogroups.com
Subject: Re: [Homebrew_PCBs] 4 Layer boards
From: "Piers Goodhew" <piers@...>

Not that I have the remotest plan to make a 4-layer board, but with the
epoxy sandwich method, how do you get any electrical connection between
either side of the fibre glass cloth?

PG

On Wed, Dec 16, 2009 at 3:55 PM, Adam Seychell <a_seychell@...>wrote:

>
>
> Trevor White wrote:
> >
> >
> > Hi all.
> >
> > I have been looking into making 4 layer boards. I have most stages
> > sorted but the big stumbling block is the bonding of the layers
> > together. It seems a special press is required that has enough pressure,
> > exhausts the gases away and generates the correct temperatures. I am
> > surprised its so difficult. I have been told second hand presses cost
> > \ufffd4000 to \ufffd5000. This is totally out of budget for me. Has anyone any
> > experience/thoughts /comments on bonding boards together?
> >
> > Thanks
> >
> > Trev
>
> Hi, I've done a number of 4 layers, its very laborious (about 5 hours
> was my record time to make a 4 layer from printing the artwork to the
> final solder mask bake)
> As for lamination, you can use room temperature curing liquid epoxy
> and a layer of fibre glass cloth for separation of inner layers. My
> method cheats, and uses two DS 0.8mm boards bonded together, so inner
> layers are about 0.2mm apart.
>
> 4 layer steps (over simplified)
> ------------------
> * cut two pieces of 0.8mm double sided board
> * protect outer layers from etchant with self adhesive book covering.
> * drill 4 or more registration holes
> * print and etch inner layers (clean, photoresist, expose, develop & etch)
> * place one board on aluminium plate
> * place sheet of class cloth,
> * pour epoxy
> * place 2nd PCB, align with pins on holes
> * place 2nd aluminium plate
> * place some bricks
> * place on top of electric hot plate and heat PCB to 90C to 110C, 30
> minutes.
> * cool, trim sides, remove self adhesive film.
> * process like an ordinary plated through PCB.
>
> To improve copper/epoxy bond you can make it black oxide by wiping it
> with strong bleach solution with 10% NaOH for a few seconds. this
> improves bond a lot.
>
>  
>


[Non-text portions of this message have been removed]

On Wed, 16 Dec 2009 09:46:44 CST, you wrote:

>The big boys use some kind of through-hole plating... solder should work

Just solder making the connection between layers is not a good idea.
If the connection is from a trace to a pit (what would be a blind via,
not all the way through), I'd use a wire and solder it.  

If the connection is through a board (full via) connecting top and
bottom layers, solder is not the way to go.  There's no way to
guarantee that the solder flows down into the hole and makes the
connection.  

Wires can be a very good connection method, if there are no plated
through holes.  The solder that helps hold them in place can wick down
the  wire, but is not the prime connection.

Harvey

>-----Original Message-----
>Date: Tuesday, December 15, 2009 11:26 pm
>To: Homebrew_PCBs@yahoogroups.com
>Subject: Re: [Homebrew_PCBs] 4 Layer boards
>From: "Piers Goodhew" <piers@...>
>
>Not that I have the remotest plan to make a 4-layer board, but with the
>epoxy sandwich method, how do you get any electrical connection between
>either side of the fibre glass cloth?
>
>PG
>

Harvey White <madyn@...> writes:
> Just solder making the connection between layers is not a good idea.
> If the connection is from a trace to a pit (what would be a blind via,
> not all the way through), I'd use a wire and solder it.  
> 
> If the connection is through a board (full via) connecting top and
> bottom layers, solder is not the way to go.  There's no way to
> guarantee that the solder flows down into the hole and makes the
> connection.  

Having tried these, I can pretty much guarantee that the solder will
*not* go down the hole, or into the pit, or whatnot.  Way too much
surface tension.

> The solder that helps hold them in place can wick down the wire, but
> is not the prime connection.

In my experience, it doesn't wick down the wire into the hole.  You've
seen it happen with *plated* holes, but without plating, it just stays
at the surface.

On 16 Dec 2009 17:18:49 -0500, you wrote:

>
>Harvey White <madyn@...> writes:
>> Just solder making the connection between layers is not a good idea.
>> If the connection is from a trace to a pit (what would be a blind via,
>> not all the way through), I'd use a wire and solder it.  
>> 
>> If the connection is through a board (full via) connecting top and
>> bottom layers, solder is not the way to go.  There's no way to
>> guarantee that the solder flows down into the hole and makes the
>> connection.  
>
>Having tried these, I can pretty much guarantee that the solder will
>*not* go down the hole, or into the pit, or whatnot.  Way too much
>surface tension.

Yep, not a good idea.  Doesn't work.
>
>> The solder that helps hold them in place can wick down the wire, but
>> is not the prime connection.
>
>In my experience, it doesn't wick down the wire into the hole.  You've
>seen it happen with *plated* holes, but without plating, it just stays
>at the surface.
>

No plated through holes, you have a problem making connections.  I've
used eyelets, but those get filled (and perhaps overfilled) with
solder.  Main thing is to make sure that the eyelet has sufficient pad
around it to make a good connection.  Ditto with feedthrough wires.
You need to lace the wire through the board, then solder.  Otherwise,
you solder the wire and it drops out on the second side.  Awkward at
best.

Harvey

>
>------------------------------------
>
>Be sure to visit the group home and check for new Links, Files, and Photos:
>http://groups.yahoo.com/group/Homebrew_PCBsYahoo! Groups Links
>
>
>

Several (at least three) strands of very thin wire will wick much better than a single strand (which barely wicks at all.)  Of course, one needs to make the through holes large enough to handle several strands.

Donald.

----- Original Message -----
From: "DJ Delorie" <dj@...>
To: "Homebrew PCBs" <Homebrew_PCBs@yahoogroups.com>
Sent: Wednesday, December 16, 2009 5:18:49 PM GMT -05:00 US/Canada Eastern
Subject: Re: [Homebrew_PCBs] 4 Layer boards


Harvey White <madyn@...> writes:
> Just solder making the connection between layers is not a good idea.
> If the connection is from a trace to a pit (what would be a blind via,
> not all the way through), I'd use a wire and solder it.  
> 
> If the connection is through a board (full via) connecting top and
> bottom layers, solder is not the way to go.  There's no way to
> guarantee that the solder flows down into the hole and makes the
> connection.  

Having tried these, I can pretty much guarantee that the solder will
*not* go down the hole, or into the pit, or whatnot.  Way too much
surface tension.

> The solder that helps hold them in place can wick down the wire, but
> is not the prime connection.

In my experience, it doesn't wick down the wire into the hole.  You've
seen it happen with *plated* holes, but without plating, it just stays
at the surface.

Adam Seychell wrote:

> 
> Hi, I've done a number of 4 layers, its very laborious (about 5 hours 
(...)
> * process like an ordinary plated through PCB.

>From your messages describing your PTH chemistry i did not identify a
desmear or etchback bath. Are you using a etchback for multilayers or is
possible to connect tracks in the inner layers without etchback the
holes in a multilayer board? 


One idea to print all the layers at once using thin fr4 in a unmodified
laser or inkjet printer stucks on not having less steps and being
impossible due to etchback. Imagine this, printing all layers in a
20x30cm 0.3mm thick board. All layers in one sheet with registration
holes, plus 2 drill marks on sheet opposite corners to align the print
in a cnc drilling them first and then the registration holes in each
part. Cut and etch the inner layers only and then glue the thing up.
Then procede like one double layer but already printed on out layers. 
It requires negative print outer layers and pattern  plating with
metallic resist so it's nothing easier after all. And fails on that
etchback solutions will remove any useful ink printed on outer layers. 
It made me search for acidic desmear/etchback things. For just resin
smear removal only, concentrated sulfuric acid room temp its said to do
the job, but not etchback. There is also a version of permanganate
desmear with sulfuric acid at room temp instead of caustic soda at 90�C
but is not safe. And EG or PEG with sulfuric acid but slower than with
MEA. 
So what do you use as desmear/etchback?

Simao Cardoso wrote:
>  
> 
> Adam Seychell wrote:
> 
>  >
>  > Hi, I've done a number of 4 layers, its very laborious (about 5 hours
> (...)
>  > * process like an ordinary plated through PCB.
> 
>  >From your messages describing your PTH chemistry i did not identify a
> desmear or etchback bath. Are you using a etchback for multilayers or is
> possible to connect tracks in the inner layers without etchback the
> holes in a multilayer board?

I have no desmear removal step. Never had an inner layer electrical 
problem yet. I think because professionals use 100k+ RPM drill speeds 
there is much more probability of epoxy smear. My home made manual drill 
press runs at 15k RPM max.
98% sulfuric is an excellent  epoxy etch. In hot H2SO4 for few hours 
your can remove all epoxy and just have nothing but bare copper traces.

> 
> One idea to print all the layers at once using thin fr4 in a unmodified
> laser or inkjet printer stucks on not having less steps and being
> impossible due to etchback. Imagine this, printing all layers in a
> 20x30cm 0.3mm thick board. All layers in one sheet with registration
> holes, plus 2 drill marks on sheet opposite corners to align the print
> in a cnc drilling them first and then the registration holes in each
> part. Cut and etch the inner layers only and then glue the thing up.
> Then procede like one double layer but already printed on out layers.
> It requires negative print outer layers and pattern plating with
> metallic resist so it's nothing easier after all. And fails on that
> etchback solutions will remove any useful ink printed on outer layers.
> It made me search for acidic desmear/etchback things. For just resin
> smear removal only, concentrated sulfuric acid room temp its said to do
> the job, but not etchback. There is also a version of permanganate
> desmear with sulfuric acid at room temp instead of caustic soda at 90�C
> but is not safe. And EG or PEG with sulfuric acid but slower than with
> MEA.

Adam Seychell wrote: 
> 
> I have no desmear removal step. Never had an inner layer electrical 
> problem yet. 

Maybe i didn't explain with the right English words. Etchback is used to
from etch resin sidewalls of holes in order exposing a copper anular
ring of innerlayers. There are good images on the web to explain this
but i didn't find any good right know, choose this because is Australian
and has other useful content
http://www.precisioncircuits.com.au/cid/011.html
on bottom 'etch back hole' but not easy understanding 
This is used to improve surface contact between plated copper and
innerlayers and should be used. For double sided board desmear or good
cleaning with air compressor or high pressure water on the holes is good
enough but it surprises me you could do it without even desmear for
multilayers, but just proves your genius skills again.

A good easy to use and maintenance free homebrew plating chemistry,
should not require high temp bath to be always ready to use or
incompatible consecutive baths for avoid long water rinses and its
disposal, and bath longer life. I don't like either microetch. So have
decided for palladium based. Palladium tin colloid with vanilla to
improve tin stability. The one i have used from bungard from what i know
now is like this and uses NaCl instead of HCl for Pd/Sn suspension i
guess to don't attack copper. The next bath is K2CO3 which removes
vanilla and with it tin and salt leaving only the palladium bonded to
resin surface. Or at least is what i can understand. 
It uses a mix of MEA and PEG600 (not really sure) for desmear at 60�C
with slow 3hour heat time. And a 25% NaCl and 1% HCl bath because of the
incompatibility of the desmear and the activation. 
I would like to make my homebrew version of desmear with a acidic
version compatible with the activation, i would love to use sulphuric
acid with EG antifreezer or DOT3/4 brake fluid for desmear but i can't
be sure it will work and it affects the bath compatibility. But the
K2CO3 at 40�C bath i don't know any replacement so it still requires a
water rinse and heat time here before the plating bath.
Your genius developing tank with nylon brushing fabric answered me
numerous ideas. Using it with NaOH as a degrease cleaning instead of
manually scrubbing is a saver way and time saver (no hands on chemicals
or gloves necessity), followed by one acidic desmear can de-oxidize the
board also, and be a bit more compatible with the activation.
My question, can you comment/help/improve the compatibility between this
NaOH degrease -  H2SO4 desmear - Pd/Sn organic activator - K2CO3 'salt
remover' - H2SO4/CuSO4 based plating, to minimize water rinses, heat
time, compatibility (residues from one bath don't affect next one) ? I
hope to have all this and be doing it in February finally.


> I think because professionals use 100k+ RPM drill speeds 
> there is much more probability of epoxy smear. My home made manual
> drill press runs at 15k RPM max.

I already suspected that you was able to do pth boards without a cnc (or
desmear) but is breath taking know that you said it...  Congratulations
i guess... Don't know what to say.

About drill speed, it's not that fast, and depends on drill diameter, a
0.5mm drill is recommended to run at 60.000rpm, 100krpm are for 0.2mm or
so, but a 2mm should run at 20krpm. But there are under 0.1mm drills and
above 200krpm spindles (high pressure air bearings). Pcb laminate
manufacturers say something about 'drill surface speed per minute' but
my few experiments with a 30-60krpm spindle said faster is not necessary
better. Maybe is just a recommendable limit speed and not recommended
speed.
http://www.isola-group.com/images/file/ISOLADE104iPGOriginalRelease3608.pdf
Check this file is from a common used laminate here in Europe, the file
has speed drill tables, and a pressure/temp vs time in multilayer
press. 

> 98% sulfuric is an excellent epoxy etch. In hot H2SO4 for few hours 
> your can remove all epoxy and just have nothing but bare copper
> traces.

Hot sulfuric acid shouldn't etch copper too?



Thanks.

Simao Cardoso wrote:
>  
> 
> Adam Seychell wrote:
>  >
>  > I have no desmear removal step. Never had an inner layer electrical
>  > problem yet.
> 
> Maybe i didn't explain with the right English words. Etchback is used to
> from etch resin sidewalls of holes in order exposing a copper anular
> ring of innerlayers. There are good images on the web to explain this
> but i didn't find any good right know, choose this because is Australian
> and has other useful content
> http://www.precisio ncircuits. com.au/cid/ 011.html 
> <http://www.precisioncircuits.com.au/cid/011.html>
> on bottom 'etch back hole' but not easy understanding
> This is used to improve surface contact between plated copper and
> innerlayers and should be used. For double sided board desmear or good
> cleaning with air compressor or high pressure water on the holes is good
> enough but it surprises me you could do it without even desmear for
> multilayers, but just proves your genius skills again.

Thanks for the interesting link. I've been looking for a good 
explanation of the various multilayer techniques but never found 
anything. My method is similar to "3) Double-Sided Cap Sheet Lamination"
  except I use liquid epoxy with glass cloth instead of pre-preg.
I do not claim desmear is unnecessary, but so far I've found it 
unnecessary in my situation. I am *assuming* because of my slow drill 
speed, that epoxy is less likely to melt and less or no smear over the 
copper annular ring.


> A good easy to use and maintenance free homebrew plating chemistry,
> should not require high temp bath to be always ready to use or
> incompatible consecutive baths for avoid long water rinses and its
> disposal, and bath longer life. I don't like either microetch. So have
> decided for palladium based. Palladium tin colloid with vanilla to
> improve tin stability. The one i have used from bungard from what i know
> now is like this and uses NaCl instead of HCl for Pd/Sn suspension i
> guess to don't attack copper. The next bath is K2CO3 which removes
> vanilla and with it tin and salt leaving only the palladium bonded to
> resin surface. Or at least is what i can understand.
> It uses a mix of MEA and PEG600 (not really sure) for desmear at 60�C
> with slow 3hour heat time. 

I wouldn't expect either of those to even touch epoxy. Epoxy is pretty 
inert. Desmear is normally a hot strong alkaline permanganate solution. 
I forget the exact composition. I found it turns the epoxy surface 
slightly brown, which I suspect is manganese dioxide.

And a 25% NaCl and 1% HCl bath because of the
> incompatibility of the desmear and the activation.

> I would like to make my homebrew version of desmear with a acidic
> version compatible with the activation, i would love to use sulphuric
> acid with EG antifreezer or DOT3/4 brake fluid for desmear but i can't
> be sure it will work and it affects the bath compatibility. But the
> K2CO3 at 40�C bath i don't know any replacement so it still requires a
> water rinse and heat time here before the plating bath.

> Your genius developing tank with nylon brushing fabric answered me
> numerous ideas. Using it with NaOH as a degrease cleaning instead of
> manually scrubbing is a saver way and time saver (no hands on chemicals
> or gloves necessity), followed by one acidic desmear can de-oxidize the
> board also, and be a bit more compatible with the activation.

I use it all the time and it made the developing process a *LOT* easier. 
Just turn on the heater, dip, wait 30 seconds, move up/down 10 strokes, 
and all done.

> My question, can you comment/help/ improve the compatibility between this
> NaOH degrease - H2SO4 desmear - Pd/Sn organic activator - K2CO3 'salt
> remover' - H2SO4/CuSO4 based plating, to minimize water rinses, heat
> time, compatibility (residues from one bath don't affect next one) ? I
> hope to have all this and be doing it in February finally.

Forget heated degrease bath, and desmear. After drilling, I remove the 
drill guide paper off the board (stuck with sticky tape), put on gloves, 
and place board in plastic tray, apply a little alkaline degreaser, 
scrub with 200 grit paper to remove drill burrs, scrub with ScotchBrite, 
then rinse. Its now ready for through hole 'activation'. If your 
activator contains HCl then I'd pre dip in HCl of same concentration 
without a rinse in between. Thats a simple room temp bath with almost 
unlimited shelf life.

> I already suspected that you was able to do pth boards without a cnc (or
> desmear) but is breath taking know that you said it... Congratulations
> i guess... Don't know what to say.

I've learned from previous projects that its sometimes best to attempt 
something in most simplest way first. Then progressively increase 
complexity until it works. I knew about drill smear long before doing 
PTH PCBs, but I omitted this step to see if it worked. I'm sure every 
step, tool and process used by professional PCB manufactures is proven 
to be absolutely necessary for reliable production. Us hobbyists are 
operating on a much smaller scale, and so it quite possible to take 
shortcuts in the process. For example, I found continuous filtration 
unnecessary in the Cu plating tank. The tank is covered 100% of the time 
and only clean PCBs are allowed to enter.


> About drill speed, it's not that fast, and depends on drill diameter, a
> 0.5mm drill is recommended to run at 60.000rpm, 100krpm are for 0.2mm or
> so, but a 2mm should run at 20krpm. But there are under 0.1mm drills and
> above 200krpm spindles (high pressure air bearings). Pcb laminate
> manufacturers say something about 'drill surface speed per minute' but
> my few experiments with a 30-60krpm spindle said faster is not necessary
> better. Maybe is just a recommendable limit speed and not recommended
> speed.


The faster the RPM the faster the feed, which means more holes per 
minute, which means more PCBs per machine, which means less machines per 
factory, which means less $ invested. I'm not sure if there is a lower 
limit. I guess a you could drill a 1 RPM , but it could take days to 
drill one hole.



> Hot sulfuric acid shouldn't etch copper too?
> 

I meant warm ~ 50C, but doesn't seem to attack copper. It was only a 
quick experiment I did some time ago. Its scary stuff to use. Would need 
some thinking to work out a safe way of applying it if you want to go 
down this path. H2SO4 is extremely hygroscopic too, so it won't last 
long in an unsealed tank.

Adam Seychell wrote: 
> 
> 
> Thanks for the interesting link. I've been looking for a good 
> explanation of the various multilayer techniques but never found 
> anything.

Replace the '11' in that url by '12' or start from '01' or click this
for the all thing.
http://www.precisioncircuits.com.au/pdf/cid_booklet.pdf
But there are better docs on the web and youtube videos, manuals from
lpkf bungard and mega have some nice info too. And 'Printed Circuits
Handbook' has almost 100 pages about it, but drawings are not much
explicit.

> My method is similar to "3) Double-Sided Cap Sheet Lamination"
> except I use liquid epoxy with glass cloth instead of pre-preg.

In you method you stack two double sided board leaving glue on the
center. Multilayers are normally stacks around one double sided board.
Adding prepreg and copper sheets around and press, just like making a
blank pcb laminate. I write about a method to print one time only,
following one idea of stacking single sided boards, which i think is
much easier for the homebrew. If the method i described didn't had
fails, since i can get 18panels 20x30cm 0.3mm thick by 32eur with tax
and postage will be a very cheap (and easy on print) method for
producing 1 to 6 layers small boards.


> I wouldn't expect either of those to even touch epoxy. Epoxy is pretty
> inert. Desmear is normally a hot strong alkaline permanganate
> solution. I forget the exact composition. I found it turns the epoxy
> surface slightly brown, which I suspect is manganese dioxide.

There are better desmear than permanganate, since it needs to be
replaced every time. Its used in much plating chemistries because of
oxidizing properties, where the activator bonds only to the resin hole
sidewall because its oxidized, and not to all board avoiding microetch
before plating. But are glycol ethers  desmears to use before
permanganate to make it last longer, and there are additives to make
permanganete regenerate, being Sodium hypochlorite (Bleach) the most
popular. The common setup from what i know is 90�C at pH around 10. A
example for one litre is  Potassium permanganate 60g; Sodium hydroxide
50g; Sodium hypochlorite 10g .
But 90�C is a no way for homebrew i don't want to wait hours until use
it. The room temp version is potassium permanganate and sulphuric acid
which is explosive, maybe few grams on 1 liter water won't explode but
it needs to be fresh all time so it fails on the same, only useful if
you have a multimillion dollar machine with concentration sensors and
precision pumps constantly renewing the thing.

About being MEA and PEG600 is what i think '2-aminoethanol and secondary
alchool ethoxylat (C12 C14)' from one MSDS means. Long shoot? I think
not. But i don't want MEA either, slow heating required and the vapor
makes you talk funny. 

> 
> Forget heated degrease bath, and desmear. After drilling, I remove the
> drill guide paper off the board (stuck with sticky tape), put on
> gloves, and place board in plastic tray, apply a little alkaline
> degreaser, scrub with 200 grit paper to remove drill burrs, scrub with
> ScotchBrite, then rinse. Its now ready for through hole 'activation'.


It's not hot, and i really need it! I should explain, i am not fat but
have this big strong warm hands which leave big marks on everything.
Since i am a lazy and don't wish to scrub or use gloves all the time,
your great tank idea answered this problems, i could degrease it on the
titanium holder without touch it. Your tank idea made me search for the
burrs problem (i hated the need to sand the boards when i worked with
them, and the expensive cnc was really against me on burrs matter), it
seems the machines for deburring  use high pressure water jets, i have
one water gun with a 3hp thiphase motor so i should try it.  
And btw do you know a homebrew version of this expensive ScotchBrite
roll brushes? They are used in pcb before lamination and after plating.
http://3m.hillas.com/pc-71200-867-3m-scotch-brite-printed-circuit-cleaning-brush-4-in-x-14-in-x-1-14-in-6s-sfn-1-per-case-custom.aspx


> If your activator contains HCl then I'd pre dip in HCl of same
> concentration without a rinse in between. Thats a simple room temp
> bath with almost unlimited shelf life.

What do you say about the Pd/Sn NaCl and vanilla activator i described?
I still didn't find the patent of the chemistry i used or the company
name which makes it. (bungard cuts the manufactor logo from labels on
bottles and sells then with the x-acto marks on the plastic! nice
business hein?!) I have metallic palladium here, should i convert it to
PdCl with aqua regia like in the books, with HCl and H2O2, bubbling air
in HCl for days, or electrolise in HCl?

> I've learned from previous projects that its sometimes best to attempt
> something in most simplest way first. Then progressively increase 
> complexity until it works. I knew about drill smear long before doing 
> PTH PCBs, but I omitted this step to see if it worked. I'm sure every 
> step, tool and process used by professional PCB manufactures is proven
> to be absolutely necessary for reliable production. Us hobbyists are 
> operating on a much smaller scale, and so it quite possible to take 
> shortcuts in the process. For example, I found continuous filtration 
> unnecessary in the Cu plating tank. The tank is covered 100% of the
> time and only clean PCBs are allowed to enter.

Continual active filtration and eductors are used like/insted the anode
bags for the same reason, removing copper shunks. By jetting filtered
plating solution against the pcb, only clean liquid is in contact with
the pcb. Is not about contamination, or it works or it doesn't work with
the contaminates you put in, no filtration can repair that. The machine
i have used don't have filtration either (but by the price should have
it and also a robotic arm to change pcb between tanks). But one time i
slit some anode bag content in to the tank when replacing anodes and was
a mess, hours of plating until that copper 'sand' disappear.

Is offtopic but see this gigantic machine, maybe the most advanced in
the world, in the video http://www.youtube.com/watch?v=5z8pV1ReuMs on
minute 5, no anode bags! because active filtering! (ok and some very
crazy chemistry - see the darken copper balls?)



> 
> I meant warm ~ 50C, but doesn't seem to attack copper. It was only a 
> quick experiment I did some time ago. Its scary stuff to use. Would
> need 
> some thinking to work out a safe way of applying it if you want to go 
> down this path. H2SO4 is extremely hygroscopic too, so it won't last 
> long in an unsealed tank.
> 

I would prefer a low concentration of sulphuric acid with EG or
derivates form things like anti freezer or brake fluid but this is still
a long shoot for me. I should keep to the things wrote by who really
knows. Just to get thiourea was a nightmare! To get all this will be
easier going to that Atotech representative at 40km from me.
Other thing, do you think i can mix easy to get NaHCO3 and KOH and heat
to get some K2CO3 in solution (along with Na2CO3), will it work for
vanilla-tin removal?

Another long email and is already really late in this side of the world,
very thanks.