Inkjet printing to copper - finally the ink resists! (like crazy!)
2006-08-01 by Stefan Trethan
Hi,
It seems i finally cracked it! Made PCBs with no more through-etching and
ink so tough not even acetone can take it off!
I spent a few hours cooking PCBs today. The curing was definitely my
problem.
Had to make a more scientific mechanism to carry out some experiments.
Basically an old electric hot plate with a aluminum plate on top and a
temperature meter attached. This way i could adjust the temperature within
+-10C and more evenly than with the hot air gun or the stove.
Then i printed many test PCBs and heated them to different temperatures,
from desiccation to delamination. It turns out that around 230C is the
sweet spot. keep that for several minutes (at least 3 i'd say for now but
more research needed). This incidentally is exactly the point where the
copper will start to go from just minimal yellowish oxidation to a purple
one (This is probably what Volkan is using to judge the temperature). It
is well away from damaging the PCB.
If you are much below or above this temperature the ink will not resist so
well. Round about 225C the ink starts to change somehow and can no longer
be cleaned off with acetone.
Anyway, that's what i can tell you today: cure on a 230C hot metal plate,
+-10C, for about 4 minutes and it will work.
I will make photos of the PCBs tomorrow, i need daylight my camera does
not make useable flash photographs.
There is definitely much more to be discovered here, but i finally found
repeatable parameters that so far work every time, all the time. There may
be other or even better ways to do this, but i finally have something that
works for me, and that's great!!
James, let me know what exactly would be required to qualify for the $50
prize ;-)
Below a writeup i started a few days ago while waiting for a PCB to etch,
mainly concerning the work on the printer. I want to put this on
<http://dipcb.schtuff.com/>, because i think a wiki might be a good way to
help further development of this process. If we write a wiki together it
will allow newcomers to get up to speed (and help us to remember).
Suggestions for that welcome.
Any my thanks again to Volkan!
Printer Modification:
*)Stripping down
First all plastic is removed around the C84 that is not required for
printing. This is the side caps, the main body, and some smaller loose
bits. At this point the printer will still print normally.
*)Preparation
make sure the printer is in good working order, and the parking station is
clean. With all used C84s you will find a badly clogged parking
station/vacuum mechanism that MUST be cleaned. Do a nozzle check to make
sure the head is working on all nozzles. I opted to disable the vacuum
mechanism by removing the small gear on the rough transport roller and
instead manually provide vacuum with a syringe. You can leave the pump
engaged if you do not mind unnecessary cleaning.
This is also the best point to install spongeless carts with auto-reset
chips and test them on paper. Fill yellow MSIPRO ink in the black cart
(you may also use the yellow cart the C84 can not tell which cart is where
as long as all are present). Add a few drops of black mispro ink for
better visibility.
*)Rear Paper path
The paper feeder is removed from the back, taking out the optical paper
sensor.
This sensor is glued to the side of the paper slot (in the metal chassis),
so that the material passing through here will activate it.
*)Front Paper Path
Remove the pizza wheels from the output. You may also remove the black
plastic paper guide and that ink sponge at the front. Rip out the whole
spiel, you only need to keep that greased sheetmetal rail where the head
rides on. It is easier to leave the shaft with the rubber output rollers
in place since it carries a pulley for the belt. It has no further
function and you could remove it if you tension the belt some other way.
Now a flat sheet will pass through the printer.
*)Central Transport Roller
Remove the middle pressure roller assy. Now a small PCB will fit through
the middle. If you cut it in half and replace the outer pressure rollers
with the half sections you get even larger PCBs through. You must hold the
springs up with something like a strip of metal or PCB to make up for the
cut-away half. If one needs full with even narrower wheels could be
attached at the very outside, but usually it will not be needed.
If you go for a rail under the carrier for 2-sided PCB alignment you need
to take out the rough transport roller and grind it down in the middle to
allow this rail to pass. Works well with a bench grinder. Don't grind it
too thin you need some strength.
*)Head Height
To allow for the increased media thickness we must adjust the head height.
The easiest method i found is to cut through the chassis just over the
paper gap. You could cut exactly through the paper gap, but that would
lift the pressure rollers as well and require a change there. The chassis
can be cut in place with nibbler pliers.
On the left (stepper/belt) side of the printer you need to add a small
piece of sheetmetal (ideally an angle piece around the rear edge) that
holds the cut-off part at about 3mm higher than it originally was. On the
right (head) side you can simply put nuts or washers under the two
mounting screws (and use longer screws) to achieve the same increase in
head height.
You also need to raise the head parking station, lift up the plastic over
the screw mount on the inside and again use a piece of plastic or
sheetmetal with screws to raise it on the outside. There was a snap-in
mechanism originally not allowing you to simply put something under a
screw. The inside in not held as firmly now, but that's fine and you could
use a longer screw and some strip of metal to clamp it down if you really
like.
Lastly you need to raise the front head rail by the same amount. Simply
put spacers (nuts) under where the screws attach the piece and use longer
screws.
*)Carrier
You need to make a carrier for your PCBs, any rigid material like
cardboard will work. It should not be too heavy. I used a sheet of formica
and glued thin cardboard to the bottom for traction (probably not needed).
If you want a rail for alignment you can glue a piece of aluminum to the
carrier. It must be low in height to fit in the gap in the transport
roller, but high enough to provide alignment between plastic guiderails.
If you do not need the rail a simple piece of cardboard will do.
The C84 will expect a short delay between the start of the feeding
mechanism and activation of the paper switch, so you need to make a cutout
of (edit) on the carrier to faciliate this delay. That is with the sensor
right at the metal, if you place it further outward you need a longer
cutout. The C84 is quite tolerant, a tolerance range of about +-4cm or so
was found to work. (Procedure for finding the valid range upon request if
you can't work it out yourself).
*)2-sided Alignment
If you need to align printing on the second side of the PCB it would be
good to be able to print in the same spot each time. To ensure that you
can put a rail on the underside of the carrier which rides between plastic
guides on both the input and output of the printer. To mount these plastic
guides it is easier to cut out the plastic bottom of the printer to allow
for wooden mounting blocks. I have not yet installed this fully as aligned
doublesided printing is not a priority.
Printing:
*)PCB Preparation
Sand the PCB with 600 or 1000 grit paper so that he surface appears
scratched everywhere. You may also use a abrasive plastic pad sold for
cleaning copper pipes before soldering. Then wipe with acetone, twice, the
first pass should take away the copper dust. The PCB must be grease free
now and stay that way.
*)Printing
To print you must insert the carrier manually into the printer. Turn the
printer on, and wait until it has completed it's dance. Now feed the edge
of the carrier into the printer so that it jsut protrudes a few mm from
the traction/pressure rollers. Make sure it is straight so it will not run
against an edge when the printer feeds.
Now rint the artwork on your carrier to find the right position (stick
some tape or adhesive backed plastic foil on it or you may end up with
permanent marking of the carrier). You can use low resolution for this
print since you only need to know where it will end up. If you use the
fastest draft setting watch out, the printer virtually shoots it through.
Now place the PCB over this position and stick it down with tape. You may
delay the acetone cleaning of the PCB until this point if you prefer.
Again feed the carrier and print the artwork.
*)Printer Settings
When printing a purely black test pattern i found the transparency setting
does not use any ink from the black cart at all.
The other settings use ink in varying amounts, the most ink is used for
matte paper and the least for glossy or durabright setting. You need Photo
or Best Photo resolution. The "fast" setting, which prints in both
directions of travel, produces better results for me. No edge smoothing
needed.
*)Curing
-- see top of message --
*)Etching
I suspected Ferric chloride is less agressive to the ink than CuCl, so i
set up a test. I cut a test PCB in half after curing to get identical
samples, and put one in CuCl and one in ferric chloride. If anything the
ferric chloride showed more agressive under-etching.
It seems i finally cracked it! Made PCBs with no more through-etching and
ink so tough not even acetone can take it off!
I spent a few hours cooking PCBs today. The curing was definitely my
problem.
Had to make a more scientific mechanism to carry out some experiments.
Basically an old electric hot plate with a aluminum plate on top and a
temperature meter attached. This way i could adjust the temperature within
+-10C and more evenly than with the hot air gun or the stove.
Then i printed many test PCBs and heated them to different temperatures,
from desiccation to delamination. It turns out that around 230C is the
sweet spot. keep that for several minutes (at least 3 i'd say for now but
more research needed). This incidentally is exactly the point where the
copper will start to go from just minimal yellowish oxidation to a purple
one (This is probably what Volkan is using to judge the temperature). It
is well away from damaging the PCB.
If you are much below or above this temperature the ink will not resist so
well. Round about 225C the ink starts to change somehow and can no longer
be cleaned off with acetone.
Anyway, that's what i can tell you today: cure on a 230C hot metal plate,
+-10C, for about 4 minutes and it will work.
I will make photos of the PCBs tomorrow, i need daylight my camera does
not make useable flash photographs.
There is definitely much more to be discovered here, but i finally found
repeatable parameters that so far work every time, all the time. There may
be other or even better ways to do this, but i finally have something that
works for me, and that's great!!
James, let me know what exactly would be required to qualify for the $50
prize ;-)
Below a writeup i started a few days ago while waiting for a PCB to etch,
mainly concerning the work on the printer. I want to put this on
<http://dipcb.schtuff.com/>, because i think a wiki might be a good way to
help further development of this process. If we write a wiki together it
will allow newcomers to get up to speed (and help us to remember).
Suggestions for that welcome.
Any my thanks again to Volkan!
Printer Modification:
*)Stripping down
First all plastic is removed around the C84 that is not required for
printing. This is the side caps, the main body, and some smaller loose
bits. At this point the printer will still print normally.
*)Preparation
make sure the printer is in good working order, and the parking station is
clean. With all used C84s you will find a badly clogged parking
station/vacuum mechanism that MUST be cleaned. Do a nozzle check to make
sure the head is working on all nozzles. I opted to disable the vacuum
mechanism by removing the small gear on the rough transport roller and
instead manually provide vacuum with a syringe. You can leave the pump
engaged if you do not mind unnecessary cleaning.
This is also the best point to install spongeless carts with auto-reset
chips and test them on paper. Fill yellow MSIPRO ink in the black cart
(you may also use the yellow cart the C84 can not tell which cart is where
as long as all are present). Add a few drops of black mispro ink for
better visibility.
*)Rear Paper path
The paper feeder is removed from the back, taking out the optical paper
sensor.
This sensor is glued to the side of the paper slot (in the metal chassis),
so that the material passing through here will activate it.
*)Front Paper Path
Remove the pizza wheels from the output. You may also remove the black
plastic paper guide and that ink sponge at the front. Rip out the whole
spiel, you only need to keep that greased sheetmetal rail where the head
rides on. It is easier to leave the shaft with the rubber output rollers
in place since it carries a pulley for the belt. It has no further
function and you could remove it if you tension the belt some other way.
Now a flat sheet will pass through the printer.
*)Central Transport Roller
Remove the middle pressure roller assy. Now a small PCB will fit through
the middle. If you cut it in half and replace the outer pressure rollers
with the half sections you get even larger PCBs through. You must hold the
springs up with something like a strip of metal or PCB to make up for the
cut-away half. If one needs full with even narrower wheels could be
attached at the very outside, but usually it will not be needed.
If you go for a rail under the carrier for 2-sided PCB alignment you need
to take out the rough transport roller and grind it down in the middle to
allow this rail to pass. Works well with a bench grinder. Don't grind it
too thin you need some strength.
*)Head Height
To allow for the increased media thickness we must adjust the head height.
The easiest method i found is to cut through the chassis just over the
paper gap. You could cut exactly through the paper gap, but that would
lift the pressure rollers as well and require a change there. The chassis
can be cut in place with nibbler pliers.
On the left (stepper/belt) side of the printer you need to add a small
piece of sheetmetal (ideally an angle piece around the rear edge) that
holds the cut-off part at about 3mm higher than it originally was. On the
right (head) side you can simply put nuts or washers under the two
mounting screws (and use longer screws) to achieve the same increase in
head height.
You also need to raise the head parking station, lift up the plastic over
the screw mount on the inside and again use a piece of plastic or
sheetmetal with screws to raise it on the outside. There was a snap-in
mechanism originally not allowing you to simply put something under a
screw. The inside in not held as firmly now, but that's fine and you could
use a longer screw and some strip of metal to clamp it down if you really
like.
Lastly you need to raise the front head rail by the same amount. Simply
put spacers (nuts) under where the screws attach the piece and use longer
screws.
*)Carrier
You need to make a carrier for your PCBs, any rigid material like
cardboard will work. It should not be too heavy. I used a sheet of formica
and glued thin cardboard to the bottom for traction (probably not needed).
If you want a rail for alignment you can glue a piece of aluminum to the
carrier. It must be low in height to fit in the gap in the transport
roller, but high enough to provide alignment between plastic guiderails.
If you do not need the rail a simple piece of cardboard will do.
The C84 will expect a short delay between the start of the feeding
mechanism and activation of the paper switch, so you need to make a cutout
of (edit) on the carrier to faciliate this delay. That is with the sensor
right at the metal, if you place it further outward you need a longer
cutout. The C84 is quite tolerant, a tolerance range of about +-4cm or so
was found to work. (Procedure for finding the valid range upon request if
you can't work it out yourself).
*)2-sided Alignment
If you need to align printing on the second side of the PCB it would be
good to be able to print in the same spot each time. To ensure that you
can put a rail on the underside of the carrier which rides between plastic
guides on both the input and output of the printer. To mount these plastic
guides it is easier to cut out the plastic bottom of the printer to allow
for wooden mounting blocks. I have not yet installed this fully as aligned
doublesided printing is not a priority.
Printing:
*)PCB Preparation
Sand the PCB with 600 or 1000 grit paper so that he surface appears
scratched everywhere. You may also use a abrasive plastic pad sold for
cleaning copper pipes before soldering. Then wipe with acetone, twice, the
first pass should take away the copper dust. The PCB must be grease free
now and stay that way.
*)Printing
To print you must insert the carrier manually into the printer. Turn the
printer on, and wait until it has completed it's dance. Now feed the edge
of the carrier into the printer so that it jsut protrudes a few mm from
the traction/pressure rollers. Make sure it is straight so it will not run
against an edge when the printer feeds.
Now rint the artwork on your carrier to find the right position (stick
some tape or adhesive backed plastic foil on it or you may end up with
permanent marking of the carrier). You can use low resolution for this
print since you only need to know where it will end up. If you use the
fastest draft setting watch out, the printer virtually shoots it through.
Now place the PCB over this position and stick it down with tape. You may
delay the acetone cleaning of the PCB until this point if you prefer.
Again feed the carrier and print the artwork.
*)Printer Settings
When printing a purely black test pattern i found the transparency setting
does not use any ink from the black cart at all.
The other settings use ink in varying amounts, the most ink is used for
matte paper and the least for glossy or durabright setting. You need Photo
or Best Photo resolution. The "fast" setting, which prints in both
directions of travel, produces better results for me. No edge smoothing
needed.
*)Curing
-- see top of message --
*)Etching
I suspected Ferric chloride is less agressive to the ink than CuCl, so i
set up a test. I cut a test PCB in half after curing to get identical
samples, and put one in CuCl and one in ferric chloride. If anything the
ferric chloride showed more agressive under-etching.