double sided PCB, accuracy
2008-07-04 by Michel Baguet
Yahoo Groups archive
Index last updated: 2026-04-05 19:38 UTC
Thread
2008-07-04 by Michel Baguet
hi guys,
I got to the point where I can etch 6 mils traces and 32 mils vias with a 12 mils hole with inkjet paper. To align both side, I tried inkjet transparent but those get deformed at some places during printing and it's impossible to align the vias. A part of them are ok while other are only touching by the side.
Taping the transparent on 200g paper didn't help.
Do someone solve this problem ?
Mike.
[Non-text portions of this message have been removed]2008-07-04 by DJ Delorie
What I've been doing is including a "target" in my prints (four 25 mil squares in a square, with a 1 mil gap - that translates to a 1 mil ink crosshair when I print negatives). I print this on all photomasks and one on plain paper. The plain paper one I use to pre-drill 16 mil holes at each crosshair. Then, when exposing, I can line up the holes with the crosshairs pretty accurately, and double-check through the glass overlay.
2008-07-04 by Michel Baguet
Well, line up is not a problem. it's the transparent getting deformed when heated. --- On Fri, 7/4/08, DJ Delorie <dj@...> wrote:
From: DJ Delorie <dj@...>
Subject: Re: [Homebrew_PCBs] double sided PCB, accuracy
To: Homebrew_PCBs@yahoogroups.com
Date: Friday, July 4, 2008, 5:50 AM
What I've been doing is including a "target" in my prints (four 25 mil
squares in a square, with a 1 mil gap - that translates to a 1 mil ink
crosshair when I print negatives). I print this on all photomasks and
one on plain paper. The plain paper one I use to pre-drill 16 mil
holes at each crosshair. Then, when exposing, I can line up the holes
with the crosshairs pretty accurately, and double-check through the
glass overlay.
[Non-text portions of this message have been removed]2008-07-04 by fana cute
I never try this, but I read from some article (i've forgotten where) that problem is from your ink printer...and the article advised to use epson ink.. --- On Thu, 7/3/08, Michel Baguet <bagmik@...> wrote:
From: Michel Baguet <bagmik@...>
Subject: Re: [Homebrew_PCBs] double sided PCB, accuracy
To: Homebrew_PCBs@yahoogroups.com
Date: Thursday, July 3, 2008, 11:04 PM
Well, line up is not a problem. it's the transparent getting deformed when heated.
--- On Fri, 7/4/08, DJ Delorie <dj@...> wrote:
From: DJ Delorie <dj@...>
Subject: Re: [Homebrew_PCBs] double sided PCB, accuracy
To: Homebrew_PCBs@ yahoogroups. com
Date: Friday, July 4, 2008, 5:50 AM
What I've been doing is including a "target" in my prints (four 25 mil
squares in a square, with a 1 mil gap - that translates to a 1 mil ink
crosshair when I print negatives). I print this on all photomasks and
one on plain paper. The plain paper one I use to pre-drill 16 mil
holes at each crosshair. Then, when exposing, I can line up the holes
with the crosshairs pretty accurately, and double-check through the
glass overlay.
[Non-text portions of this message have been removed]
[Non-text portions of this message have been removed]2008-07-04 by Dylan Smith
On Thu, 3 Jul 2008, Michel Baguet wrote: > hi guys, > > I got to the point where I can etch 6 mils traces and 32 mils vias with > a 12 mils hole with inkjet paper. First, I presume you're using glossy inkjet paper for toner transfer from a laser printer. If that's the case you don't need to muck around with transparencies. Print both sides of the layout to your normal paper, place them together face in - hold them to a bright light and use the vias to align. Suitable bright lights are: the sun, a table lamp, or best, a light box where you can lay it all flat. Then tape the ends of the paper together at one end, to form a hinge. Re-check that it's still aligned. If it is, slip the copper board between the paper and get ironing. I've done lots of double sided boards this way - never had a failure.
2008-07-04 by Harvey White
On Fri, 4 Jul 2008 09:47:55 +0100 (BST), you wrote: I've always had a conceptual problem doing it exactly this way. I can see that if the joint between the two pieces of paper is exactly in the middle of the board thickness, every thing aligns. If it's not, then one layer is offset with respect to the other. Has anyone considered aligning the masks with a filler made from a scrap piece of etched board, same thickness, and then perhaps a piece of plastic to duplicate the board? Harvey
>On Thu, 3 Jul 2008, Michel Baguet wrote: > >> hi guys, >> >> I got to the point where I can etch 6 mils traces and 32 mils vias with >> a 12 mils hole with inkjet paper. > >First, I presume you're using glossy inkjet paper for toner transfer from >a laser printer. > >If that's the case you don't need to muck around with transparencies. >Print both sides of the layout to your normal paper, place them together >face in - hold them to a bright light and use the vias to align. Suitable >bright lights are: the sun, a table lamp, or best, a light box where you >can lay it all flat. > >Then tape the ends of the paper together at one end, to form a hinge. >Re-check that it's still aligned. If it is, slip the copper board between >the paper and get ironing. > >I've done lots of double sided boards this way - never had a failure.
2008-07-04 by Leon
----- Original Message -----
From: "Harvey White" <madyn@...> To: <Homebrew_PCBs@yahoogroups.com> Sent: Friday, July 04, 2008 2:09 PM Subject: Re: [Homebrew_PCBs] double sided PCB, accuracy > On Fri, 4 Jul 2008 09:47:55 +0100 (BST), you wrote: > > I've always had a conceptual problem doing it exactly this way. I can > see that if the joint between the two pieces of paper is exactly in > the middle of the board thickness, every thing aligns. If it's not, > then one layer is offset with respect to the other. > > Has anyone considered aligning the masks with a filler made from a > scrap piece of etched board, same thickness, and then perhaps a piece > of plastic to duplicate the board? That has been mentioned on this forum before, it's probably the best way to do it. A couple of pegs could be used with holes in the transparencies, to make the alignment even better. I think that something similar is used by the PCB manufacturers - they use dowels with their phototools. Leon -- Leon Heller Amateur radio call-sign G1HSM Yaesu FT-817ND and FT-857D transceivers Suzuki SV1000S motorcycle leon355@... http://www.geocities.com/leon_heller
2008-07-04 by Bertho Boman
Sticking the "tail" or joint between two half thickness spacers (strips) forces the tail to be in the middle and the artwork aligns. To use the alignment pins suggested a dedicated puncher with fixed spacing should be used and a way to accurately align the artworks to be punched. Of course the fixture pins must also accurately match the punch setup. Bertho
From: Harvey White Sent: Friday, July 04, 2008 09:09: I've always had a conceptual problem doing it exactly this way. I can see that if the joint between the two pieces of paper is exactly in the middle of the board thickness, every thing aligns. If it's not, then one layer is offset with respect to the other. Has anyone considered aligning the masks with a filler made from a scrap piece of etched board, same thickness, and then perhaps a piece of plastic to duplicate the board? Harvey [Non-text portions of this message have been removed]
2008-07-04 by Dylan Smith
On Fri, 4 Jul 2008, Harvey White wrote: > On Fri, 4 Jul 2008 09:47:55 +0100 (BST), you wrote: > > I've always had a conceptual problem doing it exactly this way. I can > see that if the joint between the two pieces of paper is exactly in > the middle of the board thickness, every thing aligns. If it's not, > then one layer is offset with respect to the other. I've never found it to be a problem. I only had bad registration on one board I've made this way, and even then it wasn't bad enough that the board couldn't be used. Most of the time the registration is pretty much spot on. If you need really fine tolerances for small vias then I suspect a piece of old board in the tape 'hinge' as you suggest would reduce the risk of mis-registration. I use 0.8mm drill bits, so my vias tend to be sized for that (except for things like jumper headers where I use 1.0mm bits), with 10 mil annular rings minimum (so that'd make my largest vias 35 mils, and the smaller ones 30 mils.
2008-07-04 by DJ Delorie
Michel Baguet <bagmik@...> writes: > Well, line up is not a problem. it's the transparent getting > deformed when heated. Ah, that. What I did was calibrate the deformity once (transfer a 3x3 square to a board and measure it) and compensate for it in PCB. To align two sides, I ended up using targets centered on the sides of the board, rather than diagonally opposite. Also, I've tried using a hand iron to tack down the paper before laminating it. Photomask is easier because there's no heat, you can see through the paper to line it up, and nothing moves during the exposure.
2008-07-04 by DJ Delorie
Harvey White <madyn@...> writes: > Has anyone considered aligning the masks with a filler made from a > scrap piece of etched board, same thickness, and then perhaps a > piece of plastic to duplicate the board? Another thing to try is to tape *three* corners, not two. The extra tension helps hold the seam centered on the thickness of the board. Not that I've ever done that; my TT paper is too opaque to line up that way. I've pre-drilled target holes in the PCB and poked matching holes in the TT paper to tack it down in the right place, with some success.
2008-07-04 by Michel Baguet
Using a filler is partway what I'm doing. I put a target at each corner of the two sheets, align them with needles then lift one of the sheet to insert a piece of PCB to get the right space between them. I place the filler outside the printed area but as close as possible from the needles. I tape the two sheet to the filler, remove the needles (the tape is supossed to keep the sheets aligned) and put the PCB to etch between the sheets, tape the sheet to the PCB and remove the filler.This is many manipulations and whatever you do, alignement is like playing with dices . Some times you win, some times you loose... Especially with 32 mils vias. This is the reason why I wanted to use transparent for inkjet. But there are two problems, the first is that the transparent get deformed, bended with the heat of the fusing unit inside the laser printer and the second, toner density is very light, even with density set at max. (laserjet 6P) And I'm afraid that the pcb will be etched with a lot of pinhole. So I believe I'll get back to blind alignement with plain paper by first driling 12 or 16 mils holes at the corner of the pcb to etch and holes in the sheets and align them with needles in the PCB. --- On Fri, 7/4/08, Harvey White <madyn@...> wrote:
From: Harvey White <madyn@...>
Subject: Re: [Homebrew_PCBs] double sided PCB, accuracy
To: Homebrew_PCBs@yahoogroups.com
Date: Friday, July 4, 2008, 1:09 PM
On Fri, 4 Jul 2008 09:47:55 +0100 (BST), you wrote:
I've always had a conceptual problem doing it exactly this way. I can
see that if the joint between the two pieces of paper is exactly in
the middle of the board thickness, every thing aligns. If it's not,
then one layer is offset with respect to the other.
Has anyone considered aligning the masks with a filler made from a
scrap piece of etched board, same thickness, and then perhaps a piece
of plastic to duplicate the board?
Harvey
>On Thu, 3 Jul 2008, Michel Baguet wrote:
>
>> hi guys,
>>
>> I got to the point where I can etch 6 mils traces and 32 mils vias with
>> a 12 mils hole with inkjet paper.
>
>First, I presume you're using glossy inkjet paper for toner transfer from
>a laser printer.
>
>If that's the case you don't need to muck around with transparencies.
>Print both sides of the layout to your normal paper, place them together
>face in - hold them to a bright light and use the vias to align. Suitable
>bright lights are: the sun, a table lamp, or best, a light box where you
>can lay it all flat.
>
>Then tape the ends of the paper together at one end, to form a hinge.
>Re-check that it's still aligned. If it is, slip the copper board between
>the paper and get ironing.
>
>I've done lots of double sided boards this way - never had a failure.
[Non-text portions of this message have been removed]2008-07-04 by Michel Baguet
the main problem is that the transparent get deformed while passing thru the fusing unit of the printer. --- On Fri, 7/4/08, Leon <leon355@btinternet.com> wrote:
From: Leon <leon355@...>
Subject: Re: [Homebrew_PCBs] double sided PCB, accuracy
To: Homebrew_PCBs@yahoogroups.com
Date: Friday, July 4, 2008, 1:21 PM
----- Original Message -----
From: "Harvey White" <madyn@embarqmail. com>
To: <Homebrew_PCBs@ yahoogroups. com>
Sent: Friday, July 04, 2008 2:09 PM
Subject: Re: [Homebrew_PCBs] double sided PCB, accuracy
> On Fri, 4 Jul 2008 09:47:55 +0100 (BST), you wrote:
>
> I've always had a conceptual problem doing it exactly this way. I can
> see that if the joint between the two pieces of paper is exactly in
> the middle of the board thickness, every thing aligns. If it's not,
> then one layer is offset with respect to the other.
>
> Has anyone considered aligning the masks with a filler made from a
> scrap piece of etched board, same thickness, and then perhaps a piece
> of plastic to duplicate the board?
That has been mentioned on this forum before, it's probably the best way to
do it. A couple of pegs could be used with holes in the transparencies, to
make the alignment even better. I think that something similar is used by
the PCB manufacturers - they use dowels with their phototools.
Leon
--
Leon Heller
Amateur radio call-sign G1HSM
Yaesu FT-817ND and FT-857D transceivers
Suzuki SV1000S motorcycle
leon355@btinternet. com
http://www.geocitie s.com/leon_ heller
[Non-text portions of this message have been removed]2008-07-04 by Michel Baguet
Do you use plain paper ? --- On Fri, 7/4/08, Dylan Smith <dyls@...> wrote:
From: Dylan Smith <dyls@...>
Subject: Re: [Homebrew_PCBs] double sided PCB, accuracy
To: Homebrew_PCBs@yahoogroups.com
Date: Friday, July 4, 2008, 3:38 PM
On Fri, 4 Jul 2008, Harvey White wrote:
> On Fri, 4 Jul 2008 09:47:55 +0100 (BST), you wrote:
>
> I've always had a conceptual problem doing it exactly this way. I can
> see that if the joint between the two pieces of paper is exactly in
> the middle of the board thickness, every thing aligns. If it's not,
> then one layer is offset with respect to the other.
I've never found it to be a problem. I only had bad registration on one
board I've made this way, and even then it wasn't bad enough that the
board couldn't be used. Most of the time the registration is pretty much
spot on. If you need really fine tolerances for small vias then I suspect
a piece of old board in the tape 'hinge' as you suggest would reduce the
risk of mis-registration. I use 0.8mm drill bits, so my vias tend to be
sized for that (except for things like jumper headers where I use 1.0mm
bits), with 10 mil annular rings minimum (so that'd make my largest vias
35 mils, and the smaller ones 30 mils.
[Non-text portions of this message have been removed]2008-07-04 by Michel Baguet
The transparent is already deformed when passing thru the fusing unit of the printer and the the deformation are not regular. For exemple, the corners align well while the top center is too low, the center is on the left and the center bottom is on the right... and so on. Different print give different deformation positions. --- On Fri, 7/4/08, DJ Delorie <dj@...> wrote:
From: DJ Delorie <dj@...>
Subject: Re: [Homebrew_PCBs] double sided PCB, accuracy
To: Homebrew_PCBs@yahoogroups.com
Date: Friday, July 4, 2008, 4:07 PM
Michel Baguet <bagmik@yahoo. com> writes:
> Well, line up is not a problem. it's the transparent getting
> deformed when heated.
Ah, that. What I did was calibrate the deformity once (transfer a 3x3
square to a board and measure it) and compensate for it in PCB. To
align two sides, I ended up using targets centered on the sides of the
board, rather than diagonally opposite. Also, I've tried using a hand
iron to tack down the paper before laminating it.
Photomask is easier because there's no heat, you can see through the
paper to line it up, and nothing moves during the exposure.
[Non-text portions of this message have been removed]2008-07-04 by Michel Baguet
What do you mean by "some succes" ? Is-it often, 3 out of 5, or less ? Is there a big off-registration ? --- On Fri, 7/4/08, DJ Delorie <dj@delorie.com> wrote:
From: DJ Delorie <dj@...>
Subject: Re: [Homebrew_PCBs] double sided PCB, accuracy
To: Homebrew_PCBs@yahoogroups.com
Date: Friday, July 4, 2008, 4:09 PM
Harvey White <madyn@embarqmail. com> writes:
> Has anyone considered aligning the masks with a filler made from a
> scrap piece of etched board, same thickness, and then perhaps a
> piece of plastic to duplicate the board?
Another thing to try is to tape *three* corners, not two. The extra
tension helps hold the seam centered on the thickness of the board.
Not that I've ever done that; my TT paper is too opaque to line up
that way. I've pre-drilled target holes in the PCB and poked matching
holes in the TT paper to tack it down in the right place, with some
success.
[Non-text portions of this message have been removed]2008-07-04 by Leon
----- Original Message -----
From: "Michel Baguet" <bagmik@...> To: <Homebrew_PCBs@yahoogroups.com> Sent: Friday, July 04, 2008 5:40 PM Subject: Re: [Homebrew_PCBs] double sided PCB, accuracy > the main problem is that the transparent get deformed while passing thru > the fusing unit of the printer. I think it only works properly when printing transparencies on an inkjet printer. Leon -- Leon Heller Amateur radio call-sign G1HSM Yaesu FT-817ND and FT-857D transceivers Suzuki SV1000S motorcycle leon355@... http://www.geocities.com/leon_heller
2008-07-04 by DJ Delorie
Try using laser transparencies instead of inkjet transparencies (assuming the toner will release afterwards). The pin trick is easier if you pre-poke the holes from the toner side, so you can tell where they are.
2008-07-04 by DJ Delorie
Michel Baguet <bagmik@...> writes: > What do you mean by "some succes" ? Is-it often, 3 out of 5, or > less? Is there a big off-registration ? Keep in mind I use 6/6 rules with 13 mil via drills, so "some success" for me might be "just fine" for you. Success for me is within 4-5 mils of dead center; I use my 13 mil drills as the pins. It's a lot of hassle, but if you take your time it does come out correct.
2008-07-04 by Michel Baguet
indeed for uv exposure but I'm doing toner tranfer --- On Fri, 7/4/08, Leon <leon355@...> wrote:
From: Leon <leon355@...>
Subject: Re: [Homebrew_PCBs] double sided PCB, accuracy
To: Homebrew_PCBs@yahoogroups.com
Date: Friday, July 4, 2008, 5:13 PM
----- Original Message -----
From: "Michel Baguet" <bagmik@yahoo. com>
To: <Homebrew_PCBs@ yahoogroups. com>
Sent: Friday, July 04, 2008 5:40 PM
Subject: Re: [Homebrew_PCBs] double sided PCB, accuracy
> the main problem is that the transparent get deformed while passing thru
> the fusing unit of the printer.
I think it only works properly when printing transparencies on an inkjet
printer.
Leon
--
Leon Heller
Amateur radio call-sign G1HSM
Yaesu FT-817ND and FT-857D transceivers
Suzuki SV1000S motorcycle
leon355@btinternet. com
http://www.geocitie s.com/leon_ heller
[Non-text portions of this message have been removed]2008-07-04 by Michel Baguet
A few days ago, I read on HP website that they are using a special mean to fuse the toner with the transparent so no possible release . --- On Fri, 7/4/08, DJ Delorie <dj@...> wrote:
From: DJ Delorie <dj@...>
Subject: Re: [Homebrew_PCBs] double sided PCB, accuracy
To: Homebrew_PCBs@yahoogroups.com
Date: Friday, July 4, 2008, 5:49 PM
Try using laser transparencies instead of inkjet transparencies
(assuming the toner will release afterwards).
The pin trick is easier if you pre-poke the holes from the toner side,
so you can tell where they are.
[Non-text portions of this message have been removed]2008-07-04 by Michel Baguet
Good idea, using the drill bits as a guide to align the sheet, same diameter as the holes should avoid misalignment --- On Fri, 7/4/08, DJ Delorie <dj@...> wrote:
From: DJ Delorie <dj@...>
Subject: Re: [Homebrew_PCBs] double sided PCB, accuracy
To: Homebrew_PCBs@yahoogroups.com
Date: Friday, July 4, 2008, 5:51 PM
Michel Baguet <bagmik@yahoo. com> writes:
> What do you mean by "some succes" ? Is-it often, 3 out of 5, or
> less? Is there a big off-registration ?
Keep in mind I use 6/6 rules with 13 mil via drills, so "some success"
for me might be "just fine" for you. Success for me is within 4-5
mils of dead center; I use my 13 mil drills as the pins. It's a lot
of hassle, but if you take your time it does come out correct.
[Non-text portions of this message have been removed]2008-07-04 by Bertho Boman
Why blind alignment? Crank up enough backlight and you should be able to see through the paper. Bertho From Michel Baguet Sent: Friday, July 04, 2008 12:39 ----> snip So I believe I'll get back to blind alignment with plain paper by first drilling 12 or 16 mils holes at the corner of the pcb to etch and holes in the sheets and align them with needles in the PCB.
2008-07-06 by Philip Pemberton
Bertho Boman wrote: > Why blind alignment? Crank up enough backlight and you should be able to > see through the paper. Perhaps a strobe light would be powerful enough. Assuming you could get used to the flickering :) But seriously, what about this -- 1 piece of FR4, 1.6mm (assuming that's what you're using to make the PCBs). Remove the copper, either by etching or sanding. Cut it into four 2"x0.5" strips. Drill a tiny - as in roughly the size of the needle - hole in each corner, about a quarter inch in from the top and left edges. Do the same thing for all four strips. Push four thin needles into the holes. Ideally these should push in and fit tightly. Also put some double-sided tape on the FR4. Both sides :) Now print your pattern with alignment holes ~2x0.5" away from the corners. Take one sheet of paper and push it onto the needles. Check alignment and remove the tape from each piece of FR4 in turn. Stick them down. Now put your PCB in the middle and do the same for the other side, align as before, stick down and remove the needles. Iron the board and cross your fingers... I'm still working on finding a decent TT paper that's still in production. Seems HP Everyday doesn't work any more... Still, it works in the inkjet (a rebuilt Epson C86 loaded with dye based ink) so it's not a total waste... -- Phil. ygroups@... http://www.philpem.me.uk/
2008-07-06 by Michel Baguet
I just find a way to solve the print geometry problem. Now I'm going to test blind alignment with needles and holes in the PCB of the same diameter. --- On Sun, 7/6/08, Philip Pemberton <ygroups@...> wrote:
From: Philip Pemberton <ygroups@...>
Subject: Re: [Homebrew_PCBs] double sided PCB, accuracy
To: Homebrew_PCBs@yahoogroups.com
Date: Sunday, July 6, 2008, 7:45 AM
Bertho Boman wrote:
> Why blind alignment? Crank up enough backlight and you should be able to
> see through the paper.
Perhaps a strobe light would be powerful enough. Assuming you could get used
to the flickering :)
But seriously, what about this --
1 piece of FR4, 1.6mm (assuming that's what you're using to make the PCBs).
Remove the copper, either by etching or sanding. Cut it into four 2"x0.5" strips.
Drill a tiny - as in roughly the size of the needle - hole in each corner,
about a quarter inch in from the top and left edges. Do the same thing for all
four strips.
Push four thin needles into the holes. Ideally these should push in and fit
tightly. Also put some double-sided tape on the FR4. Both sides :)
Now print your pattern with alignment holes ~2x0.5" away from the corners.
Take one sheet of paper and push it onto the needles. Check alignment and
remove the tape from each piece of FR4 in turn. Stick them down. Now put your
PCB in the middle and do the same for the other side, align as before, stick
down and remove the needles.
Iron the board and cross your fingers...
I'm still working on finding a decent TT paper that's still in production.
Seems HP Everyday doesn't work any more... Still, it works in the inkjet (a
rebuilt Epson C86 loaded with dye based ink) so it's not a total waste...
--
Phil.
ygroups@philpem. me.uk
http://www.philpem. me.uk/
[Non-text portions of this message have been removed]2008-07-06 by Bertho Boman
Try a regular good flash light or a halogen desk lamp. It works surprisingly well. A related trick if this method is used is to not use two identical targets as is normally done. Instead use complimentary targets. For example, on one layer use concentric circles of 5 mil lines and 5 mil spacing. On the other layer use the same dimensions but start with a space and then a black line. In other words, you will have a positive and a negative target. When placed on top of each other and viewed with light from below they will turn all black only when perfectly aligned. Any error is very obvious. Another comment: Taping just a hinge allows a slight movement of the layers since the hinge allows a microscopic movement. Cut off the corners of the top layer where the hinge will normally be placed. Placing tape on those corners really locks the artwork together without any movement. Finally, place two half thickness spacers one on each side of the tail to force it to be in the center when the board is put under the glass for exposure. For ironing use just one half sized spacer on the bottom and push the tail down on it to keep the top and bottom aligned. Bertho =================================
From: Philip Pemberton Sent: Sunday, July 06, 2008 03:45 Bertho Boman wrote: > Why blind alignment? Crank up enough backlight and you should be able to > see through the paper. Perhaps a strobe light would be powerful enough. Assuming you could get used to the flickering :)