Homebrew_PCBs

Hi all, great group. I'm a relative newbie using Eagle to make two
sided boards. Have build a cnc machine that will does double duty to
drill the holes for me and use TT with 10mil tracks with magazine
paper and even regular hi grade copier paper on a HP IIIP with fair
success. Tried milling isolation but that's no good for more complex
or densly packed boards.

Here's my problem I'm doing a lot (for me) of prototype boards with
db25 connections right angle and 0.100 dual row connectors. What's
the best way to deal with the problem of through holes for these?
That is I often find that I end up with a bottom and top track
meeting at a pin on one of these connectors and I can't solder the
top connection on all the pins?

I'm finding that I have to account for that as much as I can routing
but not always able to, drill over size and wire through before
fitting the connector or using a lot of via's. I'm presently trying
to produce 3.5*2." LPT opto isolator with two db25's, 4 9 pin RN's,
four 16 pin opto's, one 4 pin opt and a bunch of led's and my amatuer
status is showing and it looks pretty ugly right now......


Any obvious options that I've missed? the press in through hole
grommit things and tooling seem to be more than I want to pay.

thanks..

fyffe555 wrote:

> Here's my problem I'm doing a lot (for me) of prototype boards with
> db25 connections right angle and 0.100 dual row connectors. What's
> the best way to deal with the problem of through holes for these?

Get away from the right angle connectors if you can. It is much easier to
route and mount standard solder cup connectors with the board in between the
rows, and one row on top and one on bottom, and simply solder straight to the
board. Eagle has layouts for this, look at M25D for an example. If you can use
it it's much easier than the other DB-25 layouts.

Alan

heh heh, through hole plating can be done but I've never tried it. I
basically try to avoid through hole connections on the component
layer. I use the routing costs to favor the bottom layer and then
fix up any inaccessable solder locations with vias. Sometimes I'll
first route the board only on the bottom layer which will only get so
far and then restart the routing but allow the top layer too. then
fix with vias. Just dont forget to solder the via wires first like
I've done.

Overall, its usually not too bad.


--- In Homebrew_PCBs@yahoogroups.com, "fyffe555" <ap_leapyear@h...>
wrote:

> Hi all, great group. I'm a relative newbie using Eagle to make two
> sided boards. Have build a cnc machine that will does double duty

to

> drill the holes for me and use TT with 10mil tracks with magazine
> paper and even regular hi grade copier paper on a HP IIIP with fair
> success. Tried milling isolation but that's no good for more

complex

> or densly packed boards.
>
> Here's my problem I'm doing a lot (for me) of prototype boards with
> db25 connections right angle and 0.100 dual row connectors. What's
> the best way to deal with the problem of through holes for these?
> That is I often find that I end up with a bottom and top track
> meeting at a pin on one of these connectors and I can't solder the
> top connection on all the pins?
>
> I'm finding that I have to account for that as much as I can

routing

> but not always able to, drill over size and wire through before
> fitting the connector or using a lot of via's. I'm presently trying
> to produce 3.5*2." LPT opto isolator with two db25's, 4 9 pin

RN's,

> four 16 pin opto's, one 4 pin opt and a bunch of led's and my

amatuer

> status is showing and it looks pretty ugly right now......
>
>
> Any obvious options that I've missed? the press in through hole
> grommit things and tooling seem to be more than I want to pay.
>
> thanks..

In a message dated 3/3/2004 11:06:18 AM Central Standard Time,
ap_leapyear@... writes:
What's the best way to deal with the problem of through holes for these?

Assume you have a "pad" on top, as well as bottom, for each "through-hole"
you need? As the pins you mention are generally square (not round, filling the
hole), you can usually get away with some bits of (sliver-plated, usually!)
wire-wrap wire, #28 or #30 ga., stripped.

Put the end of a length through the hole, bend it over sharply on top, and
solder it to the pad, NOT filling the hole. MIGHT take some practice,
particularly if you are coordination-challenged or have less than twenty years
solder-fiddling. You MIGHT even want to insert a temporary NON-solderable pin of some
kind (black drill-bit shank?) which fills the hole, and then bend that thin
wire around that bit, over the pad, so you can solder "all around". Then clip
off the wire below so that a bit of it remains below. Repeat for all "needed
PTH holes". THEN insert your connector, and solder a couple of pads that do
NOT have these "through wires", to HOLD the thing in place. Now neatly bend
these stubs of thru-wires around the pins, and solder all conventionally. There
you have a "hidden" jumper soldered on top as well as bottom. Works for me!
However, the tiny rivets are SO much nicer that I generally at least
TRY to use those where I can. BUT, if I ever do, I ALWAYS solder them on top,
being sure solder "wicks" under the heads! Yes, they DO require larger holes
drilled for them, but sometimes that is just "life"! Also, if their heads
are TOO large, sometimes they are "too close" for 0.1" IC spacing, etc. Just
part of the hobby!

One other thing that makes all this top/bottom soldering easier, AND makes
the boards look and STAY-looking nicer is to use "Tinnit" solution (PER
instructions!) to Sn-plate the copper. Some do this as "resist" and some do it AFTER
etching, and even "re-flow" with heat, as the professional boards are
generally done. Sadly, the Tinnit solids don't seem to have much shelf-life! SOME
vendor should WORK on that for us, doncha think?????!!!!!!!!!

Jan Rowland


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

> I'm finding that I have to account for that as much as I can routing
> but not always able to, drill over size and wire through before
> fitting the connector or using a lot of via's. I'm presently trying
> to produce 3.5*2." LPT opto isolator with two db25's, 4 9 pin RN's,
> four 16 pin opto's, one 4 pin opt and a bunch of led's and my amatuer
> status is showing and it looks pretty ugly right now......
>
>
> Any obvious options that I've missed? the press in through hole
> grommit things and tooling seem to be more than I want to pay.
>
> thanks..
>

Most simple solution:
Do not make a design with a connection on the top side.
Do everything to avoid this.
it IS possible.

you can solder the two-rows connectors on the top, use one with longer
legs and keep the plastic 2-3mm above the pcb top.

the other things (wires etc) were suggested. i would not use eylets
(too expensive).

DB25 are available in a version where you can solder the top two rows
also,
they are the ones with no plastic angle and turned, bent pins.
I have not seen them for a while now, maybe they are obsolete.

ST

You can also see pic's of the process here:
http://www.pic101.com/pcb/thru.htm
Rick

JanRwl@... wrote:

>
> What's the best way to deal with the problem of through holes for
> these?
>
> Assume you have a "pad" on top, as well as bottom, for each
> "through-hole"
> you need? As the pins you mention are generally square (not round,
> filling the
> hole), you can usually get away with some bits of (sliver-plated,
> usually!)
> wire-wrap wire, #28 or #30 ga., stripped.
>
> Put the end of a length through the hole, bend it over sharply on top,
> and
> solder it to the pad, NOT filling the hole. MIGHT take some practice,
>
> particularly if you are coordination-challenged or have less than
> twenty years
> solder-fiddling.

In a message dated 3/3/2004 3:57:23 PM Central Standard Time,
rixy@... writes:
You can also see pic's of the process here:
http://www.pic101.com/pcb/thru.htm

Amazing! Exactly what I was describing has already been "E-documented" as
Rick says!


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

Stefan Trethan wrote:

> Most simple solution:
> Do not make a design with a connection on the top side.
> Do everything to avoid this.
> it IS possible.
>

Yes that is the other excellent suggestion, a few well placed jumper wires on
the bottom out to a pad for the rest of the trace solves almost every problem
like this with very little fuss.

Alan

Unless you're designing for production. If you get too used to doing
that, it will be a bad habit to break.
Rick

Alan King wrote:

> Stefan Trethan wrote:
> > Most simple solution:
> > Do not make a design with a connection on the top side.
> > Do everything to avoid this.
> > it IS possible.
> >
>
> Yes that is the other excellent suggestion, a few well placed
> jumper wires on
> the bottom out to a pad for the rest of the trace solves almost every
> problem
> like this with very little fuss.
>
> Alan

Rick C. wrote:

> Unless you're designing for production. If you get too used to doing
> that, it will be a bad habit to break.
> Rick
>
> Alan King wrote:
>

Well few production items make sense to worry about having double sided
without through holes anyway. And I'd still disagree with that opinion on other
grounds, in general it develops your layout skills far more to work on good
topology and have the minimum number of jumpers with everything on the bottom.
Trivial to go to some easier method, so hard to consider it a bad habbit. For
the most part designing towards any goal strengthens your skills for designing
towards other goals, the particular goal for a particular case hardly matters.
I could design for months straight this way, and then still have no problem
doing something else, and I bet most other people could too.. I mainly do SM
single sided boards now for no holes because it makes sense, but it hasn't made
me bad at still doing a double sided when needed.

Alan

>> Alan King wrote:
>>
>
> Well few production items make sense to worry about having double
> sided
> without through holes anyway. And I'd still disagree with that opinion
> on other
> grounds, in general it develops your layout skills far more to work on
> good
> topology and have the minimum number of jumpers with everything on the
> bottom.
> Trivial to go to some easier method, so hard to consider it a bad
> habbit. For
> the most part designing towards any goal strengthens your skills for
> designing
> towards other goals, the particular goal for a particular case hardly
> matters.
> I could design for months straight this way, and then still have no
> problem
> doing something else, and I bet most other people could too.. I mainly
> do SM
> single sided boards now for no holes because it makes sense, but it
> hasn't made
> me bad at still doing a double sided when needed.
>
> Alan
>

I fully agree with that.
you have to work with what you get here and now, and using it the best way
possible.
Keeping things simple is no bad habit i would say.

Also keep in mind how many "low end" electronics is still on single sided
paper/resin board.

I hate it when i come across the most simple amateur circuits in the web
which need to use
both layers, with a design complexity that would allow building them on
half a layer ;-).

I think that is rather a bad habit.

I said before, routing is an art, many disagreed, but i still think it is
so.


Another case is when you make a prototype of a production series, with a
homebrew board.
then of course you can not route differently....


ST

>From: "fyffe555" <ap_leapyear@...>
>Reply-To: Homebrew_PCBs@yahoogroups.com
>To: Homebrew_PCBs@yahoogroups.com
>Subject: [Homebrew_PCBs] Through hole solutions?
>Date: Wed, 03 Mar 2004 16:58:40 -0000
>
>Hi all, great group. I'm a relative newbie using Eagle to make two
>sided boards. Have build a cnc machine that will does double duty to
>drill the holes for me and use TT with 10mil tracks with magazine
>paper and even regular hi grade copier paper on a HP IIIP with fair
>success. Tried milling isolation but that's no good for more complex
>or densly packed boards.
>
>Here's my problem I'm doing a lot (for me) of prototype boards with
>db25 connections right angle and 0.100 dual row connectors. What's
>the best way to deal with the problem of through holes for these?
>That is I often find that I end up with a bottom and top track
>meeting at a pin on one of these connectors and I can't solder the
>top connection on all the pins?

You could avoid the problem by using vias with wire links through them.

Leon
--
Leon Heller, G1HSM Tel: +44 1424 423947
Email: aqzf13 at dsl dot pipex dot com
WWW: http://www.geocities.com/leon_heller

_________________________________________________________________
Express yourself with cool new emoticons http://www.msn.co.uk/specials/myemo

Stefan Trethan wrote:

>>>Alan King wrote:
>>>
>>
>> Well few production items make sense to worry about having double
>>sided
>>without through holes anyway. And I'd still disagree with that opinion
>>on other
>>grounds, in general it develops your layout skills far more to work on
>>good
>>topology and have the minimum number of jumpers with everything on the
>>bottom.
>>Trivial to go to some easier method, so hard to consider it a bad
>>habbit. For
>>the most part designing towards any goal strengthens your skills for
>>designing
>>towards other goals, the particular goal for a particular case hardly
>>matters.
>>I could design for months straight this way, and then still have no
>>problem
>>doing something else, and I bet most other people could too.. I mainly
>>do SM
>>single sided boards now for no holes because it makes sense, but it
>>hasn't made
>>me bad at still doing a double sided when needed.
>>
>>Alan
>>
>
>
> I fully agree with that.
> you have to work with what you get here and now, and using it the best way
> possible.
> Keeping things simple is no bad habit i would say.
>
> Also keep in mind how many "low end" electronics is still on single sided
> paper/resin board.
>
> I hate it when i come across the most simple amateur circuits in the web
> which need to use
> both layers, with a design complexity that would allow building them on
> half a layer ;-).

Check this case in unnecessary design:

http://www.rigelcorp.com/r535j.htm
http://www.rigelcorp.com/__doc/8051/R535JASSM.pdf

It's a development board I got a long time ago, like 94 or 95. If you look
in the assembly manual, there's a picture in the first 2 or 3 pages. Don't see
a photo right now on the site.

Truly silly thing was it was a 4 layer board. Way back then when they were
very expensive. And totally not required, when a friend and I got ours, we
noticed this and looked at it very closely for WHY was it 4 layers before
putting components on. Not the chips, they're all fairly simple routing and
could have easily been taken care of. Instead it was the simple components with
many crossed wires, the resistors, caps, and transistors. They placed, and
didn't move a few minor items to make it a double sided board. The board was
high quality too, not laminated sheets but etched, thick layer resist, more
copper, then other etch or similar. The green highly translucent type and you
can see the quality and no thick lamination sheet for the other layers. Suppose
it could have been laminated though, with very thin sheet. Of course it was
probably never high quantity, and he may have been making his own 4 layer
boards, so may not have been too big an issue. Still would have been only a few
minutes work to reduce it to a two layer and have a much simpler board. If they
were buying these all this time though they've wasted a heck of a lot of money.
Very good board great for soldering, but much of a waste..

I still have it, will take a small pic so you guys can see it next time I run
into it. Board probably cost $25 instead of the $5 it should have back in 95 or
so, may not be such a difference now but think of the lost $20 each pop. At
that time at least they probably sold hundreds a year of these. Probably 5 or
10 K in waste a year if they didn't make it themselves. Of course not too long
after PIC's etc started getting really popular and cheap and made more sense.
Never even used the board much for this reason, got into PICs more right after
getting it.

It would be very nice to find these materials though, and get home made
boards up to this quality level if possible, it does stand out as one of the
best boards I've ever worked with. The good through hole plating may have been
much of this extra quality though, things didn't lift etc with any amount of heat.



And I wonder if they ever fixed their LCD code, had a bug in the Enable line
wait logic that made it very spotty until I fixed a couple of lines. We thought
it was something flaky with our LCDs until I found it and everything worked
after that, and I bet most everyone else with one thought the same.

I noticed on the end of the mail that it went off topic,
even off topic of the already a bit off topic discussion.
If you still want to read it go on...

>
> Check this case in unnecessary design:
>
> http://www.rigelcorp.com/r535j.htm
> http://www.rigelcorp.com/__doc/8051/R535JASSM.pdf
>

Especially for a prototype / evaluation board 2 layer is a great advantage.
you can follow the tracs and you also can make changes as needed.
i would like to see how you make changes in a 4 layer board.


You can take apart whatever consumer equipment you like, 2 layer boards are
still very popular, as are 1 layer boards.

I do not know how this will develop, the demanding boards have more layers
now,
but also the things get more and more integrated.
If you look how many layers they use for IC internal interconnection (14
or so??)
you can maybe guess how the boards will develop. On the other hand, as the
integrated
circuits get more and more functionality, requiring nearly no external
components
the external circuit should get simpler, allowing less layers there.

The requirements are so different - some things need to be as small as
possible, some
as cheap as possible, some both, some high quality, some easy to
manufacture.
the demands are very different, and i think the solutions to meet them
will be as different
in the future.

Development is not stopping now, that is the only thing you can be sure of.

I think the trend will be towards integration, but the human is not built
to interface silicone chips, it needs to use the fingers to operate the
stuff
and needs audio and video feedback. this human interface is (in this way)
impossible
to integrate into anything (on chip buttons and such ;-).
maybe the soultion to that is that the human carrys a
"Human interface device" with him all time, and the machines we build have
no direct
interface at all, and communicate wireless with the HID of the individual.

This has certain advantages, but also a lot of disadvantages.
I'm known around here for my skills with electrics/electronics so if there
is some problem
people call me to help them. I regualrly come somewhere only to discover
there is no fault, but it
is just a user "inability". some equipment is simply too complicated for
some people, especially
elderly people with disabilities like bad sight and such. If the TV or sat
receiver is operated through
a complicated menu and they simply can't remember how this works, or if
the buttons on the remote are simply
to small to read the lable (this happens) then i really question the
purpose.
I mean, back a few years each tv had a row of buttons, and that's it. just
press and look.

Now, if you follow the thought of the "universal human interface" this
could lead to two consequences:
a) everything gets far too complicated for some. this limits the market -
BAD.
b) you can program it, also in a "simple" mode. so that the stuff works
with maybe limited functionality,
but WORKS. you could design human interfaces even for people with severe
disabilities if you like.


Sure, the society changes, gets used to the technology, learns to keep up
with it, but everyone?
and with each and every machine?
It happens not seldom that one of my professors, on a technical
university, (teaching electrotechnics or much more advanced stuff) can not
manage to make the laptop
computer operate with the video beamer and switch all things on. Not that
he is too stupid for it, of course
not, if he has some time and reads the manual it would work easily, but
there and then there
is no manual, and no time to read it. He wants nothing else than a big
green ON button and a big red OFF button.
he has no interest in all else of the functionality.

On the other hand, if you know what you do exactly, and the machine tells
you "you can't do that now"
you are also upset (like formatting your system partition on the PC).

I could go on and on, but i think everyone who kept up reading so far got
my point.

There will still be much development going on, no doubt, faster, cheaper,
more functionality.
But we also need to keep in mind that the limited human needs to operate
and use the stuff. Not the designer,
which knows every function and error, the simple consumer who wants not to
waste much time with the manual.



THE END ;-)

Didn't read it after writing, there may be much errors.. you can keep them
i have plenty.
look forward to discussion and possible solutions.

ST

to continue on this thread. I've found that if I lay out my board
and then look at how the traces run, I can make a few tweaks and
significantly reduce routing complexity. For example, I put an LCD
display header on a board I'm doing now. after first routing, I
noticed that a couple of the traces were going all over the place. I
went back and switched 2 of the control lines going imto the micro-
controller and rerouted. wow, what a difference, the traces were
direct and I was able to get rid of 8 vias (4 top side wires). Its
still 2 sided but, my design is now much more manufacturable in my
little "pcb factory".

Another thing is that routers are pretty dumb. at least the Eagle
one is. I'll often route, rip-up, hand route a couple of problem
traces and then reroute to a much better result. It also pays to
understand the router cost tables and what each cost item means.
I've gotten comfortable at getting eagle to route the way I want but
it still makes some stupid choices and winds up putting waaaay too
many traces on the component side even with top layer at max cost
(for through hole stuff).

This brings up the thought of having a set of home brew design rules
(and an HB-DRC lol). I have my own informal set in my head but would
love to hear what others use.

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

> I noticed on the end of the mail that it went off topic,
> even off topic of the already a bit off topic discussion.
> If you still want to read it go on...
>
>
>
> >
> > Check this case in unnecessary design:
> >
> > http://www.rigelcorp.com/r535j.htm
> > http://www.rigelcorp.com/__doc/8051/R535JASSM.pdf
> >
>
>
> Especially for a prototype / evaluation board 2 layer is a great

advantage.

> you can follow the tracs and you also can make changes as needed.
> i would like to see how you make changes in a 4 layer board.
>
>
> You can take apart whatever consumer equipment you like, 2 layer

boards are

> still very popular, as are 1 layer boards.
>
> I do not know how this will develop, the demanding boards have more

layers

> now,
> but also the things get more and more integrated.
> If you look how many layers they use for IC internal

interconnection (14

> or so??)
> you can maybe guess how the boards will develop. On the other hand,

as the

> integrated
> circuits get more and more functionality, requiring nearly no

external

> components
> the external circuit should get simpler, allowing less layers there.
>
> The requirements are so different - some things need to be as small

as

> possible, some
> as cheap as possible, some both, some high quality, some easy to
> manufacture.
> the demands are very different, and i think the solutions to meet

them

> will be as different
> in the future.
>
> Development is not stopping now, that is the only thing you can be

sure of.

>
> I think the trend will be towards integration, but the human is not

built

> to interface silicone chips, it needs to use the fingers to operate

the

> stuff
> and needs audio and video feedback. this human interface is (in

this way)

> impossible
> to integrate into anything (on chip buttons and such ;-).
> maybe the soultion to that is that the human carrys a
> "Human interface device" with him all time, and the machines we

build have

> no direct
> interface at all, and communicate wireless with the HID of the

individual.

>
> This has certain advantages, but also a lot of disadvantages.
> I'm known around here for my skills with electrics/electronics so

if there

> is some problem
> people call me to help them. I regualrly come somewhere only to

discover

> there is no fault, but it
> is just a user "inability". some equipment is simply too

complicated for

> some people, especially
> elderly people with disabilities like bad sight and such. If the TV

or sat

> receiver is operated through
> a complicated menu and they simply can't remember how this works,

or if

> the buttons on the remote are simply
> to small to read the lable (this happens) then i really question

the

> purpose.
> I mean, back a few years each tv had a row of buttons, and that's

it. just

> press and look.
>
> Now, if you follow the thought of the "universal human interface"

this

> could lead to two consequences:
> a) everything gets far too complicated for some. this limits the

market -

> BAD.
> b) you can program it, also in a "simple" mode. so that the stuff

works

> with maybe limited functionality,
> but WORKS. you could design human interfaces even for people with

severe

> disabilities if you like.
>
>
> Sure, the society changes, gets used to the technology, learns to

keep up

> with it, but everyone?
> and with each and every machine?
> It happens not seldom that one of my professors, on a technical
> university, (teaching electrotechnics or much more advanced stuff)

can not

> manage to make the laptop
> computer operate with the video beamer and switch all things on.

Not that

> he is too stupid for it, of course
> not, if he has some time and reads the manual it would work easily,

but

> there and then there
> is no manual, and no time to read it. He wants nothing else than a

big

> green ON button and a big red OFF button.
> he has no interest in all else of the functionality.
>
> On the other hand, if you know what you do exactly, and the machine

tells

> you "you can't do that now"
> you are also upset (like formatting your system partition on the

PC).

>
> I could go on and on, but i think everyone who kept up reading so

far got

> my point.
>
> There will still be much development going on, no doubt, faster,

cheaper,

> more functionality.
> But we also need to keep in mind that the limited human needs to

operate

> and use the stuff. Not the designer,
> which knows every function and error, the simple consumer who wants

not to

> waste much time with the manual.
>
>
>
> THE END ;-)
>
> Didn't read it after writing, there may be much errors.. you can

keep them

> i have plenty.
> look forward to discussion and possible solutions.
>
> ST

On Thu, 04 Mar 2004 17:52:42 -0000, Phil <phil1960us@...> wrote:

> to continue on this thread. I've found that if I lay out my board
> and then look at how the traces run, I can make a few tweaks and
> significantly reduce routing complexity. For example, I put an LCD
> display header on a board I'm doing now. after first routing, I
> noticed that a couple of the traces were going all over the place. I
> went back and switched 2 of the control lines going imto the micro-
> controller and rerouted. wow, what a difference, the traces were
> direct and I was able to get rid of 8 vias (4 top side wires). Its
> still 2 sided but, my design is now much more manufacturable in my
> little "pcb factory".

there are 2 things:

A) placement of parts

B) exchangeabe pins

Orcad does support exchanging of pins and also of complete function blocks
(like in a logic IC).
This is very handy.

>
> Another thing is that routers are pretty dumb.

I nearly never use a autorouter. only for placement.
sometimes i place the components like i think it is best, then i let
the autorouter work and see where the problems are. then i move parts
around
to resolve most.
but i always unroute the board and do it myself. the autorouter results i
get are
simply unuseable.
Only if i know a particular board is "strictly one time only proof of
concept", just
switched on once and then dumped i will autoroute it completely and then
never look at it
again.

Autorouter just doesn't match up with my "routing is an art" opinion.
And the result also doesn't match up with what is my minimum demand.

ST