11 Sep
2012
11 Sep
'12
7:29 p.m.
are you supposed to be able to tell just by looking at them if they have multiple layers?
Assuming you have one w/multiple layers, and I guess I already know the answer, it's a
huge headache to try and draw up schematics (or duplicate the artwork rather) w/only
photographs of the top and bottom of the board? You'd need knowledge of the ic's,
which isn't that huge a deal perhaps, being I'm questioning this w/regard to very
standard ic's and glue.
?I want my APC back. I need to create copies of the cpu and disk controller cards. I
don't have ready access to these. Each of these boards is nearly the size of a IBM
5150 mobo, so would it make sense that they are anything but single layer boards? And the
APC had 2 video boards, each w/it's own pd7220.
?I got to find that yogi guy.
11 Sep
11 Sep
8:26 p.m.
Generally, the tip off is power and ground planes vs. bigger traces on the back or front.
Are decoupling caps connected to vias or to power lines?
Or, especially on older stuff, there will be a layer "key" on a board edge.
-Ryan
On Sep 11, 2012, at 2:29 PM, Chris Tofu wrote:
are you supposed to be able to tell just by looking at
them if they have multiple layers? Assuming you have one w/multiple layers, and I guess I
already know the answer, it's a huge headache to try and draw up schematics (or
duplicate the artwork rather) w/only photographs of the top and bottom of the board?
You'd need knowledge of the ic's, which isn't that huge a deal perhaps, being
I'm questioning this w/regard to very standard ic's and glue.
I want my APC back. I need to create copies of the cpu and disk controller cards. I
don't have ready access to these. Each of these boards is nearly the size of a IBM
5150 mobo, so would it make sense that they are anything but single layer boards? And the
APC had 2 video boards, each w/it's own pd7220.
I got to find that yogi guy.
8:31 p.m.
On Sep 11, 2012, at 3:29 PM, Chris Tofu wrote:
are you supposed to be able to tell just by looking at
them if they have multiple layers? Assuming you have one w/multiple layers, and I guess I
already know the answer, it's a huge headache to try and draw up schematics (or
duplicate the artwork rather) w/only photographs of the top and bottom of the board?
You'd need knowledge of the ic's, which isn't that huge a deal perhaps, being
I'm questioning this w/regard to very standard ic's and glue.
It is nearly impossible with photos of just the top and bottom of
the board. It's hard enough with X-rays. Visual inspection of the
board (if you're lucky and they're not buried under planes, you can
see the shadows of the buried traces) and buzzing out with a meter
(assuming the normal caveats and pitfalls of buzzing out) are going
to be a better bet.
Of course, I'm assuming by "multiple" you mean "more than 2"
rather
than "more than 1". 2-layer boards are generally doable with just
photos if there aren't gobs of lines running beneath ICs (and
especially tightly-packed ICs). You'd still want the board to do
continuity checks because you will invariably have lines running
beneath ICs which you will need to sanity-check; the most pesky
of those are the ones which actually terminate at a pin on the
IC under which they run.
I want my APC back. I need to create copies of the
cpu and disk controller cards. I don't have ready access to these. Each of these
boards is nearly the size of a IBM 5150 mobo, so would it make sense that they are
anything but single layer boards? And the APC had 2 video boards, each w/it's own
pd7220.
I'm assuming something like that is at least a 2-layer board, but
anything is fair game. If it's a 4-layer board, I would expect a
significant portion of the inner layers to be power and ground
planes.
- Dave
8:36 p.m.
On 11 Sep 2012 at 12:29, Chris Tofu wrote:
are you supposed to be able to tell just by looking at
them if they
have multiple layers? Assuming you have one w/multiple layers, and I
guess I already know the answer,
Hold the PCB up to a very strong light--many if not most multi-layer
PCBs have internal power or ground planes. This makes the PCBs
pretty much opaque--you can't see light shining through where the top
and bottom foil has been edched away. Vias that seeming disappear
into the PCB (i.e. no via visiable on the other side) are a pretty
good hint also.
--Chuck
9:04 p.m.
On Sep 11, 2012, at 4:36 PM, Chuck Guzis wrote:
On 11 Sep 2012 at 12:29, Chris Tofu wrote:
Or, more likely, that go in one side and the opposite side has no
connection. Blind vias have always been pretty expensive, and I
can't recall having seen them on any vintage machines (though I'm
sure there are plenty).
- Dave
are you supposed to be able to tell just by
looking at them if they
have multiple layers? Assuming you have one w/multiple layers, and I
guess I already know the answer,
Hold the PCB up to a very strong light--many if not most multi-layer
PCBs have internal power or ground planes. This makes the PCBs
pretty much opaque--you can't see light shining through where the top
and bottom foil has been edched away. Vias that seeming disappear
into the PCB (i.e. no via visiable on the other side) are a pretty
good hint also.
9:21 p.m.
I'm getting the impression that examining the visible (top and bottom) circuit traces
might be a better indication of what's "missing". If many chips lack +/-
connections, that would indicate there are hidden ground planes.
?If my buddy only got me schematics, I'm sure I could figure out the rest. I don't
have any skinny circuit board material that's as large as those boards I don't
think. But regardless the choice of an adhesive isn't all that critical, right? *ducks
for cover*
9:12 p.m.
________________________________
From: Chuck Guzis <cclist at sydex.com>
Hold the PCB up to a very strong light--many if not most multi-layer
PCBs have internal power or ground planes.? This makes the PCBs
pretty much opaque--you can't see light shining through where the top
and bottom foil has been edched away.? Vias that seeming disappear
into the PCB (i.e. no via visiable on the other side) are a pretty
good hint also.
--Chuck
C: Thanks for the the input from you, Ryan, and Dave?
?My friend Yogi is going to scope this situation out for me. As I said, he seemingly has
complete sets/s of APC documentation. If I could only convince more people that digital
camera "scans" of vintage documentation are more then adequate, we might have a
lot of this stuff archived already. But then idiots like me allow disk media to sit around
too long and in the wrong storage conditions that it negates any benefit there is to
having the documentation. What a shame.
12 Sep
12 Sep
7:08 p.m.
are you supposed to be able to tell just by looking at them if they have mu=
Not always, but there are often a lot of clues. The first is to hold the
boad up to the light. Often there is s digit on each layer -- inclduign
the outside layers -- ans the other layers are clear in that area. So by
holding it up to the ligth you can see all the digits, oen per layer.
That was origianlly done for that reason, of course, so that they could
check that no layer had been left out when the board was made. The second
trick is to look for things that must 'go somewhere, but don't appear to.
Vias that have traces on one side of the board and not the other. Pins on
ICs (particularly power and ground pins) that don;'t have traces going to
them. Things like that.
ltiple layers? Assuming you have one w/multiple
layers, and I guess I alrea=
dy know the answer, it's a huge headache to try and draw up schematics (or =
I'ev traced out schematics for many multi-layer boards. I don't find it
much worse than palin double-sided boards. What I do is :
Usolder anything that will test as a short. This means any switches that
can't be turned fully off (like anything with a changeover contact),
relays (for the same reason), most inducators and transformera, etc. Then
I manke a list of the main components o nthe board -- and their
subsections. I might have :
U1 a b c d 74LS00
U2 Z180
U3 27C128
U4 a b c d e f 74LS04
...
K1 Coil a b (that's a relay with 2 sets of contacts)
And so on.
Then I try to get some idea of the fucntionality of the board. If there's
a processor I strart by finding the reset and clock circuits, and draw
those out. As I draw a sectoion of a component, I corss it off the list
(that saves me from drawing otu the same bit twice !). I woudl then draw
out the processor chip, thus identifying the buses. THen the mmeory
devices, and sort out the address decoding (or maybe the other way round,
you can trace back fro mteh chip select pins on SRAMs and EPROMs before
you've drawn tohsoe devices). And so on. To trace the conneftions I use a
good continuity tester. 'Good' meaning it won't bee fooled by a
forward-biased diode junction, and that it's quick to respond. THat way I
can put the proble on one IC pin and stroke the other probe round every
other conencituon if I have to.
Yes, it takes time, but it's not impossible.
Some people probalbgy use a computer to keep track of this. I find a pen
and paper works perfectly well.
duplicate the artwork rather) w/only photographs of
the top and bottom of t=
If you don't ahve the board itself, you have a _lot_ of work to do. The
outside layers are a start (is the picture of the component side a bare
board, or do you have traces disappearing under components? The latter is
alot worse You can probagly work out some of the cirucity form the traces
you can see (althoguh I have seen at least one PCB where the outside
layers were the power and ground palnes, with almopst no visible signal
connections), but you won't get everything. If you don't ahve the
scheamtics, you will have to guess, and that's always dangerous....
-tony
4491
days inactive
4492
days old
7 comments
5 participants
participants (5)
-
ard＠p850ug1.demon.co.uk -
cclist＠sydex.com -
fraveydank＠gmail.com -
rampaginggreenhulk＠yahoo.com -
ryan＠hack.net