Creating a Circuit Diagram from a PC Board

Then again a decent light underneath usually shows the traces.

A flatbed scanner instead of a camera might work better, consistent scaling.

But not where they go behind parts.

Of course, that's still a problem for double sided and more of a problem for multilayer PCBs.

But I like his general approach. :)

In fact, I may go and try it as I have a single sided PCB to trace at the very moment.

--- sam | Sci.Electronics.Repair FAQ:

Important: Anything sent to the email address in the message header above is ignored unless my full name AND either lasers or electronics is included in the subject line. Or, you can contact me via the Feedback Form in the FAQs.

It will work great for the traces but may have problems depending on how tall the tallest part is. Though, before I had a digital camera, all the photographs of circuits boards in the FAQs were done on ab HP scanner.

--- sam | Sci.Electronics.Repair FAQ:

Important: Anything sent to the email address in the message header above is ignored unless my full name AND either lasers or electronics is included in the subject line. Or, you can contact me via the Feedback Form in the FAQs.

"Nelson" schreef in bericht news: snipped-for-privacy@news.giganews.com...

Well... It's a way. I used a scanner to make the pictures. As most of the times I cannot remove the components, I keep a mirrored full color picture of the solder side next to the real thing. Of course, things became more difficult over time as PCBs tend to be soldered on all sides these days.

petrus bitbyter

snipped-for-privacy@hotmail.com hath wroth:

It won't. A scanner has a very narrow depth of field. If there is anything between the scanner glass plate and the board, it will cause the board to be out of focus. I've tried it and it doesn't work.

I use a digital camera to take photos of both sides of a board. I then use a photo editor to flip over, convert to a transparency, and eventually create a composite image. If I'm ambitious, I play with the colors sufficiently to remove the uniform board and trace colors, leaving only the components. Then, it's easy to just connect the dots.

If I'm really ambitious (or have lots of time), I import the component layout BMP into a CAD or PCB layout program, overlay the components with component symbols, overlay the connect the dot traces with rubber bandable lines, and then drag the components into something resembling a schematic. I recently did this with the glow plug control box for my diesel truck, which was very simple. I don't recommend this method on anything complicated.

A few years ago, I inherited a large pile of boards without documentation. I wanted to get them working and had to generate a schematic. It was tough because of the large number of SMT parts. So, I sacrificed one board, used a propane torch to remove the components (on both sides), and ended up with a bare board. Much easier than trying to trace the board with the components in place.

Incidentally, there are companies on the internet that advertise circuit board reverse engineering services. If this is a complex board, you might want to get a quotation before attacking. For example:

Here are some links that show a sample board I did:

It's overkill for this board, I know, but it is a good example of the technique. You can see how it wouldn't work with a scanner because of the height of the components, particularly the heat sink. As it was, I had to prop up one end when photographing the traces to keep it level.

Isn't there a problem with static on a flatbed?

Why should there be a problem ?

Graham

In article , snipped-for-privacy@bigpong.com (known to some as Rheilly Phoull) scribed...

At least until you're confronted with a four (or more) layer board. ;-)

Keep the peace(es).

The typical flatbed scanner just uses a bright fluorscent lamp so shouldn't have any effect on erasable programmable parts.

I used to use an HP scanner for all the photos of reasonably planar circuit boards in the FAQs and still do. The resolution and perfect geometry are amazing for such inexpensive technology!

--- sam | Sci.Electronics.Repair FAQ:

Important: Anything sent to the email address in the message header above is ignored unless my full name AND either lasers or electronics is included in the subject line. Or, you can contact me via the Feedback Form in the FAQs.

Jim Garland had this same issue -- needing to create a circuit diagram from a PC board. He used CIRCAD - for the HRO-600 project.

His challenge was the very rare National HRO-600 receiver.

The details of his extensive investigations, repair and restoration (including finding a long-discontinued Motorola IC of the "MECL II" family for the repair)

On Sat, 27 May 2006 10:26:36 -0400, g. beat wrote (in article ):

Thanks for the pointer. Must have been a labor of love :-) CIRCAD looks like an interesting program but unfortunately I have a Mac and don't think I'll be laying out $1,000 for a CAD program anytime soon :-)

It's not clear to me from Jim's description (or from CIRCAD's web site) if you can overlay the images of the front and back of the board. Having them side by side would definitely help, but being able to superimpose just the traces on the components would seem to make it a lot easier (look at my sample photos)... at least for me. Perhaps CIRCAD can do that... I can't tell.

Hamad bin Turki Salami hath wroth:

No. Methinks you're thinking of a non-flatbed scanner such as one of the multi-malfunction desktop printer/scanner/fax/modem/shredder combinations. The ones with rubber rollers or sheet feeders might create a static charge, but always have brushes at the paper exit slots to remove the charge. Besides, I wouldn't want to run a populated circuit board between the rollers anyway. Crunch.

Drivel: Long ago, one of my products was in the prototype stage when marketeering decided to stop everything so that they could take some pictures. The boards were carefully sent to a photographer, who took numerous photographs under the bright studio lights. When I got the boards back, they were all dead. All the (2716) EPROM's had been partially erased. I don't think a flat bed scanner will do the same thing, but I wouldn't want to find out.

Going further afield ... this reminds me of the experience of a composer I knew (actually well known, but I won't identify him). He landed a performance of an extremely difficult opera he wrote. The technical difficulty limited the possible performers to a very rarified set of companies, and getting any performance of a modern opera at all is a major coup for a composer. Of course, he had a recording done one of the nights, and the performance was brilliant, beyond his hopes. The next week, the engineer had the master tapes in his studio. While mucking around for something, he placed a telephone on top of the tapes. As fate would have it, that moment the telephone rang. It was one of those old telephones with a big electromagnet driving the clapper of a bell ... and that was the end of the recording of that brilliant performance.

I just recently went through this myself. Use the scanner to image the foil side of the board, then mirror the image to make an x-ray view from the component side. (My print dialog has this as an option.) Then using a red pen, I draw the component symbols between the trace pads. Identify the DC supply and ground traces, and use various color highlighters on them. When done, it's fairly straightforward to draw the schematic of the circuit.

I've had moderate success scanning the component side of the board, too. Seems many scanners have fairly good focus at 1-2" from the glass.

Neil

is

in the