A few pics of cpu bonding 1975

[tsfarley]

I'm playing with boards from an HP9835 era desktop from the 70's.
pics are at about 150x magnification on a simple school grade microscope. I think I'll need a 10x or 100x more magnification to see the thickness of gold.

 

[GoIdman]

That is amazing, maybe some serial number? It looks fantastic.

Pete

 

[Martijn]

Makes sense. 1975 was a very good year.

Nice to see. Grainy gold on all traces. Will it become a display piece?

 

[tsfarley]

Makes sense. 1975 was a very good year.

Nice to see. Grainy gold on all traces. Will it become a display piece?

very likely yes, my daughter is an artist and wants to make something out of it.
I'm up tp 13lbs of gold traces, pins, fingers, etc. , so saving these.
I was experimenting how strong a microscope I needed to see the thickness of gold vs modern techniques of gold pcb plating. 150x doesn't come close. a few shots at 600x needs a ridiculous amount of light to see any detail likely due to the type of biology microscope I picked up for $25. lol
the mic came with a light below the specimen. and I added about 2500 lumens from above with an led flashlight a 1/2 inch away.

At 600x, I see detail, but not sure what I'm looking at. see examples.
I've read a little bit about finger specs, requiring a bevel, maybe thats what this interface is between the 2 colors?
thanks for the feedback.
Tim

 

[BShan]

I think that the way these are generally made, the entire surface of the PCB was one uniformly thick layer of copper. Areas intended to be traces is masked off, and acid dissolves the unmasked parts. This step may be what is producing that texture on the sides of the finger trace. Instead of a sanded/smooth surface, it's an etched surface.


I think the bevel you mention is much larger, and goes along the edge of the fingers, such that when you are inserting the card into its connector, the row of fingers can "wedge" its way between the springy contacts of the connector.

 

[macfixer01]

Some of the old HP equipment has very interesting hybrid IC’s. Here are a couple from a 5180A and a 3497A which I scrapped recently. Underneath the ceramic caps they’ll have multiple silicon dies and gold bond wires, but I don’t want to break them open until and if they get processed. Those brown ones have a really strange connection method, the pins are attached to the board and the chip just pushes down over them and is held in place on the standoffs.

 

[macfixer01]

I think that the way these are generally made, the entire surface of the PCB was one uniformly thick layer of copper. Areas intended to be traces is masked off, and acid dissolves the unmasked parts. This step may be what is producing that texture on the sides of the finger trace. Instead of a sanded/smooth surface, it's an etched surface.


I think the bevel you mention is much larger, and goes along the edge of the fingers, such that when you are inserting the card into its connector, the row of fingers can "wedge" its way between the springy contacts of the connector.

It’s really unclear to me what area of the fingers he’s focusing on in his closeup photos? Is it showing the middle area along where the connector pins have been rubbing off gold? Or is this along the outer edges of the fingers themselves? I can’t see a similar color variation in the non-close-up photos, except for the worn areas.

I believe the bevel he’s referring to is on the leading edge where it would be inserted into the slot connector. It would help prevent any tendency for the square edge to catch and peel, or for the sharp edge to cause undo wear to the slot connector pins during insertion.

I was just mentioning on Facebook last night, there’s an old 1969 promo video from Tektronix that details how they make their circuit boards. The way they did it (at least back then) was to pre-drill the holes, then do a light electroless plating of the holes and surfaces to establish conductivity, then do an electrolytic plating of copper to get a more robust coating within the holes and their connection to the surfaces. Afterward they masked everything except the traces and pads then gold plated all the trace and feedthrough hole areas. Lastly they removed the masking, then did a conventional etching to remove the excess copper but using the gold plating as an etch resist. I don’t know if other companies like HP who often have gold coated traces were doing the same, or if anybody is doing it this way anymore exactly? Obviously if the traces were not being gold plated then it couldn’t be done this way, and would require an application of each resist only where the traces and pads were to remain, with the background areas unprotected.

 

[BShan]

Regarding the closeups, I did take a bit of liberty in assuming that the piece was oriented the same and just zoomed in.

I've seen some RAM boards with gold under the solder mask, so the process you describe from Tektronix is definitely possible, and probably closer to the actual process used to make tsfarley's board. I'm not sure if I've ever seen a RAM board that didn't have through-hole vias. I'll search for that old tektronix video you mention, it sounds interesting!

I'm familiar only with the basic process due to DIYing circuit boards back in the day, where you start with a fiberglass board with copper cladding on both sides, and etch it away after applying a resistant coating to selective areas.