X-ray reverse-engineering a hybrid module from 1960s Apollo test equipment
righto.com
righto.com
I suspect this was done to ensure the device could only be plugged in one way.
This technique was forgotten by the time I worked on designing embedded systems. All kinds of devices (chips, caps, diodes, etc.) had symmetrical pinouts, so they could be plugged in the right way, or the wrong way (and would burn out as a result). At least the #1 pin on a chip had a slight depression over it. We always designed chip circuit boards so the #1 pin was always in one direction, so a visual scan of the completed board would verify the chips were all in the right way.
The board was also set up so the diodes were all pointed the same way, caps too.
Of course, the chips could still be plugged in off-by-one.
Sometimes mathematical symmetry is just a bad idea. I had an old Benz that had a mechanical fuel injection system that sat on a splined shaft. There was no key spline. If you got the spline off by one, you couldn't tell until they engine ran poorly. What a wretched design.
Boeing always used keyed splines if orientation made any difference. 4 hole bolt patterns were never symmetrical so the device could only be bolted on one, the correct, way.
It was so catastrophic to hook up hydraulics to the wrong ports that the supply and return ports used different size fittings, color codes, and even one port was left-hand and the other right-hand to prevent swapping them. The ports were placed so the tubes could not reach the wrong port.
Even then, now and then some mechanic would fashion an "adapter" in order to hook them up wrong, which is why they are also tested after assembly for correct polarity.
I'm not an EE, but wouldn't this have the potential to significantly complicate the layout of traces on the board? I assume you're talking about hand-assembled boards, and hand-drawn too?
It's sometimes fun to imagine the alternative reality where miniaturization never happened and spaceflight is forced to use low-tech navigation aids - precise clock, periscopes, reticles, slide rules, nomograms. (and communication with the flight control center which does the bulk of the computational work)
Excellent book on this topic.
Do you think there was any kind of primitive authentication of the received PSK signal? Even something as simple validating a preamble against a hard-code bit sequence?
For the range safety officer's destruct control, NASA did use cryptographic security because they wanted to make sure the Russians couldn't blow up the spacecraft. This was an entirely separate system. The control unit used a secret plug that encoded the "key", that was installed just before launch so nobody could spoof the command.
What was this? Never heard of it. Anywhere I can read about it? It was a remote controlled destruct??
[1] See https://history.nasa.gov/afj/ap10fj/pdf/19700015780_as-505-a... pages 14-15.
And yes the rockets had shape charges in each stage which can be activated to destroy the rocket in a controlled fashion. If you have a rocket gone haywire, it's better to jettison the crew capsule to safety and do a controlled demolition, to prevent the rocket from flying out of control and crashing into structures or people on the ground.
This is also why rocket launch sites tend to be coastal locations with ocean to the east - there's a lot less valuable stuff under their flight path.
A more general question - What's the hardest part about reverse engineering technology from the 1960's and such, from your experience?
Cheers from a former Avionics Technician. Keep up the good work, and love your blog.
Another hard part is things that are encased in epoxy or welded shut. In this case I used X-rays but its generally a problem.
Overall, the main secret of reverse engineering old stuff is to spend way more time on it than an sensible person would.
quote of the week.
so many things have been laid bare to me by just putting in the time. even just staring at the thing(s) will often help me realize something important, somehow.
A more serious question: why'd you use an X-ray instead of cracking one open?
As the modules are made of discrete components (as opposed to some thin film process) wouldn't it have been better to assume standard values, i.e.: 20k to either 18k or 22k, 4.9k to 4.7k, 155 to 150 ohms, etc. and work backwards from there?
Tooling up for such odd values would be a messy business, especially so for such a small quantity (and in that era when jigging up would have been a big effort).
Alternatively, standard components may have been used and a SOT (Select on Test) method employed to obtain those specific values (say as is done in 'film terms' matrices where exact values are of critical importance).
No doubt this could be determined by opening one of the modules and either measuring them directly or checking the color code bands on the resistors. It's very unlikely that a standard resistor made during that era would not have had the normal color coding and as far as I can recall Motorola never made resistors so it would have had to have purchased them from a manufacturer of same, Allen-Bradley (A-B), Sprague Electric, etc., so it's all the more reason the resistors would have been color-coded. (BTW, please correct me if Motorola did make resistors.)
I'm of the generation who remembers those events happening so I find it fascinating. I'd like to monitor how this project eventually unfolds.
Perhaps something to think about when writing code. It’s not always enough to know what it’s doing. Leave a comment with the why :)
The virtual AGC site has links to a lot of Apollo documents, with a focus on the AGC: https://www.ibiblio.org/apollo/links2.html
As far as scanned Apollo documents, archive.org has a bunch.