I don’t understand it because even when I was a kid, say mid 1960s, a PCB was the “right way” to build anything electronic
Cordwood construction was used (IIRC) for compactness. With the large components of the day, laying them out on a single layer PCB in two dimensions would result in a lot more surface area usage than going three dimensional with cordwood construction, and it essentially gave you a two layer PCB when such a thing didn’t exist (because you had a PCB top and bottom) making routing easier. The example you show would have to be at least five times bigger if laid out on a flat PCB (and would still be just as tall, due to the large inductor).
In those days too wire wrap was often better for high speed digital circuits because the nature of point to point wire wrap means you tend not to suffer so many issues with stray capacitance and cross talk between signals.
It’s easy to forget how fiddly routing a complex single layer PCB is (especially when you had to design it by drawing by hand!) when even the home electronics hobbyist can order one-off four or six layer PCBs cheaply today which makes routing a breeze (and a company making thousands of a thing, can have 4+ layer PCBs made for very little money per unit) and there are good open source schematic and PCB design tools.
Oh yes the days of 1-layer PCBs, especially the paxolin ones ones which would crack easily
BTW, it was possible, even in the 1960s, to do 2-layer PCBs. Before PTH, you would insert rivets in the vias and solder them both sides.
Good point about large components. But there are few in this autopilot. It looks like someone was consciously trying to avoid PCBs – or just building a 1950s design in ~1970.
Were they building so few? The Robin factory must have been historically quite busy to have built the French aeroclub fleet. With components being generally really cheap, one would always get 100 PCBs made, and that strips out much of the labour content, and then you can finally assemble the whole product in smaller batches, as the orders come in. Even if it will take 10 years to get through the 100, it is worth doing.
To hand-build one at a time is going to be incredibly expensive, as well as creating all kinds of other issues e.g. requiring absolutely superb documentation, so that – given that, over time, you can reasonably expect everybody who knows the product to have left/retired/died/etc – a complete monkey can pick up the docs and hand-build and test just one.
Were these autopilots used on aircraft other than the Robin? The technology looks like it has origins way back.
I had one of those Baden autopilots in my Robin: the system was driven of the intstrument vacuum pump, with a tube which ran to the underside of the rear seats, there was an on/off valve within reach of the pilot. The control mechanism under the rear seat was operated by a bellows connected to the aileron control wire (much like the electric motor type). Unfortunately, the bellows and rubber connecting tube were made of perishable material. I only recall it working once and decided to delete the whole system as parts and expertise were not available.