From Hackaday:
Hackaday alum [Adam Munich] shot a tutorial video on using a rate gyroscope.
Here he’s showing off the really fancy piece of ancient (technologically speaking) hardware. It would have set you back about fifteen grand in the 1960′s (inflation adjusted) but can be had these days for around $30. What a deal! These are not small, or power efficient when compared to the components that go into smart phones or gaming controllers, but they’re a heck of a lot more accurate than the ubiquitous modern parts. That’s because a rate gyroscope — which is the gold cylinder on the left — actually incorporates a spinning motor and a way to monitor how it is affected by changes in gravity. The driver/interface circuitry for this gets hairy relatively fast, but [Adam] does a solid job of breaking down the concept into smaller parts that are easy to manage.
Wondering what is different about this compared to a MEMS accelerometer? We know they’re really not the same thing at all, but wanted a chance to mention [The Engineer Guy's]video on how those parts are made.
Comments
BTW: That bulky thing is only for one axis?
Mark, that's fascinating stuff (I learned a lot about the B-58 today too). I also didn't know that was the approach to the redundancy but it makes sense. I've heard that most small aircraft pilots don't bother using VOR even though they know how because they've got GPS and it's just that good, but they can still fall back if required.
I guess for me looking at a simple arduino or IMU, even though it has an amazing amount of computing power compared to to systems back then, there is just something tantalizing about mechanical systems.
Josh, they were very high maintenance birds, like all military craft. I don't remember the exact MTBF numbers but our avionics shop had over 150 guys taking care of about 200 aircraft, F102s, F105s, F4s, etc. They make modern airliners seem maintenance free.
There was little redundancy built into individual systems, and the systems weren't integrated like today. Redundancy was accomplished by using different systems. So if the autopilot heading function failed the pilot could use the radio/radar navigation (no GPS of course), or even fall back on the compass! The systems could fly the plane using servos to control the hydraulic actuators, however all those could be bypassed by physical pilot input on the stick and rudders. Fly by wire was still on the drawing boards or more likely in engineers dreams!
The B-58, being somewhat digital, did have some redundant sensors. The separate yaw dampening system, installed in 1969-1970 was one of the first truly redundant systems. It had triple redundant sensors and electronics all the way through until the final yaw dampening flight controller which would monitor and choose the middle signal (at about a 5K hz sampling rate). If one signal was more than 15% different than the other two it was thrown out. The entire new yaw dampening system was installed for one purpose, to save the craft if one of the 4 massive engines failed on takeoff. A human could not respond fast enough to save the plane, but of course the new electronic system could. Not good to have a massive delta wing bomber with a live nuclear payload crash into the local corn field!
Jack, didn't people already build IMU's out of "vintage" gyros? How about the autopilot system in a V-1 flying bomb? Talk about analog! It was powered by compressed air.
Jack, speaking of IMU and unusual gyros. I would like to see how well a IMU using ultra high end ring laser gyros would work. Those ring laser gyros has very little drift compared to MEMS, so if you sampled rotational data accurate enough, you could theoretically keep the copter level using gyros only for a long time.
Mark that's very cool. What was the MTBF on most parts? Did you have to build in a lot of redundancy because of the mechanical nature of things?
Someday, someone is finally going to build an IMU out of vintage gyros. Maybe a future generation will make a quad copter using a mechanical IMU, just like all the tube amplifiers. It would be an interesting test of which was more accurate.
I was an autopilot specialist on USAF delta wing aircraft (B-58, F102, F106) in the 1960's. Not only were all the gyros, accelerometers (rate gyros), etc. mostly mechanical (moving parts), most of the system was entirely analog. The F102 had a totally analog flight data "computer" complete with many little "black boxes" which "computed" via gears moving pots to vary the analog signal levels. Servos were similar to today in concept (mechanical PID settings) but again worked on voltage comparators and mechanical feedback. I always get a kick realizing at my APM 2.5 has more basic ability to control an aircraft than those very expensive, maintenance heavy systems back then. That said, it is still interesting to note those systems (when working correctly..) flew those supersonic birds rock steady, could lock on to a target 40 miles away, compute an intercept point, fly the aircraft to the point and give the pilot a target lock and indicate exactly when to launch an air to air missile. Plus, at least the F102 and F106, did it without one digital signal in the entire aircraft! The B58 was one of the first digital systems, but was originally designed with discrete components and no software.
Today, even though I was a part of that world, find it hard to believe how accomplished the electronic engineering designed was for it's time. Forgive my ramble down analog lane, I'm pulling myself back and will now launch this missive off in a very digital manner!