For giving a fellow paraglider a lift back to the top
Have another beer, Lester.
Also, nice work on the simulation/testing.
Watch Video The Low Orbit Helium Assisted Navigator (LOHAN) team is still crunching numbers from last Friday's test run of the Vulture 2 spaceplane, which came at the end of an intensive week of brain surgery on the aircraft's Pixhawk autopilot. Hats off to Linus Penzlien and Andrew Tridgell for crafting and installing the …
Supercaps are not really capable of delivering large currents; they're meant as a buffer for stuff that draws little current (like RTCs and CMOS storage) for which you might not want to use a primary button cell (because of replacement issues) or rechargeable cells (additional charge circuitry, degradation).
Depending on the current drawn by the Pixhawk, slapping a reasonably large cap on its logic power supply line together with a series diode, so that when the servos cause the battery voltage to drop it won't propagate to the Pixhawk, would be my try to fix this.
errrr... supercaps ARE capable of delivering HUGE currents actually. You're thinking of conventional capacitors. A bank of 6 maxwell supercaps (58F, 16V) can have 170A drawn from it at max load, which is pretty darn big. It won't provide it for long, but this type of application is exactly what they're good at.
Well known problem in big metal ships. Apparently they used to sort it with some strategically placed magnets (or a big lump of steel), though I couldn't tell you the specifics. Probably easier to test that by just having a play around with an energised LOHAN in a big open field, followed by a beer or two for refreshment.
Well known problem in big metal ships. Apparently they used to sort it with some strategically placed magnets (or a big lump of steel)
Well known problem on airplanes as well. Even on fiberglass gliders with little to no metal components a compass will be noticably out of alignment if not corrected. (And TINY amounts of metal can already influence it.) A sunroof and aluminium roofrack are not going to cut it, there's still way to much metal in close proximity.
On aircraft the compass has a small tray for mounting small magnets or iron blocks, used to counteract the influence of the aircraft frame on the indication.
I doubt the LOHAN compass will have any sort of adjustable compensation built in. It would still be a good idea to slug it up a hill, set it on a non magnetic table (Preferably something you coble together from some old planks without using ANY nails or screws) pointing due north, power everything up and stand well back while noting what the compass reads. Then turning it 90 degrees, rinse, repeat.
On the video is also seems to me like the canard has some trouble returning to neutral after making a full deflection. Is this just an illusion or is the servo drawing much more power at that point?
Unfortunately it may be a little difficult in practice to get the Playmonaut to make the necessary adjustments on board while comparing readings with the vehicle based compass. The upside is that if an attempt to do so succeeds, venture capitalists will be falling over themselves to fund the development of general purpose micro-robots.
Looking at those spikes, wouldn't a fat capacitor across the servo battery also help? Cells have a higher internal resistance than capacitors: a 10000uF 10V electrolytic has an esr of 0.038 ohms at -10C (reducing as temperature rises). AA Lithium cells tend to have a maximum safe discharge of the order of hundreds of mA, which suggests to me that the internal resistance is fairly high.
If it was *my* project, I'd be looking more carefully into why that first servo smoked, and what was going on to cause those current spikes on the servo power rail.
Clamping zeners, supercaps, extra batteries are just band-aids. You need to get to the root of that problem, lest it resurface at flight time. If you have a second set of servos and another power unit, perhaps a bench test with a scope on the power rail is in order? Something's not right there, and it needs to be gotten to the bottom of.
Maybe a clue here?
Not clear if you're powering servos from the 3DM power module, but this page explicitly says not to:
I don't, but the design guys do. Suffice it to say, a test at 100km/h close to the ground is going to give the surfaces a much harder time that at altitude, for the same velocity. Added to that, the airflow over the canards is much more brutal when the plane's strapped down on the top of a van, so we're sure they can stand the heat.
Be aware, a BEC (battery eliminator circuit) is not designed to regulate output to the servos, but rather to the (much lower current) receiver in an r/c system. It eliminates the need for a separate receiver battery.
And the current spikes are almost certainly due to excessive friction in the canard pivot system. Miniature ball bearings may be a solution. Perhaps a pair each side supporting each canard shaft.
Great project - keep it up......
A BEC is indeed for supplying both the receiver and the servos, but the current rating has to be sufficient for the load. 2 amp for 2 or 3 standard or micro servos, 3 amp for 4 servos, etc. The main issue with a BEC is the voltage drop across it, typically they use a TO92 package voltage regulator requiring an input voltage several volts above the regulated output, this may cause the same issue due to servo load spikes causing it to drop out :(
I favor the existing setup with something like a 4700 uf capacitor connected across the the battery supply but close to the load, the issue is spikes rather than the battery capability.
Spending £1 on a capacitor and another test drive would seem to be the way to go :)
Biting the hand that feeds IT © 1998–2019