On the weekend I did a 425km flight with my Vigilant C1.
Setting a new distance record is something I’ve been slowly working towards ever since flying 301km with the Cyclops EPO plane in 2014
John Smith and Gene Robinson of RP Flight Systems supported me by providing the Vigilant C1 airframe. It has a 3m high aspect wing and loads fuselage space making it ideal for long endurance missions.
In terms of flight testing the biggest performance gain was from tuning the powertrain setup. I tested two options, a 4S battery vs a 6S battery with the stock 300kVa motor. The 6S with stock motor won hands down with a performance difference of around 25%.
I also tested half a dozen different propellers, a couple of APM parameter settings, three different flying patterns and different battery payload capacities.
I tested each parameter by flying repeated circuits of an octagonal test lap and analysing the telemetry log files to work out the power consumption. I could calculate the watts used and the distance covered to work out the best settings. I could extrapolate those figures to see how far I could fly if I used the full battery load.
One dilemma I faced was concerning the flaps. With such an efficient airframe heavily loaded I wanted flaps to control the airspeed and glideslope for landing. But I didn’t want to power the flap servos and carry the weight for the whole flight when I only needed them for 30 seconds at the very start and end of the flight. In testing I gradually reduced the flaps each landing and was relieved that I could land easily with a long low approach. I did however have trouble taking off. I used a bungee that dragged the plane along the ground. Even though I achieve plenty of speed, the fact it was on the ground meant I couldn’t rotate to pitch up. With flaps I had enough lift to take off every time but without them I couldn’t. It took three attempts without flaps the first time to get airborne and even then it only happened because I hit a bump that bounced me up off the ground. I decided to remove the flap servos and glued the flaps in place and I made a ramp out of some plywood sheets and used that for the first time for the record attempt.
With the 6S setup I used a HobbyKing power module and the figures it gave in testing seemed too good to be true. I calibrated it against the amount of energy the charger put back in and that showed the power module had been underreporting the current used by about 25%.
I watched the weather and picked the best day I could have hoped for. It was very calm for the first half of the flight with only a light wind later in the day. I flew a large loiter circle (guided mode) with a 600m radius giving almost 2km per lap.
I tracked distance covered vs battery voltage and vs battery percentage remaining and found gave very consistent results that sat just above the 400km projection line all day.
I trusted the voltage reading more than the battery percentage remaining. I knew the Li Ion batteries would be 95% depleted when they reached 3V per cell so I planned to terminate the flight at that point.
All in all it was a big day, 425km of flying taking 7 hours, 48 minutes.
What I’m most excited about is the possibilities this milestone opens up. With an airframe like this the range limitations for tasks such as pipeline monitoring, mapping and search and rescue have just been moved to another dimension altogether.
Setup Details:
- Plane Vigilant C1 V tail – fibreglass and carbon fibre construction with 3m wingspan.
- Stock Vigilant C1 motor – 300kVa
- Pixhawk autopilot
- Panasonic 18650B Li Ion batteries. 6S 9P = 30600mAh.
- Aeronaut CAM Power Prop 13x12
- HobbyKing telemetry module
- HobbyKing power module
- Distance covered: 425km (264miles)
- Flight duration: 7hr 48 min.
- Average groundspeed 15.3m/s (55km/hr)
- AUW = 5.7kg (12.5lbs)
- Goteck DA2311T Servo
- ZTW 65A Gecko ESC
Comments
I'd be really keen to see your post, and try your idea Andrew. Thinking about JB's comments above, the aggressiveness of the climb might not matter too much. If you were only half as aggressive as ideal (say because of battery power limitations) you would still get a long power off time relative to the power on time.
Does anyone have any ideas for programming the autopilot. I think you'd want a high TECS_SPDWEIGHT value. It wouldn't matter if you didn't follow the planned glide slope accurately, but it does matter if you glide too fast or too slow and have to use your motor. Even if you program it perfectly to start with, it could change mid flight with changing winds. In a head wind you'd hit the high points and low points too early and have to fly level until you reached them, with a tail wind you wouldn't reach them and would start the next leg without the full height difference.
Andrew this weekend would be very opportune! Thanks for the info. 10-20% is considerable and worth investigating further. We are doing some range testing this weekend with a competition in 10 days time, where we are allowed to go to 1500ft, and this might help in coaxing a little more range just by using a climb/glide flight strategy.
Am I right in thinking the range gains are a product of both running the prop in a better efficiency band on climb and using the "stored" energy contained in altitude to fly for a period without prop, therefore without the prop efficiency losses usually accumulated over that distance? (Gravity being much more "efficient" at moving the aircraft forwards than a underutilized prop)
Does your modeling include battery discharge performance, and how a higher current draw on climb impacts overall battery capacity? Roughly how aggressive is the climb?
I look forward to seeing your data. Thanks again.
Regards
@Andrew Rabbitt. I would be interested too to see your approach of the climb and glide.it seems you consider the gain comes from the better efficiency of the propeller in climb vs cruise. why would that be?
Thanks Naish
That is 7.8% more power for the same weight. That should give about 35km more range in perfect conditions, about 30km more in the conditions I experienced. 425 km + 30 = 455 km.
This great achievement got an article on UAS Vision:
http://www.uasvision.com/2016/09/15/400-km-flight-on-single-battery...
JB, the answer is "it depends". I think 10-20% extra is likely but if you're running a high powered high L/D it could be more. One thing I want to explore with my simulation is the sensitivity to L/D. I'll see if I can get my stuff together and post something over the weekend.
@ Andrew Rabbit
I'm very interested in any information for climb and glide long range flight. Do you have a rough indication of what sort of range improvement one can expect by flying like this, and roughly the climb pitch that should be used? It would be good to get an idea of how much better it is and then try a flight with that as a guide. (It sounds like how you should drive a car for efficiency - accelerate and coast) I'm wondering how similar the dynamics are. Thx.
-
BTW somebody should give moglos a 400km badge! He deserves it! ;-)
Here is the official manufacturer of the GEB batteries Model N: GEB8043125
Also we did some testing on it, and you can see few plotted graphs that Tridge did here:
http://uav.tridgell.net/GEB-Batteries/discharge/voltage-capacity.png
http://uav.tridgell.net/GEB-Batteries/discharge/current.png
http://uav.tridgell.net/GEB-Batteries/discharge/voltage.png
These batteries work really well at less than 1C.
Hobbyking now sells them too :
http://www.hobbyking.com/hobbyking/store/__55804__Multistar_High_Ca...
Thanks for all the support, and suggestions. I'm keen to try a few of them. I actually didn't test half the things I would have liked to. Some because I didn't have time, others because they were too hard, and of course lots that I didn't think of - thanks for the suggestions.
@ Trung - I thought about releasing/throwing it from above my head, or using a pvc tube ramp, but I worried it would hit the ground before it reached speed.
@ Alex - a mobuis or gopro could probably to take timelapse stills for that long. I actually had a ground based mobius that I planned to make a timelapse from, but when I got home it had just recorded audio. Wrong settings I guess.
@ frederic - I did a bit of thinking about that myself. The total force required to change direction around 360 degrees must be the same whether doing it in one big steady state circle or four corners (assuming the same time per lap). But one big circle would be smoother as it doesn't need to spend energy to accelerate then decelerate the plane in the roll axis to bank for the corners then level out again. I actually tested a circle vs an octagon (of the same radius) and found the circle was 11% better. That's a much bigger difference than I expected, I did a reasonable number of laps (circle n=10, octagon n=8). I also did a bigger circle lap and that was the worst (but a low sample size n=3). It was a bit windy so I'm probably seeing random error.
@ Rob - no I didn't test or research other ESCs. It is a big heavily loaded plane and I'd been using it with 4S where I think I used close to 40 A taking off. It did occur to me that with 6S I could at least save weight with a smaller ESC. What do you have in mind for a high power system that I could jettison? A rocket motor? With a good ramp/bungy launch and a very gentle climb out I think I could have kept it under 10 A. I had considered jettisoning batteries as they were used up. I thought of a pod under each wing that could be dropped when they were used up. The mechanical and electrical aspects of that might be a (fun) challenge.
@ Digital Wings - I know, I'm so glad it worked, that was the first time I tried it. I was worried it might not actually hit the ramp.
@ naish - do you have any info on "the GEB batteries"? I Google gave me a few sites, but no datasheet. I read somewhere about a new Li Ion that has a higher C rating, but the same energy density. The higher C rating might be great for multirotors, but wasn't necessary for me.
@ Andrew - we PM'd about your idea for a climb/glide strategy a while ago. I just didn't have time to give it a try. I'm keen to keep talking and give it a try.
Congratulations Moglos! Impressive effort.
Rob L makes some good points. There should be some ESC's out there that use active freewheeling that will drop the part-load power consumption a little more. I don't know if the Hobbywing one he mentions uses this.
I have done a bunch of calcs that demo a climb-glide profile is worth consideration for distance efficiency. It has some dependency on prop and glide ratio, but it doesn't seem to matter what useable prop combo you choose, it will probably be more efficient in climb than cruise. I must pull my finger out and publish this data!
Given that climbing improves thrust efficiency, an aerobatic manoeuvre to perform a 180° turn such as an immelman might be more efficient if you wanted to run straight-and-level for the most part of the flight