Please See the Next Speed Project:
Odd Flyer? Contrary to how this build looks, it is not at all awkward to fly. In fact, this is now my favorite frame to fly now that I put the camera angle at 40°!
Crashability: DO NOT build and expect this thing to survive much abuse – there are still some weak points that haven’t been addressed. I fell short trying to strengthen it on the VXT project and I actually have a better method. I hope to post the new design before I even get a chance to do it myself since I would like to see this record broken!
May 26, 2017: VX1 broke the speed record today! You can read about the details here.
Much better video (ONLY 150mph though):
- GPS Velocity Accuracy
- Test Frame
- VX Frame
- VX1 Frame Issues
- VX Frame Derivatives
- No Kits
- Motor Choice
- Bill of Materials
- Speed Record
- Nose Job and Butt Job
- Another 150+ MPH Run
I won’t go through the whole development process since the projects page gives a pretty good outline as to how I got to this point. The main feature that sets these frames apart from the rest of my projects are the arms.
The last 2 issues are finally resolved: Although I liked the fact that the C series of frames had motor mounts that were adjustable, they never fit the bill since they were still relatively weak. However, the angle adjustment wasn’t the main purpose of the motor pylons. In order to keep the frame relatively short, I had the arms as close to the battery as possible. But this meant that the motors would be too close. I used the pylons to mount the motors further forward or back (depending on location) to increase distance. Of course the best thing to do would be to have the arms coming out at an angle, but that meant the arms would no longer be a solid design (between the 2 front or 2 back motors) and strength would suffer tremendously.
The VX arms
It finally popped in my head one day a way to have the arms linked into one another which solved 2 issues:
- Motor arms could now be angled straight to the required motor locations without suffering in strength
- No more weak motor pylons
The solution was actually very simple. One arm would go through the body. The other arm would go through the body, but a section was cut out where the first arm intersects. Before the 2 pieces of the 2nd arm are put into place, a hole is drilled through the first arm and a smaller diameter support arm is put through. Once in place, the 2 pieces of the 2nd arm can be put into place. All is secured with Starbond and is exceptionally strong.
2. GPS Velocity Accuracy
After researching it, I found out that GPS is a lot more accurate than a radar gun. However, in order to get a more accurate velocity reading in Betaflight, I had to adjust one of the source files to have the gps_auto_config enable the “Airborne <2g” dynamic model mode. This is much better suited to quadcopters and testing is currently in process to see if this change will be implemented into one of the upcoming Betaflight releases.
Yes – position accuracy of the ublox M8N is only 2.5m, but the velocity accuracy is 0.05m/s. This doesn’t seem to make sense, but the velocity is calculated using Doppler calculations, not change in position (if position was used for velocity calculations, it would be a disaster). The combination of low positional accuracy and the assumption that the velocity is calculated from position data is why many people assume GPS is no good for velocity readings. However, this is contrary to the truth.
Resources on ublox M8N:
3. Test Frame
The final iteration of the C series of frames was used as a test bed for the VX1 configuration. Although flimsy, it worked well enough for a proof of concept for the body/motor configuration.
Also, this was the first time flying with the GPS set to the “airborne <2g” dynamic platform model. Results were excellent and I actually recorded a top speed of 111 mph.
Due to my lack of insight, I only posted a video of the FPV DVR footage since the Mini Mobius had a horrible view:
4. VX Frame
Although the motor arms are in an X configuration (which is important to reduce cross section), I consider this more of a Z frame since the Z frame design is not only a larger factor in reducing cross section, but also in keeping center of gravity, center of thrust, and center of pressure in alignment. Maybe the name ZX would have been a more fitting name…
The frame is constructed entirely out of carbon fiber. The main body tube is 1.75 inch I.D x 24 gauge tube (44.45mm I.D. x .70mm wall thickness). This tube on its own is relatively flexible, but once the arms are attached (with starbond and epoxy) is EXTREMELY strong and rigid. Once the openings for the ESC’s and battery bay are cut, it becomes a little flexible again, but a strip of 1.50mm x 4.50mm x 110mm carbon fiber on either side of the bay stiffens it right back up.
Features of the VX frame series:
- Strong frame – no flimsy motor mounts like the C series
- Base carbon fiber frame weighs 37g (47g including motor screws and nose/tail cones)
- 178mm frame with 5 inch props
- 5s (VX1 & VXS) and 6s (VXV) capable.
- Extremely low drag coefficient
- Extremely low profile quad frame
- Battery fits inside the body
- CG and CT aligned for max motor utilization
- Round motor arms (no thrust blocking flat arms)
- Inverted front motors to help with CG/CG alignment and reduce prop wash going to the back props
5. VX1 Frame Issues
Just wanted to note a couple issues I had with this frame:
- Camera angle – my fault for not double checking this…
- ESCs were mounted too deep – this made for a very tight fit. Not all of my lipos would fit. Regardless, it takes quite a bit of work to get the ESCs and wiring just right so everything fits in.
6. VX Frame Derivatives
VXS: Main page: VXS
I had a chance to do a little flying with this frame today – definitely fun to fly. Although I need to adjust a couple things, I was able to hit 118.6MPH.
This will be a stretch version. The stretch comes from angling the motor arms at a 30° angle (relative to the VX1 arms). This will make the optimum cross section at a pitch angle of around 60°. I am assuming this will be my favorite since it should have the best overall performance of these frames. The arm angle puts the frame size almost to the 250mm mark.
Frames in progress:
VXV: Main Page: VXV
This frame will be the speed run frame (VX1 wasn’t supposed to be). As of now, the frame is finished and can hold a 6s lipo. A few construction changes will have to be made vs the VX1, but is in general the same size except the arms will be 8mm in diameter.
VXT: Main Page: VXT
A little stronger version of the VX1 and doesn’t require a special take off platform.
7. No Kits
Selling kits looks less likely since I think this is still more of a do-it-yourself type frame… I’m not saying I wouldn’t consider maybe one day making frames and selling them, but as far as selling kits to assemble – there would be no point.
However, if anybody has any ideas or thoughts on this, let me know…
8. Motor Choice
As of now, I have 2 frames built and being outfitted with electronics. Motors are getting harder to choose, but after a lot of number crunching from motor data found at Mini Quad Test Bench, I decided upon the Cobra CP 2207/2450 motors.
The key factor that I was looking for is a motor that is efficient while still turning out a relatively high number of RPM’s since RPM’s are the largest factor in speed, not thrust. However, the Cobras are still able to throw down loads of thrust while staying relatively efficient. A testiment to this efficiency is the fact that they were able to set a new speed record using a 5s lipo (last record holder was using a 6s lipo).
9. Bill of Materials
NOTE: When buying carbon fiber, make sure it is the woven pattern, otherwise torsional stiffness suffers
- (1) 1.75 inch I.D. (inner diameter) x 24 gauge x 116mm long carbon fiber tube from Dragon Plate
- (2) 8mm O.D. x 6mm I.D. x 55mm long carbon fiber tube from Amazon
- (2) 10mm O.D. x 8mm I.D. x 96mm (minimum) long carbon fiber tube from Amazon
- (2) 10mm O.D. x 8mm I.D. x 122mm (minimum) long carbon fiber tube from Amazon (use leftover from above item)
- (2) 10mm O.D. x 8mm I.D. x 18mm long carbon fiber tube from Amazon (use leftover from above item)
- 1/16 in. x 1 in. x 12 in. carbon fiber from McMaster (used to make motor mounts – very important to use the woven pattern type!)
- (2) 0.057 inch x 0.177 inch x 100mm long carbon fiber from Prop Shop
- (1-2) plastic Easter eggs OR (2) Estes NC-80B nose cones from Amazon (a bit more sturdy than the eggs… a lot of wasted material though)
Electronics: (some things can obviously be substituted as long as they fit)
- (4) Cobra Champion 2207 2450kv motors from Prop Shop
- (4) Super Racerbee 30a ESC’s from Pitch Roll Yaw
- Betaflight F3 flight controller from Piroflip
- Anbee Circular polarized cloverleaf antenna w/ RP-SMA plug from Amazon
- Crazepony mini FPV cam from Amazon (AWESOME camera for the price)
- Any standard sized 1300-1400mah 4s lipo or 5s 1300mah lipo (5s Tattu fits)
- Small satellite Rx such as the Lemon Rx DSMX from Amazon
- 5 minute epoxy from Amazon
- Starbond from Amazon
- Clear silicone sealant from Amazon
- Aluminum heat sink from Amazon
- (2) antenna tubes with caps from Prop Shop
- Threadlock (Loctite 242) from Amazon
- Monokote trim from Prop Shop
- GoPro type lens (I had one in my electronics junk pile)
- 3mm nylon wire mesh guard from Prop Shop
- Anodized low profile prop nuts from Piroflip
Obviously, much of the construction is very “make it up as you go” depending on the electronics you use. The only straight forward assembly procedure are the arms as shown in the development section. To secure them in place, I used Starbond (I swear by this stuff now) and then topped it off with epoxy.
Be sure to add holes for the motor wires to exit the arms and be soldered to the ESCs. Add these holes before assembly.
Also, for ease of arm hole locations, here are links to scale drawings of the “unwrapped” body which can be cut out and taped onto a carbon fiber tube. I have also included the pattern for the VXS.
I used a dremel tool with a cutting disk to cut all the carbon fiber. It is an absolute mess and it is imperative that a quality ventilator mask is worn such as this one from 3M. Carbon fiber dust is not toxic, but it is extremely irritating to the respiratory system.
The ESCs are the trickiest part of the build. Cut outs must be made for the ESCs to sit in – but the cutouts must be made oversized to avoid shorting out on the frame.
The best method I found for mounting the ESCs was to use some tape to hold them in place while I epoxied them in. Double check to be sure all the batteries you plan to use have enough clearance to fit (I made that mistake!) Below are some pictures of the ESCs on the VXV project. I ended up using solid copper wiring to get more clearance and so the motor wires fit in the 6mm arm tubes.
I used 2 different mounting methods. One I used on the VX1 and the other on the VXS. After having gone through crashes with each frame, the VX1 method is the stronger of the 2 – not even a free fall from 100 feet budged the motors. For the VX1, I simply drilled holes in the arms (not all the way through) for the screws. I then drilled smaller holes on the opposite side of the arms for the hex wrench to fit through. I made the holes small as possible since the holes weaken the arm. For good measure, I then secured the screws with Starbond and added a little fillet of epoxy between the arm and the bottom of the motors.
There are probably better tools to do this, but this is how I mounted the VSX arms: I drilled a 1/16 inch hole through the arm. I then had to insert a small file and file in a 1/16 inch x 1/4 inch slot. 1/16 inch x 1/4 inch x 1 inch pieces of carbon fiber (with motor holes) were then inserted and secured with Starbond and epoxy. Spacers were used in between the motor mount and motor.
For the flight controller to fit the way I wanted, I had to file down the corners. Yes, this is scary but I have not had any issues. Double check to make sure you wont be filing away any circuit tracks and you will be fine. Be sure to clean the board of any filing dust.
Here is a picture of the FC on the C1 project:
The rest of the components are not so straight forward. Here is where I opted to place the rest of the components:
- Vtx: in the back cover
- GPS: epoxied in a cutout on the top of the body
- Rx: stripped of the platic case, protect with electrical tape, and mounted inside at the top of the body
May 26, 2017: The only differences I made to the VX1 was to use a Tattu 1300 5s battery and APC B5x4.5E-B4 multirotor propellers.
Note 1: OSD speed is in kmh. A list of the speeds recorded, direction, battery voltage, and time in video:
- 212 kmh (131.7mph) 12.5v West to East 0:31
- 230 kmh (142.9mph) 13.6v East to West 0:47
- 240 kmh (149.1mph) 15.3v West to East 1:06
- 248 kmh (154.1mph) 14.4v East to West 1:21
- 241 kmh (149.8mph) 14.6v West to East 1:34
HOWEVER – GOING FOR AS MUCH AS I CAN GET: Blackbox/gps log says 69.06 m/s which is 154.5 MPH
Note 2: Flight took place around 8pm on May 26, 2017 and winds were 5mph max. Link to the wind conditions.
Note 3: Very poor OSD reception/quality. The Eachine DVR records a black screen if there is any interference.
Note 4: After much research, I have found out that GPS is the most accurate way to determine speed. A summary of my findings: https://downanddirtydrones.com/gps-vs-radar-gun/
Note 5: Flight ended in a small crash – I thought I had turned on angle mode, removed my googles, pulled back on the stick to land it and realized too late I was still in acro.
It was interesting to see how dramatically the speed increased between the first 3 passes and how the voltage drop improved as the battery warmed up and internal resistance dropped. I’m also very surprised at how much of a voltage drop there was! The battery was very warm, but still able to be handled.
Please read the article on why GPS is more accurate than radar gun (or view the references below). Radar guns should never be used to verify GPS – nothing that is less accurate should be used to check something that is more accurate. Would you use a ruler to verify the measurement of micrometers?
GPS log also shows that max speed was not taken during a “dive”. Max speed was recorded about 3.8 seconds into a shallow climb (which lasted about 7 seconds). After analyzing this graph, it is also clear to see the correlation between higher velocity and sustaining level flight.
Freda P., and A. Angrisano, S.Gaglione, and S. Troisi, ”Time-Differenced Carrier Phases Technique for Precise GNSS Velocity Estimation,” GPS Solutions, Doi: 10.1007/s10291-014-0425-1, 2014
Hoffmann-Wellenhof, B., and H. Lichtenegger, and J. Collins, Global Positioning System: Theory and Practice, Springer, Berlin Heidelberg New York, 1992
Olynik, M., and M. G. Petovello, M. E. Cannon, and G. Lachapelle, “Temporal Impact of Selected GPS Errors on Point Positioning,” GPS Solutions, 6(1-2): 47-57, 2002
Serrano L., and D. Kim D, R. B. Langley, K. Itani, and M. Ueno, “A GPS Velocity Sensor: How Accurate Can It Be? — A First Look,” Proceedings of the ION National Technical Meeting 2004, pp. 875-885, Institute of Navigation, San Diego, California, January 26–28, 2004
Szarmes, M., and S. Ryan, G. Lachapelle, and P. Fenton, “DGPS High Accuracy Aircraft Velocity Determination Using Doppler Measurements,” Proceedings of International Symposium on Kinematic Systems in Geodesy, Geomatics and Navigation – KIS97, pp. 167-174, Department of Geomatics Engineering, The University of Calgary, Banff, June 3–6, 1997
Van Graas, F., and A. Soloviev A., “Precise Velocity Estimation Using a Stand-Alone GPS Receiver,” Proceedings of the ION NTM 2003, Institute of Navigation, Anaheim, California, January 22–24, 2003, pp. 283-292
Since the VXV project has slowed down a little, I decided to put the tail cone for the VXV on the VX1. It is basically a carbon fiber spinner for RC planes. The cone shape (as opposed to a rounded shape) will further reduce drag. This doesn’t work the same way for the front. If travelling below mach 1, a blunt round nose is best. This is why the Concorde had a pointed nose while all other airliners (travelling below mach 1) have a rounded nose.
The Easter bunny was kind enough to leave me an egg that was closer in diameter to the body than the last one I had used. The shape of the nose cone will blend into the body a bit better which keeps something called the boundary layer (of air) to stay intact and continue flowing along the body. If the boundary layer separates, it gets turbulent.
Another 150+ MPH Run:
June 8, 2017: Another calm day (weather conditions link) with winds calm, gusting to 5mph. I won’t list all speeds for each run, but again it broke 150 and hit a top speed of 153.2mph. I also got video with the Mobius Mini, but the video isn’t the greatest – I could only get the Mobius on by mounting it backwards and putting an angled mirror in front of it. The mirror wasn’t big enough, so almost the whole bottom is the view looking towards the rear of the quad. Next time, I’m going to do it without the mirror – the view looking out the back is actually pretty interesting.
Once agin, the APC props and Tattu 1300 5s lipo was used.
- Mobius Mini video with FPV video: