You may have seen the discussions in the forums about this project. Now read the first article in series of three about Dr. Andrew H. Watson's (aka Doc Watson) X35-B!
Vertical Take Off/Landing (VTOL) Electric Ducted Fan Scale Model
This story of the VTOL will be told in three parts. The following is Part 1.
Part 1 - History and Basic Design
Part 2 - Detailed Design and Building
Part 3 - Flight Testing
Part 1 - History and Basic Design
History
I have dreamed for a long time (since about 1992) of building a scale VTOL radio controlled model and the natural choice, being British, was the Harrier (you will probably know it as the AV-8B). I started researching the full-scale airplane and through various sources found that the required thrust to weight ratio required for VTOL flight was 1.15 so for a 10 pound model you would require 11.5 pounds of thrust. I took this as a starting point and looked for the most powerful ducted-fan and engine in the market at the time. I finally got an OS91 VRDF engine and a Viojett fan unit that could produce 14 pounds of thrust. I started designing the model making it as near scale as possible, the only compromise being the four nozzles that the harrier uses to vector the engine thrust. The plans worked out at about 1/8 scale (a big bird) then started designing the internal ducting. I don't want to dwell too much on the actual design, but after calculating the losses due to turning the airflow through 90 degrees in the nozzles, the available thrust dropped to around 8.5 lb. After weighing all the required components, it was a marginal design. Various other events then transpired which eventually prevented me from building the actual model.
I know of two or three people around the world who have had various degrees of success with a Harrier design. The best, in my opinion, is one produced by Eric Dainty who has successfully flown a 'full' sortie comprising vertical takeoff, transition to conventional flight, followed by transition back to hover, and then landing vertically. As a side note, I did read an article a few years ago about Eric's model being 'handed over' to Yellow Aircraft for final flight-testing, presumably to be developed into a kit. However, the model required one pound to be 'lost' before true VTOL performance could be achieved. I haven't heard anything from that point on; does anybody know what's happening?
Very important lessons were learned from the Harrier design, the first obviously being weight. It is the most critical factor when designing a VTOL aircraft, also complicated by the 'weight loss' associated with IC engines, and the possible change in CG due to spent fuel loads. The second lesson learned was that the losses due to turning airflows through 90 degrees were very bad.
So, time passed by and I spent the next few years concentrating on my PhD (working for Rolls Royce on turbine designs, and British Aerospace on 'evolving' future fighter aircraft). I still managed to build a few wet aircraft and then one day my brother showed me his latest creation, an electric powered glider. I was amazed at the power the prop driven model produced. After a little research on the web, I found a site that produced electric ducted fans. The site was WeMoTec and it was the first of three pieces of the jigsaw required for me to produce a VTOL scale model. The second and most important piece was finding this site, an absolute revelation. All these guys were producing EDF models whose performance was comparable to the IC powered DF models. I joined the E-Zone in April 2000 and flooded the EDF forum with questions about motors, the best fans, cells etc. etc. A big thanks to Andy Willetts must go here for answering my early 'greenhorn' questions and giving some sound advice on the path to producing flyable electric models.
I had the fan data I needed, and I had quickly run up the electric learning curve (with a little help from you guys). So there was only one thing left, what scale VTOL airplane would I build? The answer came as I did a routine 'VTOL' search on the Internet early in 2000. I stumbled across information on the US competition between Boeing's X32 and Lockheed Martin's X35. The winner was to produce the next generation of fighter aircraft, the Joint Strike Fighter (JSF). There were to be three variants of the aircraft, the 'A' version being conventional, the C version carrier based, and the B variant being STOVL, which would replace the Harrier.
All the information was there, so it was then a case of deciding which one to build. The decision as to which would win the contract not being made until October 2001. As some of you will know, Murphy's Law stepped in and took control. I decided to design and build the Boeing X32-B. The winner was the X35 and I was actually 50% into the build when the final decision was made. Therefore, I scrapped the X32-B and started designing the X35-B. Anyway, I'm getting ahead of myself so back to the design.
Designing The X32/X35
I did initially look at both designs and produced a fan layout that would work for both airframes (luckily!). I acquired a WeMoTec catalogue and started playing with my calculator. I quickly realized that I could power the model using three fans in a triangular arrangement with the two rear fans rotating to provide the transition to and from hover. This got around the losses problem associated with the original Harrier design, it also allowed control of the pitch and roll axis by varying the motor speeds to each motor using lightweight pizo-electric gyros. Things were looking promising. The front fan being twice the distance from the CG position as the two rear fans and remains fixed pointing down. The two rear fans rotate through 90 degrees for transition to 'normal' flight from the hover position.
I needed to optimize the fans/motors/cells combination. After about a month of playing with my calculator, I wrote a program that would provide the duration of hover and the amp draw of the motors for various model weights, cell types and numbers, fans, and motors.
The three motors would be connected in parallel to the cells, so the total amp draw from the cells would be three times the single motor draw.
The computer output looked as follows:
C
(g)
(lb)
T(g)
w
A
A(tot)
Time
2400mAh
12
2971
6.55
1089
458.0
34.7
104.1
83.0
16
3215
7.09
1179
511.0
29.0
87.1
99.2
20
3459
7.63
1268
566.9
25.8
77.3
111.8
24
3703
8.16
1358
626.3
23.7
71.2
121.4
28
3947
8.70
1447
689.7
22.4
67.2
128.6
32
4191
9.24
1537
758.3
21.5
64.6
133.7
1700mAh
12
2719
5.99
997
405.9
30.8
92.2
66.3
16
2879
6.35
1056
438.7
24.9
74.8
81.8
20
3039
6.70
1114
472.5
21.5
64.4
95.0
24
3199
7.05
1173
507.5
19.2
57.7
106.1
28
3359
7.40
1232
543.6
17.6
52.9
115.6
32
3519
7.76
1290
581.2
16.5
49.5
132.6
1300mAh
12
2647
5.84
971
391.4
29.7
89.0
52.6
16
2783
6.14
1020
418.9
23.8
71.4
65.5
20
2919
6.43
1070
447.0
20.3
61.0
76.8
24
3055
6.73
1120
476.0
18.0
54.1
86.5
28
3191
7.03
1170
505.7
16.4
49.3
95.0
32
3327
7.33
1220
536.3
15.2
45.7
102.4
VTOL Design Calculations, Dr A.H.Watson 23/04/02
The printout shows three cell types, and cell counts for 12 to 32 cells in steps of four cells. The first column shows the number of cells, and the next two columns are the Expected AUW of the model in grams and pounds. This is determined by weighing all components and estimating the bare airframe weight. T(g) is the thrust generated by one fan with the watts in and amp draw for one fan next to it. The next column A(tot) is the total amp draw for the three motors, and finally, the last column the duration of a hover only flight.
The program required detailed thrust versus Watts in curves for the fans (I had already decided to use WeMoTec HW620's). With this data collected and graphs drawn, equations of the 'best fit' were produced and plugged into the program. (Note: all WeMoTec fan curves are available on my site under the 'Fan Data' section).
As an example, if I used twelve 1300mAh cells, the AUW would be 5.84 lb, the total amp draw 89.0A, and the duration 52.6 seconds. It's an optimization problem. Fewer cells means less weight and higher current draw, and more cells mean fewer amps and more weight!
I then narrowed the cell count to 24 cells. This then enabled me to use 2x12V lead acid batteries connected in series, which could power the model with 24V power source, via an umbilical for testing and setting up the control system. I figured that I would need a long time setting up the gain values and tweaking the control system before I let the bird lose and fly freely. In addition, I didn't want to wear out the cell pack, at £130 each it was another expense I that I didn't need. The lead acid batteries are 60Ah cells so I could run the setup for nearly an hour, although the motors might get a little warm!
This then narrowed the choice to the following:
2400mAh
24
3703
8.16
1358
626.3
23.7
71.2
121.4
1700mAh
24
3199
7.05
1173
507.5
19.2
57.7
106.1
1300mAh
24
3055
6.73
1120
476.0
18.0
54.1
86.5
VTOL Design Calculations, Dr A.H.Watson 23/04/02
Duration was then the driving force so it was the 2400mAh cells that would be used. This gave a total amp draw of 71.2A and a duration of two minutes. This is of course if I could keep the model at or under 8.16 lb, a very hard task to achieve.
I decided on the following setup.
Motors: 3 brushless HP220/30 A3 S P4 motors.
Fans: 3 x WeMoTec HW620-E 86mm fans.
Cells: 24 x 2400 mAh cells, matched and zapped.
ESC's: 3 x Schultze 35bo controllers.
The motors can produce up to four pounds of thrust each at about 45A. The projected AUW is to be 8.19 pounds, so it looked on paper as if it can be done. Obviously, if all three motors were drawing 45A, the cells would die, as they would need to supply 135A!
From the calculations, the motors need to draw about 24A each for hover (that's about 576W per motor and a total of 1728 Watts). At 8.19 pounds AUW that's 211W/lb! Duration is around two minutes while hovering, depending on final weight (which is of course critical). On paper it looks like it could be done, but the question then was would it actually work? I wrote to Chris Golds explaining my design, and he said 'No'. Being a true RAF Navigator, I never listen to Fighter Jocks and charged (excuse the pun) ahead. All parts were obtained in 2001, including a few 'extras' that will be discussed later.
The design looked good on paper and so it was a case of 'put your money where you mouth is', and after buying the motors/fans/cells and ESC's my mouth was empty!
The control of the model would be via a seven channel TX and would be as follows:
Roll: One rear motor speed is increased and the other decreased producing the required roll control, linked through a pizo gyro.
Pitch: Control of pitch is achieved by changing the front fan speed only, again via a pizo gyro. To increase pitch attitude the front fan speed is increased and to decrease the pitch the motor is slowed.
Yaw: A thrust vane inside the front fan placed along the centerline of the plane will provide limited yaw control with another pizo gyro.
Again, this control is a leap into the unknown, so my main concern being the roll control. The nature of the design requires both rear fans to be very close together, and thereby reducing the torque effect of slowing one and speeding up the other rear fan. The yaw control is also an unknown. How big should the vane in the front fan be, and how much will the pitch control be affected by the yaw control?
The rear fans are rotated through 90 degrees by a modified Hitec HS77-BB servo, which is operated via the flaps control on the TX. The signal to the motor on the front fan is mixed with the rotation servo so that when the rear motors are fully down 100% of the signal gets to the front motor, and when the rear motor is in the 'normal' flight position, 0% of the signal gets to the front motor. At 45 degrees rear fan rotation 50% of the throttle signal would be sent to the front fan. This should allow control in hover (although, the thought of £2000 of kit flying fills me with terror), and hopefully, allow for the smooth transition to normal flight on two motors.
Bells and Whistles
I also wanted to add the following features:
Pneumatic retracts, Spring Air 600's being used
Fully working doors and hatches (there are 16 in total!) including sequenced UC doors and doors for hover flight
Scale jet pipe (once I figure out how to do it!!!)
Leading edge and trailing edge flaps for slow (and I mean slow) flight
The reason for all the extra scale detail (and work) is that I intend to enter the model into some scale Jet competitions in 2003. Who knows it might actually work!
I have rambled enough about the basic history and design of my X35-B model. It would require a whole book to describe all the features here, but I hope you at least have a 'feel' for the basic design concepts of the model.
The next article in this series of three will include detailed design and building. Current progress can be seen at my website on www.awatson1.fsnet.co.uk and I am a moderator on the E-Zone EDF chat every Saturday night at 12 midnight GMT.
Jun 03, 2002, 01:00 AM
Similar Threads
