20 December, 2020
STARTING CART
Completed project
For a couple of new gliders with impeller propulsion I needed to decide on a way to launch it. In the past I had used a bungee launch system for some EDF-models. And although this worked without issues, I never felt comfortable with it. The other option was to use a starting cart. I was not very impressed with the commercially available ones. The main weakness with these carts in my view is that the model is just supported with the wing so that model and cart can move independently from each other. Maybe this is OK when the glider is pulled by a glider tug but with a self-starting model, there is no way to control the direction. In a lot of videos you can see the carts moving in almost every direction, except straight.
In an RC-Network thread (https://www.rc-network.de/threads/die-etwas-andere-startmethode-der-startwagen.47828/) I found some good discussion and a couple of interesting solutions. Based on this I decided to build my own version. The core of the solution is that the model is not supported under the wing, but with the fuselage. Because of this, there is a good “connection” between the cart and the model due to the curved fuselage shape, making it possible to control the direction to some degree during the start. A consequence of this is that the cart is specific to a model. I did not see that as a big issue for my own use. Anyway I designed the cart so that it can be modified relatively easily for another model.
Starting point is a glassfibre saddle that is molded on the fuselage. For this, two layers of clingfilm are taped to the fuselage. In principle one layer should be enough but I am always afraid that there might be a small hole somewhere. I don’t care too much about having some wrinkles but try to avoid trapped air bubbles. Three layers of 160gr/m² are applied with epoxy resin. I make the saddle quite long, about 500mm. Approximately two thirds of the area is in front of the centre of gravity and one third behind. After cure, the laminate is removed and cut to shape. I also remove a large portion of the middle of the laminate so that only an outer contour of around 30-40mm wide remains. This is to avoid any risk of the saddle “sticking” to the fuselage due to the vacuum effect between two large surfaces. The saddle is then bonded to a wooden frame. This becomes the removable part of the cart, to be made to size for each model.
The cart itself consists of a wooden base plate that connects the axles for the front and rear wheels. The axles are square aluminium tubes with round aluminium tubes bonded in at the ends that serve as axles for the wheels. I also glue wooden rods inside the round tubes for added strength. I included one additional feature to the cart to be able to position the rear wheels higher or lower. This results in changing the incidence of the wing to adjust the take-off behavior. If a lot of elevator is needed to pull the model in the air, the rear wheels can be positioned higher to increase the wing incidence, and visa versa. My starting value is 3 degrees. For the Confluence I reduced this to 2 degrees after some starts. I realized that a consequence of this construction is that when changing the cart for another model, this incidence setting maybe also needs to be changed, requiring an additional operation during the switch. In a next version of the cart I will see how to move this incidence adjustment option to the removable part.
For the wooden parts I used 2 layers of Liteply bonded together for a “laminate” thickness of approximately 7.5mm.
The first wheels I used where cheap units bought in a DIY-store. They had a diameter of 150mm. Although they were quite OK, I had the feeling they were limiting the speed of the cart, maybe because they had no ball bearings and because they were relatively wide, resulting in an amount of friction with the grass. Looking around I could not find anything better for a reasonable price, so I made my own. My first thought was to cut wooden discs but then I wondered why not to print them, the only limitation here being a maximum diameter of 200mm on my printer. I designed a wheel consisting of a relatively strong rim and a separate “tire”. The thinking behind this being that the tire could either be printed with a flexible material, or the shape of the spokes could provide some flexibility. Not wanting to buy additional printing filament I went for the second option. The tires printed with PLA are still surprisingly stiff. Rim and tire are bonded together with epoxy. Of course I mirrored the wheels for left and right due to the shape of the spokes in the rim, because I didn’t want to cause a side force when the wheels are rotating (joking here!). In the center two ball bearings are installed. I bought some very cheap bearings on Ebay, throwing them in the garbage immediately because that was exactly what they were. I then bought some (slightly) more expensive ones from SKF, these felt like bearings should. A good example of “buying cheap is buying twice”. The size of the bearings is 6x22x7 (ID x OD x W). This seems to be a very common size.
The (almost) last item of the cart is two vertical plates that align with the front of the wing. The real final item I added after some starts is a couple of small foam pieces (I used depron) to keep the wing (approximately) level. When taking off with some crosswind I had experienced that the model could rotate in its saddle. Although this could be corrected with ailerons it increases the workload for the pilot unnecessarily.
Up to now this cart works very well for me. I have been using it for the Jeta and the Confluence. Using rudder to keep it straight during take-off has been surprisingly effective, even with the Confluence where the rudder is not directly in the airflow path of the impeller.
The pictures show the starting cart for the Confluence.
