Thread Tools
This thread is privately moderated by BAF23, who may elect to delete unwanted replies.
Jan 19, 2013, 04:10 PM
The sky is the limit
BAF23's Avatar
Build Log

FMS T28 Part 7: final assembly and wiring

Part 7: final assembly and wiring

After getting the fuselage up the cradle and removing the canopy it became time to hook up all the servo’s and get those flight controls to behave. With the gear sequencer in place (doors open first, gear leg lowers with a delay and vice versa), I noticed it got rather hot so that would have to find a cooled placed as well for final positioning. It also required a cycle or two to get everything right after power ups but luckily I had no serious jams when things didn’t go as advertised. I wanted the gear down so I could arrange the directional setup. I already had decided I would use separate channels for rudder and nose wheel steering (NWS) instead of using the provided Y cable, because I wanted to be able to adjust their neutral point (by subtrim) separately, and have serious expo on the steering because the very short distance between main and nose gear probably would cause touchy steering at higher speeds during the ground rolls.

I first tried to connect the aircraft rudder on the rudder channel, and the NWS as a mix on one of the aux channels. I didn’t want any expo on the aircraft rudder control because I want instantaneous response to cater for gusts in finals, and also prefer to have rudder trim effectiveness to cater for engine torque according to power (maybe by mix later). Having aileron rudder (mix) interconnected to eliminate adverse yaw from the (differential ailerons) it is also better to have a positive movement instead of letting most of it water through the sloppy rudder expo. I tried to get an aux channel to slave on the rudder, but was unable to create an independent aux expo on my DX10t. After intense thinking I came up with the unorthodox solution of reversing the setup.

I connected the NWS to the rudder output and programmed 90 percent expo to it. I previously successfully used that system on a rudderless Robbe F16 impeller jet which also travels pretty fast on a short coupled gear during takeoff and landing (high speed tricycle effect). So much expo gave me perfect directional control during takeoff and landings using up to hallway rudder stick deflections, and maneuvering at taxi speed in tight turning radius using the more extreme deflections. Aux 3 input was then redirected to the rudder stick, and without expo deflects the rudder in a linear way with the rudder stick. A mix was then selected between the aileron channel and aux 3 to minimize adverse yaw without any inputs to the NWS. This setup is far from standard, but has many advantages and allows completely independent trimming for neutral points, expo and max deflections.

Next item on the list were the elevators. FMS uses two separate elevator surfaces, each with their own control horns and pushrods, squeezed together by one small screw in a loop affixed on top of a single 17gram servo. I already had nightmares about that single screw untightening due to my nearby main sound speaker vibrations, and initially thought to get a form of security by getting the two rods tightened in an second ring close to the first one. If the one on the servo arm got loose, it still would be able to push the second one (elevator up movement) to get the airplane on the ground (a trimmed aircraft will always descend using bank, and at lower speed not extending flaps it needs up elevator).

This setup seemed to work but still relied on that single FMS 17gr servo. Having everything duplicated up to the servo arm, and sufficient space to install an extra 9gr servo, I preferred to have each elevator driven by its own servo, connected by Y cable to the elevator pins on the receiver. That would allow even a servo failure or loose linkage on one of the elevators to be completely independent from the other one. It was important both servo’s had the same speed and arm length. The extra servo I tried didn’t, but the original 9gr rudder servo did, and could be fit neatly underneath the higher 17gram elevator servo arm. I even could use some predrilled servo attach holes. Installing a spare 9 gram servo from the scrap box and connecting it to the rudder was easy.

I now feel much more confident as running both elevators even on just a single expensive quality servo. The airplane now has a much better chance to be safely recovered (on half the elevator effectiveness) in case of failures. The ailerons using different channels commanding separate servo’s already provided the same degree of redundancy. With twin system reliability in pitch and roll you dramatically improve the chances of recovering from otherwise fatal crashes (rudder or throttle are not considered critical in precautionary or forced landings).

As already mentioned a few chapters ago, I preferred running the throttle and sound system on separate channels to allow better synchronization between prop starts (and stops) with the startup and shutdown sound, and low idle sound together with minimum allowable prop rpm. I already had programmed a throttle curve for the engine (to allow more precise rpm control during final approach), and tried to couple that on an aux channel, using each channels’ capability to be independently programmed in subtrim and servo travel. After spending a complete evening trying to obtain the desired result, I finally gave up. Because none of the adjustments was visible (there are no servo arms moving but only electric inputs going into the ESC and sound card), everything had to be done by listening to the (propless) engine and speakers. The sound card also has to be reactivated after each shutdown by retarding the signal to well below idle (similar to arming an ESC). I was running in a dead end but could get an acceptable result without any throttle curve. I also didn’t understand why the Spectrum DX10t software denied me the use of the gear channel on the channel assign. With twin aileron, twin flap and twin throttle channels I became short on aux channels because I also planned on having the cowl flaps on a separate channel as well. When I ordered the Spectrum Dx10t system it included a separate plus8 receiver module to increase the number of channels, but this could not be plugged into the AR10000 receiver.

Next morning I took good notes on the connections and took the AR10000 10 channel receiver out, replacing it with the (also in the package) AR9020 9-channel receiver to which I plugged the plus8 module. I then plugged a cowl flap into the x+1 channel, and tried with the sound system into the x+2. After another hour of experimenting I understood the note in the +8 manual that these extra channels were only for accessory drive, not for flight controls. Linear throttle movements and trim adjustments just weren’t transferred to those supplementary channels. Even after binding the new receiver and plus8 extender my transmitter stubbornly kept assigning the left flap actuation to the aux5 channel, which is nowhere to be found on the AR9020 receiver (the pins being used to connect the plus8 receiver). My brains were put to a serious test to resolve all those problems.

I freed up the gear channel by assigning my gear selector knob to channel x+2 on the secondary receiver and that worked. Aux 4 was reprogrammed with the throttle as input, and I connected it with a BEC eliminator cable to the sound card. Reason for that is I wanted the sound system to get its power through the engine battery (through a Deans Y-cable). I did not want to risk problems with the sound unit power by getting a different power input through the receiver which will be fed by a separate 10Amp Bec through a dedicated 2S flight control battery (I like redundancy and don’t trust incorporated BEC’s on aircraft with so many servo’s). This time the adjustments between the throttle inputs and electric motor and sound system could be made to work well after eliminating throttle curve and getting a linear input again. I programmed a 2 sec servo speed in the sound receiver channel and that really eliminates any audible (clicked) throttle response, sounding more like a real radial engine accelerating or decelerating with all its built in inertia (without sacrificing instant throttle response to the engine itself). Throttle curve can be made active again by activating a single command, and probably will be used on the first flight(s) without sound system connected, until I get a good feel of engine power adjustments versus flap settings for landings.

Then came the time to finally install the 3 speakers into the fuselage and start getting the rest of the wiring sorted out. None of the speakers got firmly attached, the main speaker is held up by gravity pushing it against Velcro around the receptacle in the fuselage bottom, the fuselage cross member rests on the top of the assembly preventing it to move up during negative g’s or rebounds). The side speakers are held in their custom carved cutouts by inserting foam bits behind, to keep them in place. That allows more vibration freedom and hopefully will minimize the transfer of those vibrations to the fuselage. A final test after winding the sound input cable around a ferrite ring, proved I eliminated any previous interferences and the system worked even at full (sound) power without resonances. I then was able to get a proper fixing for the ESC in the extra air channel by means of cutting some spare pieces of foam to hold it in position but still being removable in case of replacement or access to the left speaker is needed). In the picture below you can see the ESC is kept in the cooling gutter (from the now functional oil cooler intake) by two top compressible foam pieces. Behind it , one of the satellite receivers. To the right the AR9020 receiver with plus8 unit behind, close to the side speakers but apparently causing no interference.

Connecting the second cowl flap through a Y-cable I noticed both servo’s were not balanced, one door remaining slightly open while the other one was closed. Forcing it closed by adjusting the servo travel caused the other one to buzz. Adjusting the servo rods is impossible (without forcing the whole mechanism out of the foam and already covered fuselage) so I opted to connect the second cowl flap to the x+3 channel, and having that controlled by the same transmitter switch as the x+1 channel. I know FMS and most people choose to connect the cowl flaps to the wing flap channel, but in real aircraft there is no common logical link and both have to be operated completely separately. On the model it wouldn’t be very visible in flight but you have to remember I will use this model as a step up towards a Windrider Boeing 737 for getting used operating all switches and various systems. On the B737 I intend to use spoilers (in front of the fowler flaps as on the real aircraft) and want to deploy those after touchdown (not in flight). I want my fingers to be trained performing such actions and thought opening the cowl flaps on the T28 was a safe way to exercise that. On the grounds it would be more realistic to control those (fake) cowl flaps with a rotary or sliding knob on the transmitter, but the way I plan to operate the B737 spoilers will not allow partial deployment without potential dramatic asymmetrical aerodynamic results. On my transmitter I control gear and flap operation with the knobs to the left of the throttle (I fly mode 2). Figuring nobody will see cowl flap operation on the T28 in flight, but wanting to exercise their deployment directly after touchdown while I’m busy controlling the nose wheel steering with the left stick, I elected to assign the extreme right button on my transmitter to the cowl flap operation (similar to spoiler deployment on the B737 by a handle at the right of the console?). With both cowl flaps on separate channels I now easily can adjust their end points individually, a luxury now available by having 17 channels on the receiver.

Remember I mentioned Aux 5 being assigned for left flap control? Having no Aux 5 pins available I tried to slave the whole thing to one of the remaining free channels like x+4 but whatever I tried, I never got it running in syncro with the right flap. After the previous experiences with the throttle input problems, I figured flaps are also considered as flight controls my Spectrum, being programmed in a separate setup page with own deployment time, percentage deployment depending on 3 position switches, and elevator compensation (all on that page), independent from servo speed or mix functions for the different available channels. I thus had to find a free channel on the main receiver and started scratching my head until I finally was able to assign the left flap function to both aux 5 (inhibited) and the gear channel. Within half an hour I got both flaps to lower in syncro, final touches being made by adjusting servo trims in the takeoff position, and servo travels for up and down positions, all by eyeballing from the front middle of the wing straight along the flaps, and from the back compared to wing root and ailerons, for up and take off positions.

It was a hell of a job getting all those systems setup individually but everything now works and is ready to be integrated into a fuselage wing mating, but I also want to rearrange the lights wires, and not connect everything together at once. I want to keep 3 separate bundles, one for the fuselage, and one for each wing, so they still can be separated for maintenance or in case of crash. With this aircraft and its abundance of connectors (lots of Y-cables and gear sequencers, a receiver plus augmenter and 2 satellite antennas, 2 batteries and a separate BEC, 2 extra sound speakers) a proper solution has to be found to get everything neatly packed and routed in the available space, unless you want to end up with an apparent can of worms when you open the cockpit compartment to connect the batteries.

For starters I connected the upper and lower rotating beacon together at the back of the cockpit. I still keep open the option of later commanding both those lights through one of the free channels (as in real aircraft switching the beacon on just prior engine start). Because in the meantime all lights will be on as long as the battery is connected, I used one of the now redundant supplied 4to1 Y cables to connect all the lights into one free x+ channels. The plus 8 auxiliary receiver can be powered either from the main receiver’s power source through use of a jumper plug, or by a separate source. My plan is to initially fly short hops with minimal use of fancy systems, from a single 4S4000mah 25c(50c bursts) battery. After center of gravity, flight control deflections, aileron rudder mix and various expo’s have been adjusted (in a no flap gear down configuration), that battery will only be used to power (through the Deans Y cable) the engine, sound system, and auxiliary receiver. This will be done by adding a full Y servo cable on the ESC power/command exit, after the split reduced on one side to red and black towards the auxiliary reveiver’s battery plug (powering all lights and cowl flaps), and on the other side with the red interrupted to the main receiver for engine control (BEC eliminated). An independent 2S950mah battery through a Castle creation 10Amp Bec will be connected to the main receiver to power it and all the important servo’s. Only from then on will the flap operations be tested against the elevator compensation (to eliminate ballooning upon extension) and the gear be raised and lowered (expecting higher current drains for those), lights burning all around and cowl flaps operated.

With the successful ground tests and setup of the various groups of systems, I separated and bundled all fuselage wires through both sides of the fuselage under the cockpit. I then assembled the wings without gluing anything. Although I do not plan to remove the wings for any reason, I like to keep the option to dismantle both wings and fuselage in 3 separate parts with their own wiring lump. With both wings kept together by the two sliding in hollow spars and the bottom plastic joining pieces, all the wing wiring got pushed through their dedicated holes into the fuselage. Servo cable extensions were custom cut (you definitely don’t want extra lengths in that rats nest) and inserted where needed, and all landing gear and doors controls fed into the sequencer which was affixed by Velcro perpendicular to the airstream on top and at the back of the nose wheel box. I figure that way it will benefit from extra cooling when the gear is traveling or down, and still be in the airstream behind the engine with gear up. The cooling air forced by the propwash through the oil cooler channel into the ESC also finds its way out along the sequencer. With 3 gear legs and doors all controlled individually through Y cables you can imagine the unavoidable spaghetti concentration in that small area. Aileron and flap servo’s were each individually connected to their respective receiver channel. Wing navigation lights were not merged enroute (so as to keep the wings independent) and the landing light also got a separate connection because I still toy with the idea of having it controlled by the landing gear position later on.

The wing wiring lumps got rearranged as well and I was surprised how neat I got everything packed with room to spare. Using the patterns of the long ago eliminated wing walk strips, I cut new wing walk areas out of black grade 100 abrasive paper for steel (not on paper back) which I glued in the correct place (as seen on the banked airplane shot in part 6) with good old Pattex contact glue. The result is much more realistic as just applying black vinyl strips, but I realize lift over that slightly corrugated area will be less as on a smooth surface. I applied the same in previous Dynam DC3 and Art-Tech T6 models (both also
I noticed after powering up some things still moved so after checking everything operated correctly without interferences, it was time to bind the system again in the fail safe mode. Although I will stow and transport the model with gear up, I always bind my models with the gear down. Reason is that when you are parked on the field (on your gear), you can change batteries as much as you want without risk of temporary commanding gear movements during connections, (always with all transmitter switches forward to eliminate errors). Only during transit (when the model will be on a cradle on the field) the gear knob has to be positioned differently during connections or before final disconnect. Flaps and cowl flaps also got bound in a up/closed position (switches forward again) to minimize drag during eventual power loss times. Once the model is on its wheels, I lower the flaps (as on the real T28 because the flaps incorporate steps to crawl onto the wing), and open the cowl flaps for the sake of showing off features (will be used on the B737 flaps/spoilers as well). Next came the weight and balance, but that is for the next and final build chapter.

Click here to read the final part eight of the T28 build log
Last edited by BAF23; Jan 25, 2013 at 03:17 AM.
Sign up now
to remove ads between posts
Jun 15, 2013, 07:27 AM
Let's fly and enjoy
Just Fly26's Avatar
Very nice indeed. I admire your ability to modify this nice Trojan T-28, i've got one as well but evrything still stock. I wish i could do the same as well.

Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Discussion Goldberg Eaglet 50 7 year project finally done! turbojoe Balsa Builders 12 Feb 13, 2016 10:21 PM
Wanted Wanted: Tail rotor assembly for Falcon 3D/400.. also other parts Maduro Aircraft - Electric - Helis (FS/W) 0 Jan 15, 2013 03:04 PM