The Model (Phase 4)

#61
On the servo side of things we decided to try the large Seiko servos, loads of torque but a little slow. They have a tie bar with a slot, this allows any disparity between the servo matching without loading the servo because they are fighting each other. It also allows movement of the elevon with one servo failed. This at least gives you a fighting chance depending on where in its travel it fails… a difficult situation but better than no control at all!

2ADBDE86-5AC5-438A-AF80-9F42D0618494-2
by Alex Jones, on Flickr

C8E443E8-328F-4C2B-80D9-B2B7CC3C8BFA-2
by Alex Jones, on Flickr

I have also designed a bolt on tray so we can try some other servos and decide what is best.

936BC4C8-6F9D-4952-AD55-F011F3E55A02-3
by Alex Jones, on Flickr
 
#62
Where does the time go? With the kids now off for the summer I can spend my mornings working rather than home schooling, I hope the circumstances will allow for a return in September.

So, onwards with the build…

Regarding the elevon mechanism I’ve been in close contact with Dave Wilshere and Phil Clark throughout the build and will of coarse do whatever the LMA require.

Our first thought with the elevon servos was 'torque' so we decided to try the Tonegawa Seiko PS-050 servos which do provide plenty of torque although they lack speed.

A video of the servos operating.


They are slow as expected but also give feedback when the elevon is moved down and unexpectedly they appear rather notchy. They don't work well with the gyro due to all of the above. They also require a separate battery and BEC increasing the weight of the model by 2 kgs. We decided not to use these and go with another solution. Dave Wilshere did warn that we might encounter problems but sometimes you just need to see for yourself. They won’t be wasted as we have a few projects in the works where these will be appropriate.

A video showing the Tonegawa servos operating the left elevon.


The next set of servos to try are the MKS HBL 388.

BE0A0D9B-E87A-4B9C-A371-6AB1CC1486D7-3
by Alex Jones, on Flickr
 
#63
I decided to try MKS because it's a brand I have experience with and the specification was within what we had calculated so thought I'd give them a go. I designed the servo trays to be removable so its very easy to try differing brands and specifications if these prove to be problematic.

Flight loads aren't really the problem... I'm interested to see how they react to the elevon hitting the water during our water based testing. Testing is booked for early September as long as nothing changes with the current travel restrictions... fingers crossed! Notice also that I have added internal lights within the hull. It's such a big area that for maintenance this seemed like a good idea, we also need to check for any water ingress using a borescope between runs and the lighting should help.

We've replaced the Tonegawa Seiko PS-050 servos with four MKS HBL 388. These give 68kg/cm (944 oz/in) at 8.4 volts and a speed of 0.14s/60degrees.

A short video showing the elevon, and mechanism.


These worked well so we will see how they perform during the water based testing.
 
#64
If you remember way back I did a series of posts on the ski mechanism. This final video is where we are at with the model as of today with only a few scale additions to add, the velcro is just temporary! before anyone mentions it... Also the tubing will be black once it arrives from Festo.

It could do with being a little faster, I'm able to change the speed using manual meter valves within the fuselage so it's just a matter of tuning the setup.

I've used high flow festo valves for the actuation as the standard model valves couldn't handle the flow rates required. With these new valves installed and working from the transmitter it was all tested before the final electrical and pneumatic installation begins.

 
#65
To round off the ski installation I thought I’d add this picture taken by one of the documentary guys to add scale to the model.

The skis are close to 5ft long and it occurred to me that there aren’t many models that you could stick your head in the wheel well… in this case ski well. There are other models approaching 4.5 meters (14’ 9⅛”) and larger but the sheer bulk of this always surprises me when assembled. It makes my 1/4 scale F104 look small even though they have similar fuselage lengths.

If anyone is ever near the Sun ’N’ Fun Museum in Lakeland Florida, take a look at the SeaDart near the entrance as it’s worth a visit.

A663EF20-9848-448C-92E3-3FB148EC78EE
by Alex Jones, on Flickr
 
#66
With the skis and elevons complete I can get on with the rest of the model.



I’m not a huge fan of after burner rings as they don’t look that great in photos and for my taste aren’t very realistic when operating, they just add weight for not a lot of benefit but its all personal preference. I did try and convince the clients to go without but their argument was that they are hoping for film and photos during those golden hours the rings may illuminate the spray during takeoff so wanted to give them a go. I can always remove the rings and wiring at a later date if they don’t work out.



4910B3BA-26FB-410B-B9F2-DE05AB8B5460
by Alex Jones, on Flickr



The first job was to try and make the ring splash proof, and disguise its presence which was just a matter of masking each light and spraying black.



The A/B lights are installed on a carbon ring that sits inside the main duct which allows a through flow of water. I’ve then designed a static part mimicking the look of the petals on the real engine that slots over this, you cant quite see the lights but should get the reflections off the exhaust shroud at least that’s the plan.



E88C3727-8172-4912-B9F4-76BA7EE24FF7
by Alex Jones, on Flickr



The carbon ring is also used to support the thrust pipe which is mounted inside the main duct using the bell mouth. This duct has a seal which compresses onto the aft turbine bulkhead when bolted allowing water to flow through the intake system. We expect sudden deceleration of the model when the skis submerge from the planed state so expect the following wave to enter aft the aft sections of the intakes. The full-size added considerable amounts of power during this phase to blast the water back out.



CB384615-E2ED-4F0B-820E-9598CE7AD960
by Alex Jones, on Flickr



669E03D7-0185-4349-A1B1-88EE0D620E1F
by Alex Jones, on Flickr



The main duct extends exactly 2mm aft of the last fuselage bulkhead to allow the addition of an o-ring which compresses into a recess in the exhaust shroud when this is bolted up, hopefully providing an effective seal.



3207BD95-4C0F-4168-8E45-0BA96A3EBDBB
by Alex Jones, on Flickr



The problem with a hand built pattern which is then digitised and an internal structure designed in CAD is that you end up with a few compromises. Ideally I would have liked the diameter of the main duct to exactly match the diameter of the shroud but this wasn’t the case due to the large radius applied to the duct for moulding and its not exactly circular. The benefit is that it does create a small ridge to block water moving into the intake ducts caused by general surface conditions when the model is just floating waiting for use. It does however catch the light and draw the eye away from the scale parts of the rear end.



E73CAE32-D2FD-4F84-BE8B-785F6745156C
by Alex Jones, on Flickr



Weathering applied to the shroud and inside edges of the thrust pipe help pull the eye to the static petals. Hopefully this creates a more realistic looking exhaust section.



D443A24B-1290-4590-866A-EC96ED1BD284
by Alex Jones, on Flickr



The lighting is courtesy of my sons Philips Hue gaming lights… I was just playing around with effects wondering if a better way A/B lighting would be to get light up the inside of the thrust pipe somehow… anyway, that’s as far as my thoughts went for the time being, perhaps on another model.
 
#67
While we’re on the backend I should probably mention the water rudders…

These use a 3D printed hinge box that is tied into a carbon structure. This is then covered with a carbon molding and all bonded together with Hysol.

E14BA8DD-C618-4056-9F28-0B6A1E420DF1-2
by Alex Jones, on Flickr

654E923F-7C47-4E14-82C6-ECA7B4F1FFC8-2
by Alex Jones, on Flickr

E0E15785-275D-4059-80A3-745F6E7F22C9-2
by Alex Jones, on Flickr

We decided to use waterproof (IP67) 50kg Savox servos for the rudders. These were tested for several hours submerged and kept working so we shall see how they cope being used in the real world.

If you don't know and are interested in what IP67 means… Water and dust proof connectivity products are defined by their Ingress Protection (IP) numbers. ... IP67 equipment is the most commonly found in the connectivity market. It is 100% protected against solid objects like dust and sand, and it has been tested to work for at least 30 minutes while under 15cm to 1m of water.

0C7C33E2-529C-43D9-8CA9-04DD7E0B9D5B-2
by Alex Jones, on Flickr

The geometry is set in such a way that with the rudder fully extended it's mechanically locked in that position reducing the load on the servo in the fully extended position. It will be interesting testing the turning radius with the rudders and without, we also have the ability to add asymmetric thrust.
 
#69
I like the systems side of things and always enjoy this part of any installation but before I start on the electronics I had to accept that no matter how hard I try water will find it’s way in. This maybe from wet hands while servicing, seals below the waterline or spray getting inside the hatch seals, either way it’s worth protecting the electronics. If I accept that it could get wet inside then all the electronics have to be splash proof so I designed splash proof boxes to protect the electronics. These are 3D printed SLS items with an O-ring seal and perspex lid. Below are some images from the CAD work I did…



43696B13-B225-4C32-9D19-752CB334EB62
by Alex Jones, on Flickr



A3DECA13-799D-40AF-A224-F8C3F2E24892
by Alex Jones, on Flickr



I have provided cooling for the CB400 and pneumatic boxes which will maintain these areas at ambient. Tests with the equipment in the centre fuselage show an increase above ambient of 10 degrees which is well within the operating range of the electronics so I decided not to include cooling in these areas. We’re well within a decent safety margin if operating the aircraft in temperatures that don’t exceed 40°C/104°F
 
#70
I started the main installation by removing everything from the engine test stand and installing this into place inside the fuselage. I used new Festo 6mm tubing but other than that everything is exactly as it was on the stand. All the associated turbine wiring was run through the fuselage. In the CAD images above you can see the carbon tubes that tie the formers together, these are 18mm tubes and also act as conduits for the wiring. Before going to the effort of making everything neat the turbine electronics were all tested to ensure everything worked as expected. I have two Digitech CTU’s on each turbine which have already proved to be very useful.

B3BC8409-6A38-4F3D-A42F-570DAE3D7E46
by Alex Jones, on Flickr

9061F850-5FAD-4671-8D7F-772C2B459E99
by Alex Jones, on Flickr

It’s then just a matter of running all the other wires, this took some time….

A555436D-8AAD-4DFC-AEE3-1EAC549C0FAE
by Alex Jones, on Flickr

ADD8A596-0E29-4163-A133-B0F532AB3ED8
by Alex Jones, on Flickr
 
#71
With all the electrics in the aft section tested and working as planned I can make things a little neater. In the electronics box just forward of the hoist point we have a Jeti REX7A receiver, after burner light ring control, hull lights control and battery to power these lights. Ignore how the antennas are placed as these will eventually be correctly polarised and poke outside the box. Further aft (not pictured) we have a Jet RSAT 900 with the antennas placed on either side of the hull.

91C0B3AA-4986-4BE7-B77D-ADA24B32DF02
by Alex Jones, on Flickr

6ACA108B-2088-4003-9F9C-C49169185A1B
by Alex Jones, on Flickr

D907F9C4-F347-4191-A458-F66A343BAFAC
by Alex Jones, on Flickr

The turbine electronics have their own individual boxes with the ECU and CTU inside. The Deans connection is just temporary and I will be replacing these with IC5 connectors which are more suitable for anything with the possibility of getting wet. They have proved to be excellent on my Son’s RC cars.

2819349B-6FA4-4F5F-A149-BFE0ED65AF8E
by Alex Jones, on Flickr

One last test of all the electronics and peumatics before I split the fuselage to work on the electrical connection splits. There are twenty cables running through to the front which need to be easy to connect, secure and splash proof.

641E02A3-CC9B-41CA-BA9A-E69564499586
by Alex Jones, on Flickr

0FE5D92F-7699-4B47-8532-1F292D5AFBE5
by Alex Jones, on Flickr
 
#72
Hi Dave,

These are the fuel air traps, one for each fuel system.

With the fuselage split I can get on with what seems to be a never ending list of jobs for the back end. Starting with the auxiliary air intakes. These are prototypical in that they do feed into the main intake. They remain open until the aircraft is >15kts and >60% thrust, when these conditions are met they close. These aren’t required from an operational point of view so are purely a scale feature although a little extra intake area won’t hurt. We expect this model to be relatively slow in the water at high power settings until the skis un-port and then get up onto the plane. We will play around with the close settings so it looks nice and scale during the takeoff run.

The intakes are constructed from various 3d printed components using SLA and SLS printing along with carbon hinges and actuators connected to some small MKS servos mounted at the top of the housing.

B675BF8F-778E-4939-AFB2-5C315DF60963
by Alex Jones, on Flickr

A733FEF0-AE19-4C19-920A-04183F94B9BD
by Alex Jones, on Flickr

8B375AE3-7211-4FED-9D41-21E2A8B4B14A
by Alex Jones, on Flickr

9062EFA7-DB4D-452B-9049-AC26815A3C4A
by Alex Jones, on Flickr

A video showing their operation during a dry fit before we had the model painted.

 
#73
The intakes have been causing me a headache as they are very difficult to remove. They also require the turbines to be removed which very quickly got on my nerves. During the build these need to be installed and removed frequently which wasn’t time efficient and more importantly I need to push forward the intake just in-front of the turbine to get the vertical stabiliser bolt in place. With all this in mind I decided to split the intake again over the tank hatch. This allows me to easily remove the intakes and easily maintain the fuel tanks and ski actuators without spending hours removing all the intake system.

I made some 3D printed intake rings with a seal grove and split the forward intake as shown.

6D78E78A-2B6F-49F2-9124-C968B37B8A2D
by Alex Jones, on Flickr

2DC45718-1A0B-41B7-8BC5-0907A2BA49E3
by Alex Jones, on Flickr

Now its just a matter of removing these bolts on this section and I can access everything I need quickly. While at the lake I only have to do one side to facilitate the installation of the vertical stabiliser which speeds things up dramatically.

819DBF90-C5C2-40F4-A665-3C8B11A2297F
by Alex Jones, on Flickr
 
#74
The tail wheel on the SeaDart was a castor so it could be easily manoeuvred by ground crew while on its beaching gear. I decided to make a fixed version as this could be easily integrated into the structure. The beaching gear wheels are located forward of the CG so it does have to takes its fair share of the load.



0335B8CC-67D5-4FDE-AA4B-32F356307A23
by Alex Jones, on Flickr



3mm carbon was used to tie the wheel into the keel beam.



EB7B785B-BD30-441B-8B85-BFB81546F791
by Alex Jones, on Flickr



IMG_4113
by Alex Jones, on Flickr



Slotted over this carbon structure is a carbon composite pre-preg fairing bonded in place.



485AAE50-FCEA-4098-8999-0762A6E2F8D7
by Alex Jones, on Flickr



The wheel is Aluminium with a 3D printed tyre running on a Stainless Steel shaft on Phosphor Bronze bushings. On top of this we made 3D printed fairing that simulates the look of the castor fairing.



E6AE5AB4-37DF-4EC8-9D12-FFD7777A80BB
by Alex Jones, on Flickr



With a working tailwheel we can have the aircraft on the scale beaching gear and in theory it can move down a slip way into the water under its own power. Like the real one there will need to be a rope attached to the beaching gear so it can enter the water at a sensible speed.
 
#75
While things are easy to access I tested the fuel systems on both tanks. It’s a good opportunity to check for any small leaks, it beds the pumps down especially as they can run at full thrust voltage and lastly I can fine tune the pump factors so the fuel state telemetry is accurate. It is not always necessary on models to do this as access is easy and it’s not a big job to change things although this isn’t the case in this instance. Usually leaks appear at higher voltages and if you’ve purchased a used turbine pay close attention to the barbs on the pumps as people like to cut tubing off with a scalpel. This often leaves very small indents on the barb ridges and it will leak on thrust setting above idle.



Following the theme of splash proofing anything electrical I have used high capacity RC car batteries in this case Overlander 7.4v 2S 5300 Lipos. These have worked well on our ARRMA cars which get drenched so should cope with this type of use well.

They are held in with a 3D printed latch, you just squeeze the red parts into the battery and it releases.



1FC90CFF-E03A-48DB-8622-B9A2FBC9406E
by Alex Jones, on Flickr



I’ve mentioned before we are using CTU’s from Digitech and the twin turbine dashboards works well. To make use of the fuel telemetry the pump factor needs to be calibrated. After running fuel through each pump I then adjust the pump factor so the fuel remaining reads zero as the main tank starts to empty.



A short video shows the result… The pump factor settings are found in the CTU menu.






3328429A-6733-4BF5-8F7D-1BAEA8D103B4
by Alex Jones, on Flickr



I normally only fill the fuel systems on my jets at a rate equivalent to full thrust, It can no doubt cope with more but this way I’m sure I won’t over pressure the tank.



1A0D14DD-C199-4FDE-9E68-637F9F545D88
by Alex Jones, on Flickr



While Ive been working on this I have come up with a plan and designed some components that will allow us to easily assemble the fuselage halves and join the electrics easily and seal them from any water. The start of the process below.



34362D71-D36E-4949-A7F0-F639D9D142BD
by Alex Jones, on Flickr
 
#76
Like the 3D printed splash proof boxes for the electronics I decided to go a similar route with the fuselage electrical split. I designed two 3D printed housings either side where I am able to push the female PowerBox servo connector into the housing for an engineered fit. These housings are SLS printed nylon which are then tumbled and died black.



E04AE66B-364D-419A-87CC-6C2AA50F1195
by Alex Jones, on Flickr



67093BBD-6209-430F-ACA5-53D3CAB2B470
by Alex Jones, on Flickr



The rear section layout finished and tested.



554EB5BF-4ABB-43AA-B4B5-095389AC4DF3
by Alex Jones, on Flickr



This moves us onto the forward section which has 3D printed housings that slide and seal into the aft section housing thereby protecting the connections. To allow room to connect the wires I have around 200mm of excess cable which slides back into the long forward housing as the two halves are joined.



DF325C6A-E51C-45D8-828A-CE609648C2E5
by Alex Jones, on Flickr



SeaDart Fuselage Join Checklist
by Alex Jones, on Flickr
 
#77
The electrics in the forward section are run through 3D printed conduits that feed into the main box where the CB400 is mounted. There is another box on the bottom about 25mm deep that allows any excess wire to sit in, the bottom is capped with 2mm acrylic.

DEC08FE5-91E0-450D-A7A8-0677F3D89FCE
by Alex Jones, on Flickr

I ran all the appropriate cables to the fuselage split ready to test each one individually, as mentioned earlier these are plugged in and the excess wire moves into the housing as the fuselage halves come together.

396F7E26-E34F-4D4D-ABCD-3B2F1F428EE7
by Alex Jones, on Flickr

2B8C02AD-1200-42F9-86E4-D9DF013A8B48
by Alex Jones, on Flickr

9FA06F20-7172-4907-9D6F-979384D1B7EB
by Alex Jones, on Flickr

Everything was connected up and tested for the first time with all 24 channels needed. We have a second transmitter that controls the pilots head, canopy actuation and locking so it has 30 channels in total.

0D08A676-31F7-47D3-A75F-DB21F2780E18
by Alex Jones, on Flickr
 
#78
I’ve been putting off doing the Pitot for ages as it was one of those jobs that had the potential to cause some damage if I got it wrong. I started by making a hole in the top of the vertical stabiliser which had to be exactly in line and 35mm deep to allow the pitot to have a snug push fit into the stabiliser. This took me a long time!

3393456A-1EB5-4171-815B-3721EC6A6989
by Alex Jones, on Flickr

Once complete I used aluminium tube which will form the base structure of the probe. With the whole made (Phew!) and a straight cut along the stabiliser to mount the probe I traced the outline scanned this into CAD and started making the fairing. This was printed using SLA allowing me to get a nice finish on the component. A Jeti Pitot is mounted in the end and everything bonded together, painted and I now have a push fit working scale Pitot.

AD0E2319-8253-4248-A8D6-0915C00AA3AF
by Alex Jones, on Flickr

1B92F074-F60D-428A-B2D1-B392BD4FD66A
by Alex Jones, on Flickr

D037DE2A-63A5-487A-9862-4C922407D91C
by Alex Jones, on Flickr

32A1981B-F9AA-47BD-9884-BBA425240B90
by Alex Jones, on Flickr

Just the end to paint silver to complete the scale look and being removable it stands less chance of being broken off.
 
#79
Moving onto the beaching gear… This will be really important for us and was designed in partnership with PES Performance Ltd. This can work as a stand allowing us to test the oleo extensions/retraction retract the skis and run the engines while safely sat on the beaching gear. Then can be wheeled down a slipway into the water. It has hydraulic brakes when in the full 5 wheel version or get broken down to the scale beaching gear the full-size had. The two versions can be seen in these computer renders.



3EA6C98F-9757-4837-B422-37123F210F15
by Alex Jones, on Flickr



D50EDF52-F976-4313-B26B-7289BE930F83
by Alex Jones, on Flickr



With the designs and drawings complete it’s just a matter of machining the components. These are all ready for hard anodising… (bar the SS stuff)



DB4815A4-2602-4C4F-B6A2-3515633CB5DE
by Alex Jones, on Flickr



98C28C2D-F8EC-4AE0-87BE-1DB8E59B59B8
by Alex Jones, on Flickr



13CEA21B-693F-449D-99DC-2712561B7EE8
by Alex Jones, on Flickr



3AAE4813-9499-4352-880C-9CC2E86585A8
by Alex Jones, on Flickr



1EA88432-E2F3-4C68-ACD3-CF6DFE1FA4DB
by Alex Jones, on Flickr
 
Top