PDA

View Full Version : CNC help required for final year Uni project



ChristoT
18-10-2014, 01:14 PM
Afternoon!

I know I've already posted this up in the RFQ section, however at JAZZCNC's suggestion, I'm throwing the thread up here so people can discuss it.

For my final year project, I am working as part of a team on a UAV for an international competition: the BMFA Payload 3 Challenge. In essence, the team of four from my University have to design, build, test and compete with a UAV with a 2m wingspan, carrying up to 4kg of dead weight payload. All powered by a single tiddly e-flite10 electric motor.

To me, the answer is obvious: make it out of carbon fibre. Thanks to my placement (http://www.e-goaeroplanes.com/), I can make carbon fibre parts and their moulds. However, we need patterns. I'm going to approach several firms to ask for sponsorship for tooling block (hopefully we'll get some from Trelleborg), but my team needs a CNC shop who can mill the block for us to our CAD models as the Uni's equiment is simply too small. Just to give an idea, each wing will be just shy of a metre long, and will need 2 moulds (and hence 2 patterns) for each. We're going for a radical aerodynamic shape with a blended wing and body design (http://www.wingco.com/atlantica_design.htm), which means we will need comparatively fewer moulds for the entire aircraft.

Before someone suggests sandwich composites (foam wing, CF skin) - that's too heavy for our purpose. Last year's winner had a more powerful engine, and a dry weight of 960grams. We're throwing every weight-saving technique at the job here, and the wing will most likely have several chord length ribs, a single, central spar, and a single CF skin over the top. Strong enough for our purpose, and exceedingly light!

If someone can help us, but can't manage 1m long patterns, it's less critical, as my team can always pin them together, and make a mould that way. The main concern is getting the wing and fuselage shapes out of the 3D CAD package, and into a tangible object we can take moulds or parts from.

What I'm looking for is a firm or individual who can sponsor us with the CNC work. We need 10 patterns (8 if I can design the fins symetrically); or if the quality can be assured, and if each pattern can be made full size in one piece, we could skip the pattern stage and go straight to moulds - something that would save my team a lot of time, money and effort for the same work done by the CNC firm. We have a tiny budget though, and I'm lucky if I'll get back the fuel for the car, as I'll inevitably play taxi for raw blocks and finalised pieces!

What do you stand to gain by this? Hopefully if you've read thus far, you're interested in the project! Unfortunately, due to our budget, I highly doubt we can offer financial reward or payment for your work. However, you'd be working with a University who lead the way in contests such as Formula Student (although admittedly our current BMFA record is less stellar - no composites yet, that's why!), and you would gain advertising and publicity from your markings on our UAV. If someone were to support us, and the partnership was a success, that might encourage the Formula Student team, and others maybe to use you too - not that those necessarily need to be sponsored too!

If anyone can help, please let me know! We're hoping to have the design finalised by mid November, so we'd be looking at around then for the moulds/patterns.

Thanks in advance!

magicniner
18-10-2014, 06:46 PM
Lost foam doesn't look like your solution, apologies, it does look like you need to get someone else to help make it for you,

- Nick

JAZZCNC
19-10-2014, 10:49 AM
Ok well I asked before what material but on reading again I can see your thinking to use resin tooling block.

What time frame do you have for this because 10 large high quality and accuracy 3D toolpaths will take many hours of machine time even for a industrial machine. I think you will be hard pressed to find any machine shop who could afford to tie up a machine for the length of time needed for these and do it FOC.

charlieuk
19-10-2014, 10:52 AM
I dont think i would totally dismiss the foam thing some of the super light but still strong gliders i used to fly were based around a glass/carbon d box and trailing edge with very light carbon caped ribs. with then the main area of the wing covered in mylar. You could very easily do something similar by hot wiring cutting you wind in blue insulation foam then vacuum bagging it with glass or carbon and then machining away pockets to leave a rib type structure. That way you wouldn't have the expense of creating moulds and could try different designs easily and while carbon sounds great it is also 5 x the price of glass as i have found in the past you can often get away with using very small amounts to get the required effects in weight and strength and substitute the rest for standard glass keeping the costs down.

Do a search for highlight rc glder and you will see the design i mean which will give you by far the strongest and lightest structure suitable for what you want.

ChristoT
19-10-2014, 12:29 PM
Hi Jazz,

Looking again, I reckon I might be able to make the fuselage and fin moulds in house, meaning only the 4 1m long wing moulds need outsourcing.

That would probably make more sense, as my current project plan is looking at a fortnight for the pattern manufacture, and another week for mould manufacture. If someone can machine each wing moulds in one hit, that's three weeks for the 4.

Charlie, I am very familiar with the technique you describe, as that is how e-Go Aeroplanes (where I was on placement) lay up their wings. They use controlled polystyrene cores, with large sectional cutouts. They have to wet layup, though, which isn't ideal due to weight considerations. Whilst you can cure pre-preg outside an autoclave, the stuff we're going for will need autoclave treatment due to the resin/fibres ratio. Obviously, polystyrene can't take autoclave temperatures! We are looking at a wing loading of about 3.57kg/m2 - which is nothing. Makes more sense to get the weight down early, especially as the electric motor we're provided with is, well, pathetic.

During the design phase, we'll be running a simulator alongside the design, tweaking as required. Another reason why we need the moulds to be spot on - as the BwB has no horizontal stabiliser, it is nigh-on impossible to trim. Unless the centre of gravity and centre of lift line up, you have an unstable aircraft. So you have to add ballast weight to trim, and your advantage of a high-lift, low weight aircraft start slipping away...

We are trying to scrape every single gram off - the contest is highly competitive. Last year, the teams were using the more powerful glow ignition engines (about 2x as powerful), and the winning entry weighed under 1kg - nearly half the weight of my University's design. If we want to win, we need to be one of, if not the, lightest plane on the flight line.

ChristoT
19-10-2014, 12:29 PM
Bother, wrong button. Attempt 2!

ChristoT
19-10-2014, 12:30 PM
Ok well I asked before what material but on reading again I can see your thinking to use resin tooling block.
I think you will be hard pressed to find any machine shop who could afford to tie up a machine for the length of time needed for these and do it FOC.

Who says it has to be one machine shop?

charlieuk
19-10-2014, 01:27 PM
Hi Jazz,

Looking again, I reckon I might be able to make the fuselage and fin moulds in house, meaning only the 4 1m long wing moulds need outsourcing.

That would probably make more sense, as my current project plan is looking at a fortnight for the pattern manufacture, and another week for mould manufacture. If someone can machine each wing moulds in one hit, that's three weeks for the 4.

Charlie, I am very familiar with the technique you describe, as that is how e-Go Aeroplanes (where I was on placement) lay up their wings. They use controlled polystyrene cores, with large sectional cutouts. They have to wet layup, though, which isn't ideal due to weight considerations. Whilst you can cure pre-preg outside an autoclave, the stuff we're going for will need autoclave treatment due to the resin/fibres ratio. Obviously, polystyrene can't take autoclave temperatures! We are looking at a wing loading of about 3.57kg/m2 - which is nothing. Makes more sense to get the weight down early, especially as the electric motor we're provided with is, well, pathetic.

During the design phase, we'll be running a simulator alongside the design, tweaking as required. Another reason why we need the moulds to be spot on - as the BwB has no horizontal stabiliser, it is nigh-on impossible to trim. Unless the centre of gravity and centre of lift line up, you have an unstable aircraft. So you have to add ballast weight to trim, and your advantage of a high-lift, low weight aircraft start slipping away...

We are trying to scrape every single gram off - the contest is highly competitive. Last year, the teams were using the more powerful glow ignition engines (about 2x as powerful), and the winning entry weighed under 1kg - nearly half the weight of my University's design. If we want to win, we need to be one of, if not the, lightest plane on the flight line.

if you are carefully doing a wet layup on a layup table and then vacuum bagging with all various extras and not just mylar a long with then removing a majority of it through pocketing you can achieve results incredibly close to a moulded or autoclaved product at a fraction of the cost and you can still post cure with xps or eps to get every last bit of strength.

For what you are doing carbon for the entire build will not only be expensive but also over built and heavy for the purpose. we used to do builds comps using micro cars for all the components which were like doing small scale payload challenge as the were heavy brushed motors and copper coil actuators, the size you can cut everything down to get the strength to fly if you relay try is crazy we were using something like .5 mm carbon caps on wing ribs were still ott a lot of the time

While moulding is great and you can get some fantastic results if you have a large budget and a proven design if you are doing a one off or prototyping it is a expensive way of doing it to get the same result as perhaps a little more time but significantly cheaper and faster method that allows you to be flexible with your design.

ChristoT
19-10-2014, 01:50 PM
The pre-preg I hope to source hardly weighs anything, because the amount of resin mixed into the fibres is the bare minimum for structural strength. A single CF spar down the middle, with ribs in, say 4 positions made up of 3 ply pre-preg, with a single carbon skin over the lot will weigh mere grams. It's an expensive method, true, but one I am familiar with. Not to mention the fact that the University has an autoclave on site we plan to use!

Truth be told, I am not as confident working with foam: all my experience at e-Go was either working with moulds, or applying patches in wet layups for small areas. It's also worth mentioning that as each wing will be 1m long, and has fins and rudders mounted at each end, there's going to be a certain dynamic load on the wing in multiple directions (certainly a torsional element from the rudder input), one which I'm not sure the foam can take.

Wet layup over a pocketed foam core certainly is a process worthy of consideration; in fact, it was the initial plan for the wing. However, due to the loading, I figured that the spar, rib and single skin method would work out substantially lighter. If no CNC sponsor steps forward, that may well be the process we have to go for, out of necessity rather than choice.

Anyone here able to hot wire cut foam? :highly_amused::highly_amused:

charlieuk
19-10-2014, 02:39 PM
if you are worried about twisting ext from other aspects then when you do your layup you are going to have to orientate your fibres in more than 2 directions to stop this which will mean multiple layers in places. i think you will find a carbon skin for the hole wing will be over kill and just add wait. try to only use carbon were you need the strength ie in the spar and in the dbox possibly. you can purchase carbon T section that will provide you with more then enuff strength for your needs and only way a few grams and then its just a matter of crating the wing shape as light as posable so using a lighter weight material like mylar were you are just need to cover the structure. There was a program by guy martin not long ago about building a human powered aircraft and they used the same foam mylar technique

Hot wire cutting foam can be as easy as a length of stainless steel wire and a 12v battery and a bow of some kind and then just two templates at ether end of the shape you want you also can get a range of densities of eps ranging from just under 1lb cubic foot to 3 or 5 so lots to choose from.

EddyCurrent
19-10-2014, 05:56 PM
@ChristoT, did you look to see if there was a FabLab near you, that might be another way to do them yourself.

http://www.fabfoundation.org/fab-labs/

ChristoT
19-10-2014, 09:35 PM
Didn't know about them, but now I see there's one in London - thanks! Now to see if any near me have a CNC machine big enough. :eagerness:

ChristoT
24-10-2014, 09:23 AM
Annnnnnnnd... No luck there. Their machines are too small for our purpose. The University has a mid-size Denford machine (I believe it has a a 500*250mm deck - something in that order of magnitude anyway), but we're trying to avoid mucking around with 6 different sections - as it is, the fuselage will probably take between four and 8 sections of tooling board to make!

Jonathan
24-10-2014, 11:08 AM
On a similar note to Eddy, have you spotted Hackspace? e.g.
https://london.hackspace.org.uk/

Have you got a drawing / image you could post showing what you want machining, or just something similar to show the general idea? My machine is big enough...

ChristoT
24-10-2014, 11:37 AM
Yep, come across Hackspace, but I don't think their machine is up to size.

The wings will be in two moulds (top and bottom), and be around 1m long each. Not sure what the chord is yet, that depends on the wing designer. The shape of the mould will be the negative of the wing, so a effectively a long curve. A quick google throws this up :
http://static.rcgroups.net/forums/attachments/5/4/6/7/8/t2952713-129-thumb-WingMoulds1.jpg?d=1261911797

The only difference I can see is that in that example, material is removed around the shape of the wing - we'd be looking for the shape of the wing to be removed from the billet instead so that we can layup directly in it.