In this article I want to show you my personal workflow and tooling for designing and manufacturing a micro quadcopter frame from start to finish. I will show you the steps I take, things I consider during the design process, the tools I use, hard- and software and some tips and tricks to make things easier.
I will show you all of this based of a real world example, my latest frame design. Designing frames is a lot of fun and a great challenge. This article is meant as an introduction and my intention is to show you that you too can have your own frame and that it does not have to be as complicated and expensive as you might think.
Disclaimer: Obviously I am not a professional frame designer. I like to tinker with things and customize them to my liking. I will not describe the road-map to a commercial product. The information here is rather intended for hobbyists who want to do all of this from the comfort of their home.
- Concept and Goals
- General Design Principles
Concept and Goals
I first start out with a concept and try to answer the following questions:
- What do I want to achieve with my design?
- Which problems am I trying to solve?
- What kind of improvements do I want to make?
- Which parts do I want my design to be compatible with?
In my latest design I wanted to replicate the EMAX Tinyhawk 2 Freestyle - one of my most favorite micro quads of the past year about which I raved in my review.
The things I want to achieve and improve are:
- Enable to mounting of 11xx motors with a 4 hole, 9mm mounting pattern
- Intended for 2.5” props
- Be able to use a toothpick sized AIO flight controller instead of the proprietary form factor that EMAX uses
- Be able to run a dedicated receiver
- Be able to run a dedicated video transmitter
- Flexibility for a different range of nano sized FPV cameras - although they all have the same width, a lot of them differ in length and I want all off them to be protected by the cam mounting plates.
- Top mounted battery
- Be able to use off the shelf M2 hardware and standoffs
- Same - or at least - similar weight as the EMAX Tinyhawk 2 Freestyle. Obviously this is not only dependent on the frame, but keeping the weight of the frame in-line with the original one is a good starting point.
To have a clear idea of what you are trying to achieve is extremely important since it will answer a lot of questions that will come up during the design process.
The parts for the frame will in the end be made from carbon fiber sheets - 2mm thick in this instance. For prototyping, test fitting and size comparisons I will 3D print all the frame parts in PLA.
You do not necessarily need a CNC router for cutting your carbon fiber, but I highly recommend you at least have a 3D printer to test everything out before sending it away for cutting.
3D printed TPU parts will also be involved: the mount for antenna, video transmitter and receiver. Instead of the 3D printed TPU parts you could also use zip ties and double sided foam tape, but since I want everything to be neat and clean I am opting for the slightly heavier TPU parts.
To quickly iterate the designs and check the fit, screw holes, slots and general sizes, I simply print the parts in PLA. I print in draft quality (fast speeds), 0.2mm layer height (I use 0.2 because it turns out to be more precise in height than using 0.3, especially when the height are “full” mm) and the pieces are usually all printed pretty quickly - the bottom plate on this particular design prints in a bit over 30 minutes.
Regarding the printer - I am personally a big fan of the Prusa printers. But basically any printer will do here. Should you not have a 3D printer yet, the Ender 3 V2 is a pretty solid entry level option that can be bought on banggood for about 250$ and is also a great option if you not only want to print PLA, but also TPU and other materials.
I generally think, that if you are into the quadcopter hobby, you sooner or later want to invest into a 3D printer, unless you have a friend you can outsource the printing to. It makes things so much easier if you can print your camera mount and protective pieces at home, not having to wait for them being shipped to you.
Obviously you could skip 3D printing your parts and simply print your designs on paper to compare and fit them this way. But having a physical representation close to the end-product will greatly help you to get a feel for the real thing.
I personally use a MPCNC - mostly printed CNC to cut my carbon fiber plates in a water bath. The water bath is to contain the carbon fiber dust which you should absolutely not breathe in. This machine is not in any way comparable to a professional CNC, but provides enough accuracy and features to cut your own frames at home without exploding the bank - I bought the materials for mine for less than 400$. You might already have some of the stuff, so it could easily turn out cheaper than that.
The working area of my DIY solution is 36x36cm in the water bath and about 55x55cm without the water bath. I purpose build this machine for cutting frame parts and wanted to be able to cut a uni-body frame for a 5” build. I am not really limited with the height of the carbon fiber sheets - tried everything from 1 to 5mm and since I am using passes of 0.5mm there are absolutely no issues with even the thicker sheets.
A semi-professional machine will easily run you 10 times the price. Not including a dust handling solution.
Using the DIY solution limits me in a couple of ways:
- The smallest end-mill I can use is 2mm in diameter, everything smaller than that tends to break during routing, even with super slow feed-rates.
- Tool changes are only possible in very limited ways. The spindle I am using has a collet which is a bit cumbersome to open and swap tools. I always tend to lose my position and everything is then slightly off. My workaround for this is to simply use a 2mm end-mill for all operations. Technically, you really should use a drill for holes, but it works well enough for me not to bother - a 2mm diameter end-mill is 2$ - so I might need to exchange end-mill more often, but it allows me to get my carbon parts done quickly.
For a waste-board I use MDF - I can stick the carbon fiber plate to it with double sided sticky tape and the MDF itself to the water container. I also tried a particle board, but it soaks up the water really quickly, and the double sided sticky tape does not stick to it at all.
I additionally use small 3D printed clamps to hold the carbon fiber sheet down to the MDF board. This comes in handy when multiple parts have already been cut and helps the carbon fiber plate to stay flat.
Probably the best solution for a waste board would be some kind of a plastic board - but I could not find anything on the cheap to experiment with. I would imagine Delrin or similar materials could be quite a good choice.
MDF seems to be a good middle ground for a waste-board: Not too expensive, does not soak up water too quickly (it eventually will) and is readily available from the hardware store. I go with 240mm x 340mm x 12mm high MDF board, that runs me about 0.9$ each. Obviously you should chose your MDF board according to the dimensions of stock you are using.
I tend to cut all parts in one session and then use a shop vac to suck off the water from the water bath. This way I can stretch the longevity of the MDF a bit since I do not let it sit in the water for longer periods of time. I can usually do two or three sessions with one waste board before it is soaked up and warped too much to be usable. And after two sessions I usually only have small pieces of the carbon fiber sheet left that I cannot really use for anything useful.
To finish up the cut carbon and get rid of the tags I highly recommend a small needle file - a set with a couple of different shapes can be picked up on Banggood for less than 10$. Also a sheet of fine sandpaper is recommended - I use 120 grit - to give the edges a nice finish. This should obviously also be done in a water bath or under running water. Just make sure not to breathe in the carbon fiber dust.
Some people seem to like to put a lot of time into finishing their frame, some even paint the edges. For me it’s simply not worth the time, I’ll crash it sooner or later (probably sooner) anyway and after a couple of sessions it will not look pristine anyway. But I have to admit, they do look super sexy after being freshly built.
Depending on your tooling you will obviously need different materials:
- PLA: If you want to rapidly prototype with the 3D printer
- Carbon Fiber sheets: This is where it gets interesting - there are a million options of different weaves, thicknesses and colors. The cool thing about carbon fiber, it is really easy to work with and to machine.
Choosing the right Carbon Fiber Sheets
First of all: the top and bottom layer is only decoration and the weave does not tell you anything about it’s internal structure. Sometimes the decorative layer is representative of the underlying ones, but it does not have to be - this is something that a lot of people get wrong and which always triggers me a little bit.
The best way to learn about the internal structure of the carbon fiber parts is to bend and twist them.
With quadcopter frames we want most of the strength to be along the arms. This makes designing unibody frames a bit tricky, especially if you are not going for a uni-body true X design, where the arms are offset exactly 90° to each other.
How are Carbon Fiber Sheets made?
Carbon fiber sheets are a compound of carbon fiber fabric layers - also called enforcement - stacked and fused together with resin. The carbon fiber fabric comes in different shapes and forms:
- Unidirectional: All fibers run in the same direction.
- Woven: Fibers are intertwined with each other in multiple directions, bi-axial but also tri-axial (and more, although not very common) is also possible. Common angles hereby are 0°/90° and -45°/+45 degree for bi-axial and -45°/0°/45° for tri-axial. The degrees are an indicator how the fibers in the fabric are offset to each other.
Another spec to look for is the weight (g/m²) of the carbon fiber fabric. The lower the weight, the thinner the layer and the more layers you will need to reach the desired height of your carbon fiber sheet.
Those fabric layers are then stacked to the desired height and “glued” or fused together with resin. There are different processes of how this can be achieved, here a list of some of them that are relevant for us:
- Wet layup: every layer is stacked by hand, brushed with resin before the next layer is stacked on top.
- Resin infusion: The layers are stacked on top of each other and the resin is then infused under a vacuum, oftentimes in a bag, but it could also be in a mold.
- Pre-pregs: The carbon fiber layers come pre-impregnated with resin (usually on one side), are stacked and pressed to achieve the intended thickness. The resin gets activated above a certain temperature so those pre-preg layups are also often cured in an autoclave/oven and under pressure. Carbon fiber parts manufactured this way are the most temperature resistant ones.
Those stacks are then cured either at room temperature, in an autoclave or oven. Oftentimes a vacuum is pulled before initiating the curing process.
The carbon fiber sheets you buy off the shelf have - with a high chance - been manufactured by some kind of resin infusion method or using pre-pregs.
If you want to learn more about different manufacturing methods for carbon fiber (and general composite material) parts, I highly recommend this in depth article on compositesworld.com.
The fabric layout, weight and the way direction it is stacked define the strength of the resulting carbon fiber plate.
A very common layup pattern is UD 0°/90°. This simply means that unidirectional carbon fabric sheets are used and each layer is offset by 90° to the previous one. This results in maximal strength when bending it in length and width.
It is also strong when twisting it in diagonal, but then again it is fairly easy to twist in length and width.
You need to be careful when buying your stock in order to get what you are looking for. Usually the layup pattern and obviously the height is specified.
UD 0°/90° is probably the most common layup used with quadcopter frames and carbon parts for RC in general.
Other layups are also available and custom layups can be manufactured perfectly designed for your application.
The finish are the top and bottom layers of the carbon fiber plates and is basically only aesthetic. There are different options: matt, glossy, different colors and different weaves. Choose whatever you like the most.
Although the colorful ones look really nice, I found them to be a bit more involved when it comes to post processing since the colorful interweaves do not seem to get cut cleanly - at least on my CNC and you need to put in a bit more time into the cleanup.
What is 3k Carbon?
3k refers to the amount of fibers in a tow of woven carbon. In the case of 3k carbon, there are 3000 fibers in one tow of carbon. Other nominations are available too, starting from 1k all the way up to 48k. The smaller the tow, the thinner the resulting fabric will be.
3k is what you know as the “classic” carbon look and is usually the weave that the top and bottom layer will come in.
Learn more about the k rating of carbon fiber fabric on estcarbon.com.
Before starting out with your design you should make sure that all the hardware you are planning on using is actually available off the shelf - if you are planning on using standoffs, make sure in which lengths they are available, also check their diameter.
When it comes to screws I like to use as few different screw lengths and sizes as possible. This is pretty straight forward with micro frames, since you will basically use M2 for everything.
I prefer to go with screws that fit a 1.5mm Hex screwdriver - this is the same size that is also used for mounting the props on motors with a T-Mount. I also make sure that all the screws can be operated with the same tool - there is nothing more frustrating than needing 5 different screw drivers to put a frame together
When it comes to mounting the flight-controller, or stack, I prefer to go with one long metal screw that goes through all the boards instead of having to deal with nylon standoffs, that for sure will break during the first couple of crashes, and are very cumbersome to replace. A screw will in the worst case bend which can easily be swapped after the session, but you will still be able to fly in most cases.
General Design Principles
There is a couple of things you should keep in mind when designing a quadcopter frame which is independent of software used.
Avoid Sharp Corners
Sharp corners are where the stress accumulates, try to avoid sharp corners by rounding them off. Sharp corners are often the place where a frame will break on impact. This is something that you can easily test and verify in a FEM analysis.
You can clearly see that the maximum stress on the part with the rounded corners is less since it is distributed over a larger area as compared to the sharp corners.
Carbon Fiber Directionality
While designing keep in mind which carbon fiber directionality you are working with and along which axes the carbon fiber is the strongest. This is especially important when it comes to the arms. Make sure that the fibers run along the arms or you have some struts in place that will enforce the arms accordingly.
This can be especially tricky when working on a uni-body design where the arms are not detachable.
I want to emphasize this point, since this is pretty hard to test via FEM simulations - so you will have to account for this manually.
When making screw holes, for example for the motor mounts, make sure to leave enough padding around the screw holes. Too little padding around the screw holes can easily lead to the screw being ripped out during a crash.
While designing your frame, one of your top objectives will probably be to keep the overall weight down. Still you need to keep in mind that not all measures you take to keep the weight down are actually worth it. Cutting slots might save you a couple of grams, but it might also add additional stress to your frame - here again FEM analysis will help you to make decisions if you are doing the “right” thing.
Know the manufacturing limits
When designing your parts you have to keep in mind that you cannot have 90° angles inside of your part. All the pockets will be cut with a round tool, so all of your holes and cutouts need rounded corners. The radius depends on the smallest tool you can use. For me personally this is a 2mm diameter end-mill, so the smallest radius I can use for corners is 1mm.
This is also important for slots - if you want parts to slot into each other, you need to either have extra cutouts at the end - accounting for the diameter of your tool, or on the sides.
Try to answer the following questions before you start your design, otherwise you might end up with something that can not be manufactured:
- What is the maximum possible part size?
- What is the smallest tool radius that can be used?
- How deep can the tools cut?
For designing my frames and 3D prints I mostly use FreeCAD - this offers all the tools that are needed to design a frame from start to finish, from CAD (Computer Aided Design) to CAM (Computer Aided Manufacturing), all in one package and available for all operating systems. FreeCAD might seem a bit intimidating at first, but it has a great Wiki and a super helpful community, where your questions will be answered quickly and a lot of tutorials to get you started.
I chose FreeCAD because I work on Linux and this seemed like the one stop shop I was looking for. You might also want to take a look at Fusion 360, solid works and EstlCAM - just to name a couple of alternatives.
From all the workbenches that FreeCAD offers I mainly use the following ones for designing the carbon parts:
- Part Design: This workbench allows you to extrude your sketch to a certain height
- Sketcher: Here you can design 2D parts which are then easy to extrude
- Part: This workbench allows you to make boolean operations with your parts, for example cut two bodies or make one body from two
- Spreadsheet: If you want to have parametric designs, I would highly recommend using a Spreadsheet. You can define values here, and when referencing them in your sketches you can easily change them all in one place later without having to go through all your sketches
- FEM: Here you can simulate and analyse stress applied to your design by adding forces to it and see where the areas with the most stress are.
- Path: This workbench is used to generate the tool-paths for the CNC router. If you send off your design for cutting, you will probably not use this workbench. If you cut yourself, this is a workbench you will spend a lot of time in. Once you go through this process you will understand why getting your parts cut will be more expensive than just the cost of the material.
For the 3D printed parts I also tend to use the Draft workbench, which is extremely useful for connecting multiple bodies to each other.
Instead of explaining each workbench in detail I link them to the Wiki, explaining the functionality of each one in detail would go above the scope of this article (and might also out-date quickly), but I will give you a couple of general pointers what to look for in each workbench and some pitfalls you might fall into.
This is up to you, but I strongly suggest starting out with a good project structure. It will make it so much easier to find everything once you come back to your project at a later point. I group everything in folders and create a “Part” for everything in advance.
I tend to set up my project in advance, just to have a clean starting point.
The part design workbench is the workbench I always start out with for each part. It will basically guide you through everything you need to get a new part going:
- Create body
- Create sketch (I usually use the XY plane, just so that everything is aligned properly for creating the paths)
Once you are done with your sketch, this is also the workbench where you can “pad” your sketch to a certain height in order to get a 3D body representing your part.
The sketcher workbench is the workbench you will most probably spend the most time in, when designing 2D parts.
Do not try to fit your whole design into a single sketch. Although this is possible, the constraint solver gets pretty slow if you have more than 150 or so constraints. I tend to have the base shape as one sketch and all the holes as a separate sketch. Unless it is a small part, then I keep everything in one sketch. You know it is time to start a new sketch when placing new constraints starts taking too much time.
From the beginning of your design process make sure that everything is properly constrained, this will save you a lot of work in the end.
My first sketch is always a “Helper Sketch” - a simple drawing of props and FC placement, the rest of my parts are based on that. This way I can quickly see if anything is blocking my props or interfering with the flight-controller.
The part workbench allows you to combine multiple bodies into one. You will use the boolean operations here a lot, for example to cut your main body and the hole body to get a resulting body with all the cutouts.
The spreadsheet workbench can be used to set variables which can then be used throughout the whole project. Cells can be named and can then be referenced accordingly. I tend to give the columns names and this way group related settings to each other. You could also use different Spreadsheets, but I usually have only a couple of variables so it is not really worth the overhead for me.
I like to use variables for everything where tolerances matter, like arm length, slots length and widths. This way I can easily update them in one place instead of having to go through each sketch to do so. During the design process you have to keep in mind that you might want to change those values at a certain point, so you need to be careful with your constraints, otherwise changing variables in the spreadsheet might result in unexpected behavior and a lot of debugging.
FEM (Finite element method)
The FEM workbench allows you to run stress test simulation and will show you where your design has stress points upon which you can improve. It is also a great tool to compare designs against each other.
This workbench was for me the hardest one to grasp until I found this video explaining everything in 10 minutes.
Obtaining the Weight of your Part
The weight of a part can be calculated if you know the volume and the density of the material. The density of carbon fiber sheets will vary, and you will have to run a calibration against a real part, but 1700kg/dm3 worked great for me.
To make things super simple, there is a FreeCAD macro that can be installed from the Addon manager, named “FCInfo” which allows you to select bodies and it calculates their weight based on a given density.
The path workbench is where you will prepare the tool-paths for cutting the carbon. You will only need to do this if you are planning on cutting the carbon yourself.
Different output formats are possible here, but the most common one - that you will most likely use - is GCode.
The biggest tip I can give you here ist, that you have to be careful with contour operations that would produce an island in your part - you need to add tags, otherwise you are running risk, that the island will get lose and mess with the rest of your operations - in the worst case this might lead to your end-mill breaking - so be sure to add a tag dress-up to your contours. Tags are small pieces of material that are not cut away, but instead left to connect your piece to the rest of the stock. In the end you need to cut those tags manually. I use a big flat-head screw driver to cut through the tags.
Even if your part does not have islands, you will need to add tags to the outer contour of your part.
When it comes to the order of operation I start with the simple ones, holes and slots, then move on to the islands and as the last operation I cut the main contour of the part. This way the carbon fiber part has the most integrity and stability till the absolute end, thus minimizing the risk of the part getting loose.
This workbench also allows simulating the operations, which I would highly recommend once you are done, just to make sure that you really got all the required paths. Nothing is more frustrating than finding out that after cutting, that you missed one hole.
When it comes to feed rates I prefer to be conservative, but on the save side. I use 3mm/s in Z and 8mm/s in X and Y. I turn up the spindle to max and go 0.5mm in depth per pass. Also I cut 0.5mm deeper than the stock actual stock is, just to be on the save side that I actually clear the whole carbon fiber sheet - so the last pass basically cuts 0.5mm into the waste board.
After you cut your parts or sent them off for cutting, the great fun part begins - testing. Build your copter and go out and fly - crash it, find out where the weak points are and start improving them.
Failure is part of the design process and although there are a lot of tools that will help you improve your design, the real feedback and behavior will only be seen in testing. Do not be discouraged when your first design explodes on the first impact - failure is part of the whole experience.
Just as a reference, I 3D print the parts about 3-5 times before I have a part I am really happy with and that fits the way I want it to. This then usually results in a good carbon cut, but oftentimes even the carbon pieces need to have a revision, so before going all out and cutting multiple frames, test one and see if it is ready for prime time, this will save you a lot of time, money and frustration.
I would also highly recommend you do some black box logging and analysis to help you understand if you have any frame resonance that might pollute your gyro data.
In my opinion designing your own frame can be a lot of fun, very challenging but also very rewarding. A lot of people do seem to like to design 3D printed frames but don’t take it the next level to the carbon frame.
Sure, 3D printed frames can - and should - utilize different techniques that cannot be achieved with carbon parts, but do not shy away to try designing and manufacturing a carbon fiber frame.
As mentioned in the beginning, I am by no means a pro, so if you have any tips or tricks, please let me know and I will expand this article accordingly.
Chris is a Vienna based software developer. In his spare time he enjoys reviewing tech gear, ripping quads of all sizes and making stuff.
Learn more about Chris, the gear he uses and follow him on social media: