Josh's Projects

Hi all! This video should hopefully give you a rough idea of the things I’ve been working on, but I also wanted to write a bit more about the technical details of each machine in case you were interested, so I’ll go through them each below in reverse chronological order. Of course, you’re probably pretty busy, so no worries if you don’t have the time to read it!

PlyBot: This is my most recent build and the foundation of my company! I began this about 18 months ago, and have gone through 8 iterations, making it by far my most time-consuming project. I’ve long been on a search for the “Holy Grail” of desktop 3D printing: a simple, cheap design that would make it a household item. There are many great 3D printers, but they’re all simply too expensive, so I aimed to make one as cheap as possible by removing the expensive linear rails and using a SCARA arm mechanism instead. I found a few projects online that attempted to implement a SCARA arm mechanism into 3D printing, but all of them were abandoned in the early days of 3D printing (2013/2014). They had a lot of problems, so I designed my own novel arm system to circumnavigate these issues, and after working on my first iteration through the summer I finally got it working in November 2017.

I saw this as a potential commercial project from the beginning, so this wasn’t just an engineering project, but a product design project. Once the arms were successfully implemented on the first machine, I spent the next seven iterations (with an eighth in the works as I type!) attempting to convert it from a school project to a commercial product. This began with making it look clean and simple, such as implementing a fishing line-drive Z axis, meaning there’s no visible drive system (like a leadscrew) that would obscure the logo, making no wires visible, and making it all self-contained with nothing exposed that doesn’t need to be. I made machines of all shapes and sizes in order to determine the optimal height/width ratios etc, and settled on the one that looked the best. Most of my time was spent making everything snap-together, and shortening the assembly time as much as possible. Everything from the hooks that are used to snap-assemble it to the radius of the fillets were experimented with and carefully selected. I was basically attempting to make the Google Cardboard of 3D printers – flatpack, simple and easy to use.

For a person with access to a laser cutter, the MDF version of the machine costs £49.22 to build (using Aliexpress parts – free shipping), making it by far cheaper than any other 3D printer. The cheapest on the market at the time of making the machine was £99, and this was with no profit – I spoke to the retailers in their London store, it’s just to get initial customers so they come back and buy a more expensive machine. This means that my machine costs around half as much as the other machines, even without manufacturing in bulk. This is mainly due to the implementation of the SCARA arm system, which subsequently results in a very low part count: Of the £49.22 one-off manufacturing cost, 74% (£37) of it is electronics, meaning that for a cheaper machine to be made without sacrificing speed (cheaper electronics would sacrifice speed), the mechanical assembly would have to be made for less than £13. As this is pretty difficult, I think I’ve made it pretty much as cheap as I can!

With this project I won the title of UK Young Engineer of the Year, and that’s when I met my business partner, Ian, and we began to change the design for the new company direction. Whilst I’d initially aimed to target this to techies and 3D printer enthusiasts, with Ian on board (and some investment!) we now had the power to aim for regular consumers, so the aim of PlyBot shifted from a simple, flatpack kit for 3D printer enthusiasts to a fully-assembled, WiFi-enabled, smartphone-controlled desktop consumer product.

I naively thought that the initial size (30x30x50cm) was a good size, but looking back I have no clue what I was thinking! I shrank the design to about 60% size and then 60% size again, determining that the middle size was closest to the optimum size, aesthetically and performance-wise. After this, we decided to shift away from plywood, and design it to be injection-moulded instead. This brings us to the current day, where I’m working with Brook (explained in my additional info section – he’s the guy that founded the big consumer 3D printing company!) to develop the machine to be mass-manufactured at a low price. The main difference with the new design is that it’s mainly plastic and the arms raise up on rods with a static bed, rather than the arms being in a fixed position with a lowering bed.

We’re also seemingly (emphasis on “seemingly”!) close to raising a round of funding over the next few weeks, which will give us much more money to develop before the impending Kickstarter, as well as trigger my move to California, which I’m extremely excited about!

DesktopWorkshop: This is a combined 3D printer and milling machine. The glaring difference between milling and 3D printing is the speed and torque, with 3D printing being fast and low-torque, whereas milling has to be slow and high-torque. For this, I used a rack and pinion drive for the XY axis with a switchable gear ratio, enabling the milling to run about 3x slower than the printing. I experimented with many different methods for switching the gear ratio, but all of them either had backlash issues or were unreliable. Eventually, I moved back to the original design – bearings pressed against the bottom of the rack can be loosened, the larger/smaller pinion placed on the rack, and the bearings retensioned. This wasn’t the most elegant solution, but it did the job! I also ensured that it was rigid, with the Z-axis sliding directly onto the four cornerposts of thick angle iron, meaning that any flex in the frame won’t affect it nearly as much as if they were on separate linear rails. I also used four lengths of threaded rod as leadscrews for the Z axis, all connected to the driving stepper motor using timing belts, ensuring they’re all completely in sync. The electronics and power supply were contained in the bottom base (made of slot-together MDF milled on my CNC router), and I aimed for the Z axis to be in there too, but it was pretty difficult without making the base double the height! The main advantage of this was the price, at under 200 dollars to make, and I did plan on publishing the files open-source online, but to be perfectly honest it’s nowhere near finished enough. I took a few shortcuts with the build that I never fully sorted out, leaving it with a few niggling issues, but all in all it was a decent machine!

CNC Router: I wanted a CNC router at the time, but they typically used expensive electronics and custom-extruded aluminium for the rails, meaning the machines cost a few thousand pounds. Instead of using extruded aluminium, I came up with a design using skateboard bearings riding on steel tubing, which cut the cost of the rails a lot. The electronics were the other major cost, due to the expensive drivers (50 USD per axis), and motors (40 USD per axis). The thing about CNC routers is they’re extremely slow, so they need high-torque, but low speed, meaning the sensible option was just to use smaller motors/drivers and a reduction gear. It sounds obvious, but it hadn’t been implemented yet, and with the declining price of these parts due to the increasing popularity of consumer 3D printers, these smaller motors/drivers were extremely cheap, making it a very sensible option. The plan was to manufacture a kit for this machine, with the customers sourcing the larger parts, the steel rails and the wood frame, for themselves. It would perform just as good as the £3000 machines, but only cost about £300. I still think this is a good opportunity, and definitely fills a gap in the market for a really low-cost CNC router.

Third 3D Printer: After the second machine, I wanted a really fast performance printer. There was a really interesting machine I saw on the internet, and I based my design on it. I started off using the same frame dimensions as it, and then using the same printed parts, and then everything else… So the end product was basically his machine – he designed it so well I couldn’t think of any improvements, so not much to see here from me!

Second 3D Printer: Whilst building my first printer, I saw many of the parts used could be replaced by much cheaper options, and built this machine to test those out. I mainly focused on the linear rails and the drive systems, as this is where the main cost savings could be made.

Linear rails:
On the Y axis, I tested steel rulers, which performed well but had a bit too much wear to be used anything past a bedroom project.
On the X axis, I used mild steel rods with V-groove bearings. I ran a program to move the carriage from one end of the rods to the other for a few days, meaning it would wear a consistent groove in the rods, and as there was a much larger surface area then it wouldn’t wear much more. Once I did this for a few days, it worked absolutely perfectly – success!
On the Z axis, I used aluminium L-channel. As the Z axis moves very little in comparison to the other axis, the aluminium worked absolutely fine, and didn’t really wear – success!

Control/Drive Systems:
On the Y axis, I used a rack and pinion. Although this is, of course, a standard method, it was such early days for the industry that I think I was one of the first, if not the first to use it on a 3D printer. This worked great, and there were no issues at all – success!
On the X axis, I used fishing line. I had a fair few issues with this. First off, it’s incredibly slippery, so I had to experiment with many pulleys, determining a simple M8 bolt to be the best! This wasn’t perfect though, and over time the line “creeped” in one direction, meaning that the print came out slanted. I fixed this by making it home the axis (recalibrate where it is by triggering the endstops) after every layer, but it was still a bodge job and I wouldn’t do it again!
On the Z axis, I used standard threaded rod, so nothing to see in that aspect!

First 3D Printer: As I explained in the video, this was my first machine, and being quite young I had no money. I salvaged everything from old paper printers and made a deal with my dad that if I could get it to a state that looks like it would work then he’d buy the electronics and extruder for me (~130 USD at the time). Most of my research was online forums, YouTube and Google Images, to be perfectly honest! I made a lot of mistakes and worked through three prototypes before finally getting this one working. I still regard this as my proudest achievement, as it was such a big jump from all of the other projects I’d been doing, and it seemed so impossible at the time.

Thanks for reading this description of my projects, although I’m conscious you’ve got a lot of applications to go through so no worries if you skipped to the end! On the off chance you want to see more about these machines, I wrote some detailed posts with photos/videos a year ago just beneath this post, so have a scroll down if you’re interested! The PlyBot is still very much a work in progress so this is changing every week, but I’ll email the admissions department when the Kickstarter launches. Lastly, after watching the video I’ve realised I come off as pretty dull – it was tricky to remember what to say, so I hope I’m less boring in real life!

Thanks again!

Josh

After returning from the Swiss Talent Forum I was set on launching a startup and identified low-cost 3D printing as my best bet. It was the obvious choice, as I had been building low-cost 3D printers for years and the only thing driving the cost down was Chinese manufacturing, with almost no change or innovation to the design of the printers.

After running through BOMs of common 3D printers I came to the conclusion that the costs of the vitamins (linear rails, belts, motors, electronics) alone were way higher than I was aiming for, therefore these would have to be compromised. I could either compromise the electronics (opt for slow, cheap steppers) or change the mechanical system to eliminate the linear rails. Using cheap steppers wasn’t an option for me as they created a terrible printer (with a max speed of 10mm/s), so I opted to eliminate the linear rails. After countless hours researching online and emailing creators of different machines, I was set on using a SCARA XY axis, however, all of the ones that had been used before had at least one fatal flaw that made them unusable for a 3D printer, with most attempts being abandoned by the creator. After experimenting with a few approaches I settled on the one I use now, which works great. All the parts were then designed in LibreCAD and either milled on my CNC Router or laser cut.

The Plybot had 4 main goals:

• Cheapest 3D printer ever.
• Pack flat in order to ship easily.
• Use only off-the-shelf parts and CNC milled plywood.
• Easy to assemble.

Cost: The design satisfied all of these aspects, with a BOM for a one-off coming to £52.04, and at larger quantities, it is expected to be less than £40. The only printer with a cost close to it is the ‘Cherry 3D printer’ at £52.64, however, this uses extremely cheap stepper motors that run extremely slowly (10mm/s) and break frequently to reduce the cost, and if I was to use them my cost of a one-off would be £34.77. The cheapest commercial printer costs about £90 to produce in bulk (I spoke to the resellers of it), so if mine was produced commercially it would still be quite profitable to price it below this printer.

FlatPack: The printer packs into a 320x320x60mm box (about the size of two twelve inch pizza boxes stacked on top of each other), so is extremely compact and therefore cheap and easy to ship/transport.

Parts: The PlyBot is completely made out of off-the-shelf parts and CNC milled 6mm and 12mm plywood (aside from the acrylic bearing holders, I couldn’t for the life of me get plywood to work).

Assembly: The frame takes about 20 seconds to assemble and is completely snap-together with no hardware used. In all, it takes probably less than 10 minutes (I put a lot of effort into minimising the number of screws and part count in general).

Speed: The printer has very little moving mass, and therefore the motors can accelerate the print head much faster, achieving higher print speeds and therefore faster prints. I didn’t anticipate or plan too much for the print speeds, as price and aesthetics took priority over it, however, I was pleasantly surprised when I bumped the settings up and the printer kept up. My current settings are Jerk:30mm/s, Acceleration:9000mm/s^2 and Max Speed for Rapids:300mm/s. This makes it one of the fastest printers out there, and therefore the bottleneck of printing would be the extruder, not the speed at which the machine can move.

Other: The printer has a build area of 190x190x165mm (although for the next iteration it should have 200x200x200mm), uses 1.75mm PLA filament, runs off a 12V 5A power supply and print quality is comparable to most commercial printers, with there being no visible differences between the PlyBot and my school’s Makerbot (at 50x the price).

I struggled a lot with many issues, including the snap-fit frame, snap fit bearing holders and extruder, having nothing obscuring the PlyBot logo, stopping the arms sagging, making an ‘invisible’ Z-axis drive, working out the inverse and forward machine kinematics and integrating them into the preexisting firmware, using only plywood (no other custom parts whatsoever), reducing visible screws, cable management and most annoying of all getting my £4 knockoff hot end to stop jamming.

 

One of the main aspects that I focused on was the aesthetics and user experience. When customers buy the cheapest product on the market they’re not expecting a sleek, beautiful product, therefore I couldn’t make the product aesthetically pleasing by using high-quality products and expensive production techniques. I opted for a raw, no-frills, functional aesthetic that above all else was clean. By minimising non-essential parts, using some unconventional design and taking great care to only place parts in certain places I personally believe that I achieved this. Some examples are:

Using fishing line with the motor mounted on the Z-axis, routing it in the same place as the Z-axis rails in order to declutter the front:

Attaching the extruder mount behind the arm mounts, making it invisible from most angles:

Routing all cables around the backboard using zip-ties resulting in pretty much no wiring being visible from the front, and also very tidy at the back:

Mounting the spool holder on the base, using a Teflon tube to route it into the extruder, which also results in minimal exposure to air, meaning that it is less likely to pick up debris which can clog the hotend:

Using a completely snap-together frame, resulting in no fasteners being used.

Using as little bolts as possible (25 in total) by having multiple uses for most of them and using alternative methods for as many fastenings as possible.

Using a power-brick style power supply instead of a traditional power supply, which also gets rid of the issue of live mains leads and also certification as they are certified by the manufacturer.

Getting rid of endstops. Endstops are a luxury, and I argue that they do more harm than good for a beginner. They were frequently an issue in almost all of my printers, (both mechanical and opto-endstops) and can be removed by simply positioning the machine to a certain place and clicking a button. This ties into my belief that a simpler machine is better for the user. Many users aren’t techies and struggle with all of the bells and whistles of a complicated printer (LCD’s, WiFi, auto-calibration, SD cards, etc), so it made sense to eliminate them in favour of usability, and an extra minute of setup before a print is much better than hours of trawling through forums (and leaving negative reviews) trying to fix issues. I also would have to either use mechanical endstops with very visible wiring or optical endstops which are quite expensive and as price and aesthetics are my main focuses on this machine I decided against it.

It’s hard to think of other methods I used to achieve this as they gradually became integrated into the design as I reiterated, but I gave almost all aspects of the design significant thought in order to declutter it and overcome many unsatisfying features. To put it in perspective a quick google of cheap 3D printers bring up many massively cluttered machines showing how much of a focus aesthetics was to me.

What Now?

I currently have two fully working machines, with the newest one tailored towards being manufacturable. My Father (an accountant) and I have run through the financial aspect of producing the machines commercially and came to the conclusion that it is definitely worth pursuing. I have run the idea and business model through as many people as I can (including two people who have run Kickstarter campaigns) in order to gain advice for what is to come, and plan to possibly launch a Kickstarter this year with 3 of my classmates. I already know the financial side should work out, and that it should be fairly profitable if I can get the sales. My plan now is to try and drum up as much attention/followers/fans as possible, and then further investigate if people would buy it based on the attention gained. If it is a positive experience and I conclude that it is worth doing then we will pursue it, however, if not then I will probably move onto my hotend concepts and experiment with them.

Future:

I have a couple concepts for changes to the hotend of the 3D printer, both of which could be extremely disruptive and profitable if they work in my opinion. I’m beginning to experiment with them now, however, most of the development of these ideas will likely be at university, hopefully in an advanced hackerspace somewhere.

The DesktopWorkshop is a combined 3D printer and milling machine with a switchable gear ration to change to and from 3D printing and milling modes. I built it as I needed to develop PCBs at home, as well as mill parts in different materials (namely aluminum).

3D printers and milling machines also share the same common components, so it made no sense for me to build another machine that would take up more room in my bedroom.

Aims:

• Be very affordable.
• Use only easy to obtain parts.
• Mill aluminium (and anything softer than it).
• Fast change from milling to printing.
• 3D print.
• Fit on a desktop.

Price: It was very affordable at £115, mainly because I used a different XY mechanism. It used rack and pinion for the XY axis, with the axis being driven on one side by a stepper motor and a pinion on the other end keeping it aligned. This negated the need for it to be driven on both sides and meant I didn’t need to use expensive linear rails.

Parts: The machine used only hardware store and eBay parts, making them easy to obtain for anyone wanting to build it. The rails can also be 3D printed (one of my earlier iterations worked perfectly with them, but I upgraded to metal ones for the Big Bang Competition). The only issue is that I used my CNC router to machine the frame, however, it can just as easily be made using rectangular parts instead of interlocking parts.

Milling: The machine mills anything softer than aluminium perfectly, and can just about do aluminium. I attempted it only once (it had to be locked up in school for moderation for months afterwards) with the worst conditions possible. I used an ALDI ‘Dremel’ with the flex shaft attachment which had a fair bit of runout, an 8 flute bit (when it should’ve been single-flute), all haphazardly attached to the machine with scrap wood and clamps. Despite this, it milled it decently, but I have no doubt that if I had a proper set-up it would mill it perfectly (albeit at a low speed).

Change Speed: It can be changed from milling mode to 3D printing mode in about 20 seconds, however, it doesn’t have an easy method of changing from 3D printing to milling easily, so this takes about a couple minutes extra, making it a failure in this aspect.

Printing: It prints perfectly, but doesn’t have a heated bed so can’t print in any other material aside PLA.

Size: It is roughly 350x350x500mm, so whilst being fairly large it fits on most desktops.

Overall, it was a decent machine. It had accomplished all of the issues I set about to resolve on paper, however, in practice it had a few underlying issues that needed more time to resolve, but with exams coming up I had to abandon the project. After exams I was too excited about the PlyBot to fix these issues, and I don’t expect that I’ll fix them anytime soon.

UPDATE for 2019 university application:

I highly doubt you’ll see this, but on the off chance that you’re interested enough to look, I’ve included a short explanation video below that was initially intended for a science fair. There was a time limit for the video and I kept messing up and going over it so this was probably the 50th attempt at it (with a dry throat and not much enthusiasm!) so I come off as super bored!

Most of my projects use sheet material as the main material. My usual way of making parts out of sheet materials was to design on graph paper, stick it onto the material using a glue stick, roughly cut it out using a jigsaw and then sand and file it down to shape. This was a very tedious method, and I saw it necessary to build a CNC Router in order to automate these tasks and also enable me to make more intricate things.

You’ve probably noticed a common theme with my machines now: they’re all cheap. This machine is no different, and I had to go about this with a budget of about £150. I used skateboard bearings riding on steel rectangular tubes and

The machine was a massive success, costing £60 for the mechanical assembly, £30 for the router and around £70 for the electronics (although I did need to replace a faulty stepper driver). It has a build area of about 500x500mm, and a Z-axis depth of about 50mm, which is perfect for milling sheets up to 18mm thick. I have used this machine extensively, with all of the parts for the Plybot and the frame for the DesktopWorkshop being milled on it.

My 1st 3D printer was a success, and right after getting it working I had moved onto my next printer. I had identified that the convention was to overbuild printers, for example, there is a commercial printer that can mill aluminium if a spindle is swapped out for the extruder, which I find ludicrous! I wanted to experiment with a few methods of getting the cost down:

Axis Drives: The printer uses braided fishing line and an M8 bolt instead of timing belt and pulleys, which worked pretty well. I did, however, find that the fishing line slipped over time, so I introduced a few lines of G-code after every layer so that it homed itself to reset its position. I concluded that fishing line could be used effectively, however, it would have to be fixed in place on the pulley to ensure that it wouldn’t slip.
The rack and pinion drive worked very well, and the only issue that I had was how difficult it was to get it perfectly parallel to the motion of the axis. The main issue is that the axis has to be built in such a way that it limits the range of movement, so if it was used it would compromise the range of motion.

Linear Rails: I saw the v-groove on flat bar as an interesting idea and wanted to test it. I used steel rulers for the X-axis and aluminium angle iron for the Z axis. Both have stood up to thousands of hours of printing with almost no wear.

This printer was a fantastic workhorse and aside from the small slipping issue it was a perfect machine.

In around 2012/2013 there was a massive hype around ‘The 3D Printing Revolution’ in the media, with Bre Pettis and a MakerBot adorning the front of most tech magazines for a few weeks. It was at this time that I became very interested in 3D printing and set about building one.

Nowadays you can buy a 3D printer online for less than £200, however, back in 2013 there weren’t any low-cost 3D printers, with just the essential parts costing hundreds. Being only 13, I had about £100 of birthday money to spend, and therefore had to be more inventive about where to gather the parts. I bought a set of 5 broken printers on eBay for £12.50, which I dismantled and gathered stepper motors, linear rails, belts and a host of other things to be used in the machine. I constructed a frame from wood and threaded rod and after a few months had completed the mechanical side of things. My father then bought me the electronics kit, extruder and hot end off eBay (the majority of the costs) after I proved to him that it worked, and after a week of resolving many software and electronics issues I got it printing.

I did have other videos of it printing, however, this was the only one I uploaded as it was the first print.

Before I became interested in 3D printing I tinkered with many things. I began by making weapons when I was a very small child and quickly branched out into making anything that I thought was cool at the time. This included many bombs, airguns, knives, video game props (multiple ‘Hidden Blades’) a machete, a sword (all heat treated using my mum’s hair dryer blowing into a charcoal fire in an old Wok). I branched out into electronics, tinkering with tasers, capacitors, Tesla coils, and many pranks involving these. This electronics knowledge was also put into good use by connecting it to Xbox remotes, enabling me to perform repetitive functions (buying and selling a Fifa item for 250 and reselling it for 300) and also helping me with in-game gameplay (rapid fire modification for Call of Duty and a Fifa modification enabling me to perform tricks that require a lot of dexterity). Most of this has faded from my memory, however, every once in a while I come across an old project in a dark corner of my garage, reminding me of some bodge job I did when I was 7.