Skiing on Jet Power

The first time I punched the throttle, there was no explosive launch, no violent shove. Instead, I just started sliding forward, smoothly, quickly gaining speed faster and faster. Before I knew it, I was flying past everyone around me. I had strapped 16 kilowatts of electric jet thrust to my arms, and it was exhilarating! This project had lived in my head for years, and I had just made it a reality.

Like many makers, I grew up watching inventors like Colin Furze strap jet engines to all sorts of things, pushing ideas to their absurd limits. Maybe this is why I have always had a fascination with going fast. I’m sure this is a common thought for a lot of makers; how to take something normal and push it to the extreme. Well, that was me with skiing. As soon as I had enough confidence to go fast downhill, there was always this lingering thought in the back of my mind that I could make something to push the boundary on what it meant to go fast on skis. I sat on this thought for years, wondering if it was possible and how I would approach it. Jet engines always seemed like the logical choice, but I always ended up thinking it would be too impractical. Thinking something isn’t practical doesn’t answer the question of if it is possible though, so one day I decided to put pen to paper and see if it could be done. Time to do some –

Research

All projects start with research, and this is no different. The search began with looking at JetCat micro jets. These are true dead dinosaur powered metallic thrust producers. However, when looking into the feasibility, it was clear it just would not work out. The engines would be ingesting a lot of foreign objects being close to the ground mounted to my skis, and they would be too hot to have close to my body. The auxiliary systems also seemed like a lot to work around. Fuel lines would have to be routed across my body, as well as high voltage wires for the ignition system. All exposed and close to the ground, just waiting to be snagged on something. Did you know that snow is just frozen water? Did you know that exposed high voltage wires like to spark on conductive things like water? Did you know that a spark can ignite flammable liquids leaking out of a snagged fuel line? Yeah, a lot of complexity, and complexity is synonymous with a safety nightmare. So I kept looking for another solution.

I have seen electric ducted fans (EDFs from now on) used on projects. If you aren’t familiar, they are basically really powerful drone motors put in a tube with a high blade count propeller for moving lots of air. They are cool, but they never wowed me… they just never seemed to stack up with my past experiences of what a real jet engine could do. That being said, EDFs come in a wide range of sizes and power, and after looking at specs online, it became apparent that my opinion of them was tainted by people choosing smaller-sized EDFs and commercially available batteries with limited output for their projects.

EDFs would be great if they could work out. It removes a lot of complexity, and also means that I don’t have to worry about heat, thus I can mount them to my arms. Keeping the EDFs away from the ground solves the foreign object ingestion problem, but it also means that I don’t have to “re-learn” how to ski. With the thrust acting on the skis, turning or stopping might require a change in technique, meaning my existing muscle memory would fail me. However, with the thrust acting on my upper body, the mechanics of skiing shouldn’t change, and that’s a huge benefit that can’t be ignored.

I built a spreadsheet of all the largest 120mm diameter EDFs. Any bigger and you get into engines meant to go on gliders that carry people, which also means the cost jumps by quite a bit due to certifications. I quickly narrowed it down to just a few choices based on two main criteria; thrust per dollar, and thrust per watt. Both of these are important metrics. A motor with a good thrust per dollar but bad thrust per watt means a bigger and more expensive battery, and will probably offset any cost benefit that the motor itself has.

In the end, I settled on the Ejets Jetfan-120 ECO with a HET 800-74-590 motor, an EDF that produces 20.5lbs of thrust (9.3kg) at 53.9 volts and 124 amps (or about 6.7kW). Doing the math with the peak voltage of the battery I planned to build, that changes to 58.8V and 135A, a whopping 8kW of power per motor! That’s just shy of 11 horsepower per EDF! In other words, the equivalent of having a Honda Grom motorcycle under each of my arms as far as power output is concerned. Now we are talking! This seems like it just might work!

I did some quick math in my notebook to check the validity of my intuition. I found some coefficients of friction for skis on snow by searching research papers online. I then plugged the range of values I found into two equations. I wanted to see what my velocity would be after 10 seconds, as well as the holding angle where the EDFs perfectly counter the force of gravity. The reasoning here is that any mistake in one equation should become apparent if the answer to the second one didn’t seem to match. With the velocity at 10 seconds, I got 21 mph for an average friction, and 32 mph for a best case. Similarly with the holding angle, I got 5.7 degrees average, and 8.6 degrees best case. This passed the sniff test for me.

As a final check, I turned to the online maker community and asked them to fact-check my work. It had been a handful of years since I had graduated university, so I was a bit worried that my physics was a bit rusty. Thankfully, someone realized I was mad enough to do this project, and it wasn’t just me asking for homework help. They confirmed my math, which gave me enough confidence to start ordering parts. Big shout out to the online maker community. It really does take a village, even when it looks like it’s just one person building something cool.

More powah!

The first challenge was testing all the parts after I got them. It’s the small things you don’t think about, but how do you test a ~60V motor that pulls 8kW? I have a lab power supply, it can output 60V. You know what it can’t output though? Yep, 8kW.

Chances are you are familiar with lithium-ion (Li-ion) batteries already. These batteries power your laptop, power tools, and even electric cars. They win out on energy density, or the amount of energy you can store per unit of volume or weight.

The next contender is LiPo, or the batteries most commonly used on drones and RC aircraft. LiPo batteries win when it comes to high output applications. They can dump way more power at once than a lithium-ion battery, but they don’t have as much total energy. A way to think of this is that our Li-ion battery is a 2L bottle of water with a small hole in the cap, but the LiPo is a normal water bottle with the cap fully removed. The smaller LiPo water bottle could dump more water at once, but it has less total water than its 2L Li-ion counterpart. This would probably be the battery to go with for a quick YouTube video, but I wanted these to be usable for a full day of skiing. Oh, also, LiPo batteries seem to light on fire a lot easier than Li-ion, which is something I would like to avoid.

So, the first thing I tackled was the battery. There were three main battery chemistries I wanted to look at, Li-ion, LiPo, and LiFePO4.

Last, we have LiFePO4. These are what you might find in a home power wall, or at a solar farm for energy storage. They can hold a lot of power, but they end up being heavier than Li-ion for the same amount of energy stored. They also have a slightly lower cell voltage, so I would be leaving some performance on the table. The main reason I was looking at them is because the battery cell I was considering (the Headway 38120) has screw terminals, so building a battery pack would have been a lot easier with them compared to spot welding.

In the end, I went with some 18650 Li-ion batteries. Samsung 25R batteries to be exact. I settled on a 14p, 14s battery configuration. This means that the assembled battery has blocks with 14 batteries in parallel, and then 14 of those blocks will be linked together in series. A total of 196 18650 cells. The final battery should have about 10 minutes of full-throttle runtime, which given the above speed calculations should be plenty! To be honest, in retrospect, I wish I pushed the batteries a bit past their rated amperage output and went with a 14s12p configuration. The 10-minute runtime was plenty, but the battery weight is just borderline bearable, so shaving 28 18650 cells off the battery would probably have made it just a bit more comfortable to wear.

Contrary to the LiFePO4 batteries that had screw terminals, the 18650s required spot welding. The rating on the thickest nickel strip I found was 10 amps. I think my math might have been wrong, but I calculated that a 30s burst would heat the nickel up 100s of degrees C for its measured resistance. I did find some copper strip that was nickel coated, but could not get it to spot weld with any spot welder I tried. Instead, I found that you could get flat braided copper wire (sold as copper drain cable for grounding electronics), and spot weld that to the nickel strip to both lower the resistance, as well as act as a heat sink. This worked wonderfully, but was very time consuming (hence the mention of the screw top batteries). I haven’t seen it done before, so I thought I would throw it out there for anyone else trying to make really high output batteries.

Also, spot welders heat up FAST. I ended up using two cheap spot welders so that I could trade off and let one cool down while not interrupting the spot welding flow. Don’t underestimate the benefit of some additional equipment for large builds like this!

Anyway, with the battery built, I could actually power the EDFs up and get to testing!

It’s all just LEGOs. Very powerful LEGOs.

The actual build is fairly simple. The battery has an off-the-shelf battery management system. This connects to an off-the-shelf electronic speed controller (a Flyfun 160A v5 ESC, the cheapest one I could find). This could talk to an off-the-shelf RC receiver, though that’s not what I did. Anyway, my point is that at this point, it was smooth sailing, and it’s incredible how plug and play most things are now a days.

I wanted the controls of the project to be integrated, and not use a separate RC aircraft remote. Because of this, I deviated away from RC airplane land, and moved into the more comfortable (for me) micro-controller territory. My one complaint is that RC aircraft parts often speak RC aircraft lingo in their docs, and don’t have details you might want if you are DIYing your control scheme. For instance, the PWM frequency or value range for the throttle was not listed anywhere online or in the physical manual the speed controller came with. This took a bit of time to figure out, but eventually what I realized is that these speed controllers expect you to calibrate them every time you turn them on. You input max throttle on power up, it beeps, you back off to 0 throttle, it beeps again, and then it will actually turn on and start processing throttle commands. I found this out the hard way when my testing caused the ESC to start ramping up the EDF’s speed while on my workbench. The EDF ate the paper manual I had open in front of it, filling my room with an ultra-fine paper dust that took hours to fully settle out of the air! Luckily, as mentioned prior, the manual didn’t have any useful info in it for microcontroller-related integration anyway, so nothing important was lost. I did however get the opportunity to learn how paper both smells and tastes. Anyway, with the ESP32 microcontroller now being able to control the speed of the EDF, I just needed some way to tell the ESP32 how much throttle to command.

I went with a knock off Wii nunchuck controller. I absolutely love them as far as projects go. They are cheap, fit very comfortable in the hand, have a built in gyroscope, and speak i2c. If you don’t want to do the i2c handling yourself, there’s even an ESP32 library specific to the Wii nunchuck that handles it all for you! Just call wii_i2c_decode_nunchuk(data, &state); in a loop, and then use state.y to get the value of the joystick Y axis. This is mapped to the PWM range that will be output to the speed controller. In total, the code for this project came out to 60 lines, most of which are comments, or additional things like using the trigger on the nunchuck as a safety switch you have to hold down for the ESP32 to send the throttle command. If you go with an ESP32 + a Wii nunchuck, check out https://github.com/moefh/esp32-wii-nunchuk – I had no issue with it, and it made the integration of the nunchuck a breeze!

Lets get to the nuts and bolts of it!

Similar to the electrical and control side, the mechanical parts of the project are all very straightforward. While working on the project, I talked to another maker you might be more familiar with, James from Hacksmith Industries. He mentioned not using 3D prints for any crucial connections between the EDF and my body. I felt stupid for not coming to that conclusion myself, as I had planned to 3D print the bracket, but it makes sense. I threw a test 3D print in the freezer to see how bad it would have been, and I can confirm that James was correct. Looking at it the wrong way caused it to break, so even a minor bump or fall could have been disastrous.

While I could have laminated some plywood together and cut a bracket out of that, I decided that having wood on the project just didn’t fit the theme I was going for, so I opted to make as much as I could out of aluminum.

The main backbone of the project is some extruded 3030 T-slot aluminum. The EDF mounts to that with a milled aluminum mount. In reality, using a service like JLCCNC would have been smart, but I am trying to grow my maker skill set and chose to mill the parts myself. The aluminum extrusion was drilled out and tapped on the end for an eye-bolt to be installed, which then attaches to a climbing harness via a locking carabiner. This acts as the main path of force, and is using components that are all way overbuilt for the ~20 lbs of force I expect them to see. This constrains the EDFs to my body, but it does not constrain their motion. At this point, if I powered them up, they would be flying around in circles chewing up anything they could suck in, while attached to my hips!

Because of this, needed some way to attach them to my arms. More specifically, an arm mount that would keep them co-linear with my arms. The only thing I found that seemed like a good arm mount was a range-of-motion brace meant for physical therapy. Sadly, these were on the more expensive side, and I ultimately ended up destroying the fancy range of motion mechanism to just get the lower arm portion of the brace. That said, it used aluminum for a bulk of the construction, and some robust feeling plastic for the Velcro arm loops. Good enough for dealing with secondary side-to-side forces!

I whacked up some quick L brackets to attach the range of motion brace to the T-slot aluminum, and the only thing left was mounting the electronics. This was fairly straight forward. I made some small 3D-printed enclosures, put some JST connectors on some prototyping PCB board, and stuck the electronics anywhere I could find space. A few thin wires for signal, and a few chunky ones for power, and everything was tested and working.

Jet skis (but for snow)

Well, with everything built and tested, the only thing left to do was head to the mountain! It’s one thing to test the thrust standing on solid ground with normal shoes, but testing it when standing on snow with slippery sticks under your feet is a whole-nother experience.

With me being the builder and knowing the numbers, I immediately had a sense of respect for the jets as soon as I turned the battery on and heard the ESCs chirp to life. Giving it a quick test fire, I Instantly felt myself slide forward. It wasn’t an abrupt shove or anything that would scare you, but your brain understands the power involved, even if you aren’t thinking about the 250 amps being delivered to your arms. I would say it elicits a mixture of excitement mixed with fear… the same feeling you get as you begin to lower the restraints on a roller coaster, or the feeling you get before jumping from a high diving board. As for the thrust, it’s an assertive but gentle acceleration that just continues to build. It doesn’t take a long burst of throttle for you to understand that staying on the throttle for more than a few seconds would have you going pretty fast.

While I didn’t have someone out with a radar gun, I can confirm that the ballpark measurements done in the notebook are fairly close to real life. As I mentioned, I do have a proclivity for speed. I have gotten up to around 68 mph going downhill on skis before (without the jet skis), and while the conditions were nowhere good enough to reach that kind of speed safely, I could immediately tell that a long enough burst of throttle, even on a more gradual green or blue slope (beginner to intermediate-class runs) could have broken my own personal speed record; a record I set on a black (expert) run that was perfectly groomed for a race that was happening off to the side.

Another fun takeaway from trying out the project was that I definitely could go uphill with them. I suppose the slushy late season snow I was skiing on had a higher static coefficient of friction, so getting uphill required a bit of a “running start”. More accurately, a jet powered start where I used the EDFs to get some speed on the flat part leading into the slope – but assuming I was moving, they did continue me uphill. Even on the steeper uphill areas where people were having to turn and walk sideways, I was able to become near weightless with the assistance of the thrust, and walk up with no real effort.

I also tested towing a second person. I attached a long strap to my backpack, and my safety buddy I was with was able to grab on and get towed along flat sections of the mountain. I’m sure he ate more than his daily recommended amount of atmosphere, sitting in the jet wash and all, but he seemed pretty happy that he didn’t have to put in effort to cross long flat areas (cat tracks, for those that know).

So, all in all, I think the project was a success. It did end up being pretty tiring having a 40 pound battery on my back all day, so I do think that a slightly smaller battery would have been better. The 14p configuration hardly heated up even with extended use, so I do think pushing the cells a bit harder with a 12p configuration would have been fine. But hey – you don’t know until you try!

I had a blast bringing the idea to life and using it. Ultimately though, it’s not going to replace normal skiing. It’s heavy, it’s loud, and that makes it very impractical. But that was never the point. This was an impractical idea that lived in my head for years. I just needed to know if it could be done.

Now I know.

I also happened to learn that sixteen kilowatts of thrust feels incredible. I would consider that a win.