Answers to common questions: - This prototype uses a 4-layer PCB from JLCPB. - We will use an IMU and onboard camera star tracker to determine what control outputs are needed at what times. - The whole thing will be made open source eventually. - My scripts are available here: github.com/manforowicz/Magnetorquer-Calc
What program did you use to display all of the graphs from python? Super new to electronics but I’m very curious. Was it just a .py insert? If so what was it? Thank you
This is awesome...both the ingenuity and the fact that we live in a time where manufacturing something like this for the home hobbyist is not only possible, but relatively cheap and easy. Amazing!
Interesting indeed! I'm doing a video series on diy cubsats and currently exploring the power budget requirements before starting the pcb layouts, the next video currently under research is magnetic torque thrusters and if they worth it for cubesat project. This helps a lot, and I'll certainly link this video in that video as it's a great explanation! 👍
super interesting research! good job! love the non-constant width idea, however i'm not sure whether eliminating the smaller turns was advantageous for the coil. The outer tracks are the longest so they are forming the largest part of the coil's total resistance
Thanks! Your great projects helped inspire this design! While the outer tracks are linearly longer, the area they encircle is quadratically greater. Outer tracks cover more area per ohm so they're more "efficient". If I wanted to fill in the center while keeping resistance constant I'd have to widen the traces which would reduce the number of coils packed along the edges. But I don't mathematically understand why this seemingly arbitrary inner-to-outer-radius ratio maximizes area-sum given a fixed resistance. (At least, according to my Python simulation)
This is really cool! I hope this technique can be used to make CubeSats more affordable in the future because ACS can be a big economic and space cost!
I find your videos fascinating, even if I can't quite follow it all (a flaw on my part). Your video on the Kelly criterion was broken down so well that I could clearly understand it. I tried learning about it in the past to no avail. If you need any ideas for content, I would love a similar video on gasseon processes.
Superinteresting, I always wondered, but never cared to research how reaction wheels deal with saturation! Also I rarely post but its great to see awesome "DIY"/scientific content coming from PL - keep it up! Given how little space you have on the cube sat - I wonder if such magnet could be reused for other things too when its operation is not required - e.g. a really bad antena ;)
Maybe measure the voltage over it to detect if the satellite is tumbling after release. I guess if the satellite is tumbling and you switch on the torquer it might not stop the tumbling as the force is small. But if you know the frequency and phase you can switch on the torquer only at the moment that it almost lines up with earths magnetic field .. and use less power. Because when the angle is 90 degrees the force is zero i guess.. and by turnining it on and off in a smart way you might be able to stop oscillation
Good idea, we might add a voltage sensor to detect the direction of rotation. Though our primary reference will be an onboard camera tracking stars. And yes, we'll need to create an algorithm for a microcontroller to turn them on and off at the right moments.
As the Kepler Telescope reactions wheels failed, it was able to continue science observations by using the unbalanced solar torque (the photons) as a third source of torque and kept pointing with only two RWs.
would it be worth making the same design but on a multilayer pcb? so you can have two coils on one pcb to have double the force in the same space? you could get more than two layers I know but I would assume there should be diminishing returns as the efficiency would probably decrease with all those magnetic fields in the same small space.
I have been thinking a lot about this, but never got to chance to try anything. I am happy that someone is working on it. I wonder if making orbital maneuvers using the Earth's magnetic field would be possible. Similar to how the head of a hard disk drive moves sideways, maybe we could use the earth's magnetic field to change the orbit of a CubeSat.
By the way, if you try doing this, I had the idea of using a material with different magnetic permissivity to create a sideways force. If you want to try it: sandwich the coil between two steel plates leaving only one side of the coil open (kinda like putting a blanket halfway). You should cover 85% of the coil with the steel plate and only leave one side of the square open. This should create a sideways force based on my thinking. But I might be wrong. I just wanted to try it a lot, but never got the chance to.
Since Earth is just a magnet, we could only apply a linear force directly toward or away from it. If you hold a magnet over the Earth, it doesn't attract it noticeably more than just gravity. The rotational torque on a compass seems to be much greater. Earth's magnetic field is only really useful for aligning with it, rather than getting pushed/pulled by it. But I may be completely wrong.
@@marcinanforowicz8319 That sounds right .. nasa did do experiments with the space tether from the shuttle. But the hardware failed. They wanted to reel out a really long conductive wire from the shuttle. and if current would run through there might be a force in the direction of movement. Not sure how a current would run through the wire. Problem is if you want to hook the tether up with wires to have current flow .. you end up with a parallel wire to the tether with current in the opposite direction. The magnetic field of that will cancel out the magnetic field of the tether so that doesnt work.
Nice project! You say it took 3 months for your essential chips to arrive, and ascribe this to your institution's purchasing bureaucracy. This is a terrible waste of the time of your talented researchers, but surprisingly common in my experience. Support functions in large organisations often forget that they are the tail, not the dog. Lean engineering eliminates the key causes of waste, including delay/waiting time. Your procurement function needs to get on board with this and keep pace with your research. You need to escalate this issue to your VP/Head of R&D to build accountability. You will probably find they are keen to find concrete examples to push forward change, and that they have strategies to work around until then! Like, buying on personal credit card and then claiming as an expense. Don't give up, demand change and accountability, and fight your corner! You are the dog, not the tail!
@@marcinanforowicz8319 You can get electrical steel (the kind of silicon iron used in transformers) from McMaster-Carr starting at ~200USD for ~0.65 m². At a guess a little more looking would get you something closer to the size you would use for a lot less.
Interesting proposition. I wonder which design would be better? Several squared coils stacked on top off each other or merely sandwiching a ferromagnetic material between two parrelel coils? Since once ferromagnetic materials are polarized, their inertial magnetic field would be resistant to the change of magnetic polarity once the squared coils reverse the direction of the current.
To get a performance measurement it could be cool to mount it on a torque spring in a vacuum and putting some AC on the coil. That way it will start to oscillate, and finding the forces is pretty easy if you know the spring constant and the masses.
Great idea! We could also calculate its moment of inertia (based on materials and dimensions), go to a location with known/undisturbed Earth magnetic field, hang it on a loose string, and see at what rate oscillations die down to deduce air resistance. Then we'd measure the frequency of oscillation, and factor all this in to get torque.
@@marcinanforowicz8319 This isn't really my area of expertise at all, but I just wanted to say that I absolutely adore the enthusiasm and excitement you're approaching the comments section (and this video in general!) with - I don't know quite why but it just makes me really happy.
@@marcinanforowicz8319 You could also apply a large known field with a coil externally, that way you have control about the field, and the increase in magnitude will make it easier to detect movements and forces.
Nicely done. Just a thought, could you use a magneto style spinning magnet arrangement coupled with a compound coil to create an enhanced large field that was both redirectable and stronger? I'm basically thinking of something like a flux guide but dynamic.
Once You enhance the system with an adequately chosen or good magnetometer or more (to achieve near instantaneous magnetic field 3D orientation information ), I would suggest you investigate the domain where You have the unit on the string here on earth (as You have shown) - and nullify the oscillations - critically dampened point (control system terminology) - again for this particular situation. This can develop s/w code of “system oscillation and stability”. Then translate the domain according to altitude planned. LEO ~50nT to 200uT depending on day/night, but of course with variations of angle! Worst case from a satellite control aspect is at sea level where gravity,air, string are so strong. Check out MEMS magnetic field sensors
There's almost nothing to dampen that oscillation in space and very little in air too. You could try to couple it with an IMU and adjust polarity / voltage to apply breaking and dampen the oscillation. It might actually need a computer vision camera to stabilize as the IMU's compass might be useless near that coil.
If they're willing to use reaction wheels, why not gyroscopes on twist axis's? With three gyroscopes with 90 degree opposed twist axis's and it should be possible to accurately allign to anything and it should provide a lot of torque as well. It could even be a modular external deployment and spin in the vacuum with no air resistance on magnetic thrust bearings where the flywheel is a wheighted PCB induction motor rotor with the stators on the outside. That'd turn the satellite. With magnetic bearings, if you made it like a pancaked squirrel cage, if it was never over-volted, it would never wear out.
Awesome video! I wonder if you added some soft iron and used flux switching to modulate earth's magnetic field instead of generate a field of it's own, whether it would be more efficient - great work either way!
Interesting idea. One of my peers actually suggested a mode of operation which cuts off power and keeps the spirals passively connected in a loop. This would effectively act as a dampener, slowing the satellite's rotation in Earth's field.
@@marcinanforowicz8319 If the frame of your satellite is conductive like aluminium then if the satellite tumbles then the magnetic field of the earth will induce currents which in their turn creates an opposite magnetic field which will dampen the tumbling i guess..
I was thinking about this, and actually changing the orbit of the satellite using the earth's magnetic field. Since the push doesn't need to be huge in the space, but constant.
Neat. How do multi-layer boards with similar patterns and powered similarly impact the result? Could you essentially do a 24 layer board bonding different layers to create a lot of specific directions of force? Or help push the desired induction one way or another?
Great question. AFAIK the shape of the field you produce doesn't have much impact - intensity is more important. 1) For some configurations, rods fit better spatially into the satellite. 2) Although a narrow rod has a lower area-sum-to-wire-length ratio, this seems to be offset: Physical wires can be crammed tighter together than PCB traces.
How about making a square section tube out of four boards that all connect together (with s something like castellation) and the traces on each of the four boards being straigh lines, open at both ends, to connect to the next board. Essentially building a square section winding on the outside of the cube sat. Using 4 layer boards, you could use the other layer for the magnetorquer and the other 3 for the sat's electronics payload. Or a large flex pcb, that wraps around the assembly to form a tube, with soldered edge contact, like an lcd flat flex
That is a very good idea, I'll look into it! It would probably increase area-sum and magnetic moment per weight. On the other hand, having self-contained magnetorquers on separate boards increases redundancy and ease of manufacture.
@@marcinanforowicz8319 yes, it occurred to me when I was thinking about it later yesterday that it does greatly increase the potential failure points. The other thing that occurred to me it that one could make the boards with the tracks on one layer in one direction and a second layer in a perpendicular direction, and arrange 6 identical board to form a cube with built in 3 axis magnetorquer. And at the same time, using the pcb as primary structure of the sat. Damn it! Now, I want to build one myself 🤣
@@AlfieMakes any sort of radiating electromagnetic coil will interfere with the rest of the equipment. No mater the shape, size and wether it is a homemade or a bought one. But so does the wire going from the battery to the electronics, and anything carying a current. That would be a choice to make depending on the satellite's mission. Of course, if you are trying to map the earth magnetic fiel in LEO, you wouldn't use any kind of magnetorquer on that mission, or you would have a calibration routine to work around
I wonder if you could just wrap winding wire around the whole cube? If the frame at the corners was made from threaded rod you could use the threads to space the wires to get uniform and consistent geometry.
Are you using any components that have radiation hardened alternatives? I also suggest using a white solder mask as this will increase albedo and lower operating temperatures if exposed to space.
Check Carl Bugeja's channel, he got some coils made of flexible PCBs, they're very thin; maybe there could some gain in weight/magnetic power ratio with something based on his approach?
Thank you for the great video. I have a question. I am trying to generate a magnetic dipole moment of 0.2 with my PCB. How was the total length of the PCB pattern calculated? And what is the current flowing through the pattern? Was the formula M=IA used to calculate it? I would appreciate your help with this
Calculating the optimal design and characteristics is a bit too complicated to explain here, but I might write a paper or something about it in the future. Feel free to use my optimizations scripts as a reference for now: github.com/manforowicz/Magnetorquer-Calc
I wonder if you can stack multiple PCBs as the added complexity of adding another pcb might not be that much compared to the added output force, if they can be added together
Heat generated is V^2 / R. This prototype is 10 volts and 100 ohms, so it generates about 1 watt of heat which is almost imperceptible. For comparison, a microwave generates around 1000 watts of heat. However, even 1 watt has to be carefully radiated away when in the vacuum of space. Adding layers while keeping electrical properties constant would increase torque while keeping heat constant.
Stupid question but why do you have to have a reaction wheel? Quadcopters keep their rotation to target by adapting the difference in rotor speed. They can rotate as much as they want by changing the difference in 2 rotating motors. So why not just add 1 more motor and get rid of the saturation issue for good?
Once an external torque has been applied, the satellite+wheels system has a net angular momentum. The wheels can't change this since they can't push off of anything external (like air, in quadcopters). They can only transfer some of this angular momentum from the frame to themselves. Similarly, an astronaut spinning in space can't stop without using thrusters or grabbing on to something.
Good question. I just assumed room temperature since the satellite will have to be roughly stabilized at that point anyways so that the batteries don't freeze or overheat.
@@marcinanforowicz8319 OK, I posted a similar idea without reading the comments, sorry. You are right that these PCBs are expensive. I checked some quotes at PCBWay and a 50mm x 50mm 4 layer 0.4mm PCB with 3/3 mil traces is about $300. If you have never ordered from PCBWay you can get a 2 layer 0.4mm PCB and order them for $5. Then you can stack them with something like thin Mylar spacers. This wouldn't be a production piece, but perhaps good enough for testing.
For one of our instruments we did a flyback DCDC with a planar transformer - it was a 12 layer PCB (polyamide) so it would be possible in this case as well.
@@slawomirjanicki9580 @Marcin Anforowicz Look at JLCPCB, they have 8layers for just $~5. In my experience they are way easier and cheaper to work with in comparison to PCB way. It sounds to good too true until you use them.
How about a thin double sided PCB? You could print 2 slightly different coils to either give you more power or more precision. Come to think about it, how about 4-, 6-, 8-, 10-layer PCBs up to the manufacturing limit? You could simulate weight vs magnetic flux gain. Or, you could just show I have silly ideas 😀
In this video I used a 4-layer PCB as a price compromise. In a future prototype I might add more layers to increase magnetic moment given a fixed resistance. Layers don't impact weight too much since the manufacturer decreases inner substrate thickness to keep total thickness constant.
Yes, very important to make sure the scripts are configured properly for the manufacturing settings. I use a config.ini file for this: github.com/manforowicz/Magnetorquer-Calc
Interesting idea! Though I think micro convection currents and drifting toward the edges would be hard to control for. Also, if its buoyancy isn't evenly distributed, the water's density gradient would apply a torque.
Intuitively I just can't see how this could possibly be comparable to a wire winding. You can just get so much more closely spaced wires. And a square flat winding that's basically the same form factor of a pcb should not be to hard to achieve either.
You're completely right, physical wire can be crammed in more tightly. The PCB approach is lower-cost, more reproducible, more structurally robust, and easier to fit into a tight cubesat. But of course, it has many downsides too.
@@marcinanforowicz8319 Don't you need a special more expensive kind of PCB for space applications though? How much does reproducibility matter here? As long as you can wind a coil that hits your minimum required effic/force output, then you could have software fine tune down to desired output.
I love the cognitive dissonance, "photons have no mass, but have momentum", how can something "not exist = (i.e.= no mass)" but at the same time have an "effect"? Ubiquitous, but anisotropic in the presence of a mass body, the Aether shall set us free. Rotate the Michelson Interferometer vertically to find out.
Answers to common questions:
- This prototype uses a 4-layer PCB from JLCPB.
- We will use an IMU and onboard camera star tracker to determine what control outputs are needed at what times.
- The whole thing will be made open source eventually.
- My scripts are available here: github.com/manforowicz/Magnetorquer-Calc
What program did you use to display all of the graphs from python? Super new to electronics but I’m very curious. Was it just a .py insert? If so what was it? Thank you
It's rare to see an explainer video for in-progress research. Much less one this good! Definitely subscribing...
Yep +1 sub here too
This is awesome...both the ingenuity and the fact that we live in a time where manufacturing something like this for the home hobbyist is not only possible, but relatively cheap and easy. Amazing!
Lol, bro. "relatively" is doing a LOT of heavy lifting for "cheap and easy" 😂
Interesting indeed! I'm doing a video series on diy cubsats and currently exploring the power budget requirements before starting the pcb layouts, the next video currently under research is magnetic torque thrusters and if they worth it for cubesat project. This helps a lot, and I'll certainly link this video in that video as it's a great explanation! 👍
Great project, good luck!
super interesting research! good job! love the non-constant width idea, however i'm not sure whether eliminating the smaller turns was advantageous for the coil. The outer tracks are the longest so they are forming the largest part of the coil's total resistance
Thanks! Your great projects helped inspire this design!
While the outer tracks are linearly longer, the area they encircle is quadratically greater. Outer tracks cover more area per ohm so they're more "efficient".
If I wanted to fill in the center while keeping resistance constant I'd have to widen the traces which would reduce the number of coils packed along the edges.
But I don't mathematically understand why this seemingly arbitrary inner-to-outer-radius ratio maximizes area-sum given a fixed resistance. (At least, according to my Python simulation)
for space applications, I imagine multi-layer PCBs would easily be worth the extra cost.
This is really cool! I hope this technique can be used to make CubeSats more affordable in the future because ACS can be a big economic and space cost!
Keep up the videos, I might've binged many of them the past day or so..
You explain very well and keep it interesting.
Cheers
I find your videos fascinating, even if I can't quite follow it all (a flaw on my part). Your video on the Kelly criterion was broken down so well that I could clearly understand it. I tried learning about it in the past to no avail. If you need any ideas for content, I would love a similar video on gasseon processes.
Really great video hope to see a part 2
Superinteresting, I always wondered, but never cared to research how reaction wheels deal with saturation! Also I rarely post but its great to see awesome "DIY"/scientific content coming from PL - keep it up!
Given how little space you have on the cube sat - I wonder if such magnet could be reused for other things too when its operation is not required - e.g. a really bad antena ;)
Good point. The magnetorquer can be used as a heater if the satellite gets too cold!
Maybe measure the voltage over it to detect if the satellite is tumbling after release. I guess if the satellite is tumbling and you switch on the torquer it might not stop the tumbling as the force is small. But if you know the frequency and phase you can switch on the torquer only at the moment that it almost lines up with earths magnetic field .. and use less power. Because when the angle is 90 degrees the force is zero i guess.. and by turnining it on and off in a smart way you might be able to stop oscillation
Good idea, we might add a voltage sensor to detect the direction of rotation. Though our primary reference will be an onboard camera tracking stars. And yes, we'll need to create an algorithm for a microcontroller to turn them on and off at the right moments.
As the Kepler Telescope reactions wheels failed, it was able to continue science observations by using the unbalanced solar torque (the photons) as a third source of torque and kept pointing with only two RWs.
How strong is the board? could it be used as a structural panel?
I know nothing about satellites or engineering but the video was cool and surprisingly easy to follow :) Good luck with your project
This is interesting. Are you familiar with Malta-based Carl Bugeja doing similar work on PCB electromagnets?
Yes, his work in-part inspired me to do this project.
would it be worth making the same design but on a multilayer pcb? so you can have two coils on one pcb to have double the force in the same space? you could get more than two layers I know but I would assume there should be diminishing returns as the efficiency would probably decrease with all those magnetic fields in the same small space.
exactly what I came here to say. It would seem you could fit many more coils.
Indeed, more layers increase magnetic moment. The prototype in this video actually has 4-layers. Sorry for forgetting to mention that in the video.
@@marcinanforowicz8319 ah fair enough
I have been thinking a lot about this, but never got to chance to try anything. I am happy that someone is working on it. I wonder if making orbital maneuvers using the Earth's magnetic field would be possible. Similar to how the head of a hard disk drive moves sideways, maybe we could use the earth's magnetic field to change the orbit of a CubeSat.
By the way, if you try doing this, I had the idea of using a material with different magnetic permissivity to create a sideways force. If you want to try it: sandwich the coil between two steel plates leaving only one side of the coil open (kinda like putting a blanket halfway). You should cover 85% of the coil with the steel plate and only leave one side of the square open. This should create a sideways force based on my thinking. But I might be wrong. I just wanted to try it a lot, but never got the chance to.
Since Earth is just a magnet, we could only apply a linear force directly toward or away from it. If you hold a magnet over the Earth, it doesn't attract it noticeably more than just gravity. The rotational torque on a compass seems to be much greater. Earth's magnetic field is only really useful for aligning with it, rather than getting pushed/pulled by it. But I may be completely wrong.
@@marcinanforowicz8319 That sounds right .. nasa did do experiments with the space tether from the shuttle. But the hardware failed. They wanted to reel out a really long conductive wire from the shuttle. and if current would run through there might be a force in the direction of movement. Not sure how a current would run through the wire. Problem is if you want to hook the tether up with wires to have current flow .. you end up with a parallel wire to the tether with current in the opposite direction. The magnetic field of that will cancel out the magnetic field of the tether so that doesnt work.
@@apsdev Ha! That's super interesting!
Nice project! You say it took 3 months for your essential chips to arrive, and ascribe this to your institution's purchasing bureaucracy. This is a terrible waste of the time of your talented researchers, but surprisingly common in my experience. Support functions in large organisations often forget that they are the tail, not the dog. Lean engineering eliminates the key causes of waste, including delay/waiting time. Your procurement function needs to get on board with this and keep pace with your research. You need to escalate this issue to your VP/Head of R&D to build accountability. You will probably find they are keen to find concrete examples to push forward change, and that they have strategies to work around until then! Like, buying on personal credit card and then claiming as an expense. Don't give up, demand change and accountability, and fight your corner! You are the dog, not the tail!
Try to sandwich two of these PCB between a thin sheet of ferromagnetic material in order to increase the magnetic field strength produced by the coil
Exactly what I was thinking! Do you have any suggestions on where to buy an affordable soft iron ferromagnetic sheet?
@@marcinanforowicz8319 your best bet is to find something that has it and salvage it. I don't know what products would contain it though
@@friedtomatoes4946 maybe a fridge magnet?
@@marcinanforowicz8319 You can get electrical steel (the kind of silicon iron used in transformers) from McMaster-Carr starting at ~200USD for ~0.65 m². At a guess a little more looking would get you something closer to the size you would use for a lot less.
Interesting proposition.
I wonder which design would be better? Several squared coils stacked on top off each other or merely sandwiching a ferromagnetic material between two parrelel coils? Since once ferromagnetic materials are polarized, their inertial magnetic field would be resistant to the change of magnetic polarity once the squared coils reverse the direction of the current.
did you consider a multi-layer pcb tp effectively double/ quadruple the number of turns?
In this prototype I actually did use a 4-layer PCB to increase the number of coils.
To get a performance measurement it could be cool to mount it on a torque spring in a vacuum and putting some AC on the coil. That way it will start to oscillate, and finding the forces is pretty easy if you know the spring constant and the masses.
Great idea! We could also calculate its moment of inertia (based on materials and dimensions), go to a location with known/undisturbed Earth magnetic field, hang it on a loose string, and see at what rate oscillations die down to deduce air resistance. Then we'd measure the frequency of oscillation, and factor all this in to get torque.
@@marcinanforowicz8319 This isn't really my area of expertise at all, but I just wanted to say that I absolutely adore the enthusiasm and excitement you're approaching the comments section (and this video in general!) with - I don't know quite why but it just makes me really happy.
@@marcinanforowicz8319 You could also apply a large known field with a coil externally, that way you have control about the field, and the increase in magnitude will make it easier to detect movements and forces.
This is glorious!
Nicely done. Just a thought, could you use a magneto style spinning magnet arrangement coupled with a compound coil to create an enhanced large field that was both redirectable and stronger? I'm basically thinking of something like a flux guide but dynamic.
I don't know if I'd be able to pull that off, especially since it needs to survive launch forces. But interesting idea!
Once You enhance the system with an adequately chosen or good magnetometer or more (to achieve near instantaneous magnetic field 3D orientation information ), I would suggest you investigate the domain where You have the unit on the string here on earth (as You have shown) - and nullify the oscillations - critically dampened point (control system terminology) - again for this particular situation. This can develop s/w code of “system oscillation and stability”. Then translate the domain according to altitude planned. LEO ~50nT to 200uT depending on day/night, but of course with variations of angle!
Worst case from a satellite control aspect is at sea level where gravity,air, string are so strong. Check out MEMS magnetic field sensors
Brilliant ! Bravo 👏🏻
There's almost nothing to dampen that oscillation in space and very little in air too. You could try to couple it with an IMU and adjust polarity / voltage to apply breaking and dampen the oscillation. It might actually need a computer vision camera to stabilize as the IMU's compass might be useless near that coil.
Yes, those will be our next steps, integrating with data from an optical star tracker to get info about orientation and velocity.
If they're willing to use reaction wheels, why not gyroscopes on twist axis's? With three gyroscopes with 90 degree opposed twist axis's and it should be possible to accurately allign to anything and it should provide a lot of torque as well. It could even be a modular external deployment and spin in the vacuum with no air resistance on magnetic thrust bearings where the flywheel is a wheighted PCB induction motor rotor with the stators on the outside.
That'd turn the satellite. With magnetic bearings, if you made it like a pancaked squirrel cage, if it was never over-volted, it would never wear out.
not nearly an expert so i might be wrong, but photons do have mass right?
Great explanation 👍 Thank you 👍🙂
top content bro! keep going like that!
Awesome video! I wonder if you added some soft iron and used flux switching to modulate earth's magnetic field instead of generate a field of it's own, whether it would be more efficient - great work either way!
Interesting idea. One of my peers actually suggested a mode of operation which cuts off power and keeps the spirals passively connected in a loop. This would effectively act as a dampener, slowing the satellite's rotation in Earth's field.
@@marcinanforowicz8319 If the frame of your satellite is conductive like aluminium then if the satellite tumbles then the magnetic field of the earth will induce currents which in their turn creates an opposite magnetic field which will dampen the tumbling i guess..
I was thinking about this, and actually changing the orbit of the satellite using the earth's magnetic field. Since the push doesn't need to be huge in the space, but constant.
good video! Let me suggest one new video where you show us how you built the model in Godot step-by-step.
Neat. How do multi-layer boards with similar patterns and powered similarly impact the result? Could you essentially do a 24 layer board bonding different layers to create a lot of specific directions of force? Or help push the desired induction one way or another?
Why are the conventional devices wound around long skinny rods? Does the length give a better lever arm to work from?
Great question. AFAIK the shape of the field you produce doesn't have much impact - intensity is more important. 1) For some configurations, rods fit better spatially into the satellite. 2) Although a narrow rod has a lower area-sum-to-wire-length ratio, this seems to be offset: Physical wires can be crammed tighter together than PCB traces.
How about making a square section tube out of four boards that all connect together (with s something like castellation) and the traces on each of the four boards being straigh lines, open at both ends, to connect to the next board. Essentially building a square section winding on the outside of the cube sat.
Using 4 layer boards, you could use the other layer for the magnetorquer and the other 3 for the sat's electronics payload.
Or a large flex pcb, that wraps around the assembly to form a tube, with soldered edge contact, like an lcd flat flex
That is a very good idea, I'll look into it! It would probably increase area-sum and magnetic moment per weight. On the other hand, having self-contained magnetorquers on separate boards increases redundancy and ease of manufacture.
@@marcinanforowicz8319 yes, it occurred to me when I was thinking about it later yesterday that it does greatly increase the potential failure points.
The other thing that occurred to me it that one could make the boards with the tracks on one layer in one direction and a second layer in a perpendicular direction, and arrange 6 identical board to form a cube with built in 3 axis magnetorquer.
And at the same time, using the pcb as primary structure of the sat.
Damn it! Now, I want to build one myself 🤣
@@randomelectronicsanddispla1765 It can also interfere with instruments on board due to the em no?
@@AlfieMakes any sort of radiating electromagnetic coil will interfere with the rest of the equipment. No mater the shape, size and wether it is a homemade or a bought one.
But so does the wire going from the battery to the electronics, and anything carying a current.
That would be a choice to make depending on the satellite's mission.
Of course, if you are trying to map the earth magnetic fiel in LEO, you wouldn't use any kind of magnetorquer on that mission, or you would have a calibration routine to work around
I wonder if you could just wrap winding wire around the whole cube? If the frame at the corners was made from threaded rod you could use the threads to space the wires to get uniform and consistent geometry.
Are you using any components that have radiation hardened alternatives? I also suggest using a white solder mask as this will increase albedo and lower operating temperatures if exposed to space.
Check Carl Bugeja's channel, he got some coils made of flexible PCBs, they're very thin; maybe there could some gain in weight/magnetic power ratio with something based on his approach?
Wonder if you could add a laser that you could point at something to act as a tractor beam and counteract some of the forces you mentioned?
Amazing!
What do you think about using fractal shape traces?
Thank you for the great video. I have a question. I am trying to generate a magnetic dipole moment of 0.2 with my PCB. How was the total length of the PCB pattern calculated? And what is the current flowing through the pattern? Was the formula M=IA used to calculate it? I would appreciate your help with this
Calculating the optimal design and characteristics is a bit too complicated to explain here, but I might write a paper or something about it in the future. Feel free to use my optimizations scripts as a reference for now: github.com/manforowicz/Magnetorquer-Calc
@@marcinanforowicz8319 Thank you for letting me know. Have a great day!
I wonder if you can stack multiple PCBs as the added complexity of adding another pcb might not be that much compared to the added output force, if they can be added together
These PCB can cost 2 to 5$ US with 30$ shipping and total time of 7days. Why don’t you order it with your own money instead of waiting 3months?!?
The entire satellite is totally lost after uncontrolled spin?
How does adding more coil layers or even pcb stacking affect performance? 🤔
Very interesting.
how hard to heat up? is cooling needed if it's in many layers?
Heat generated is V^2 / R. This prototype is 10 volts and 100 ohms, so it generates about 1 watt of heat which is almost imperceptible. For comparison, a microwave generates around 1000 watts of heat. However, even 1 watt has to be carefully radiated away when in the vacuum of space. Adding layers while keeping electrical properties constant would increase torque while keeping heat constant.
Stupid question but why do you have to have a reaction wheel? Quadcopters keep their rotation to target by adapting the difference in rotor speed. They can rotate as much as they want by changing the difference in 2 rotating motors. So why not just add 1 more motor and get rid of the saturation issue for good?
Once an external torque has been applied, the satellite+wheels system has a net angular momentum. The wheels can't change this since they can't push off of anything external (like air, in quadcopters). They can only transfer some of this angular momentum from the frame to themselves. Similarly, an astronaut spinning in space can't stop without using thrusters or grabbing on to something.
❤ Godot!
why not use a multilayer pcb and continue the spiral in the other layers
Sqaures are good, but what about hexagons?
Considered coppers thermal resistance coefficient? Pretty significant if your operating temp is extreme
Good question. I just assumed room temperature since the satellite will have to be roughly stabilized at that point anyways so that the batteries don't freeze or overheat.
Please share the python script you used to create spiral kicad traces 😁
Of course, here you go: github.com/manforowicz/Magnetorquer-Calc
😍😍😍
Interesting stuff. Subbed.
Definitely subscribed
Hello! Coud you please get a link of this coil winder ?
I used footage of the physical coil winding machine from this video: ruclips.net/video/s6DOWAMhrVA/видео.html
Hey are you Polish? Your name sounds like you're from there
Cool!
It can be done multilayer, right?
Yes, this is a 4-layer PCB. More layers would increase its magnetic moment, but the cost rises dramatically.
@@marcinanforowicz8319 OK, I posted a similar idea without reading the comments, sorry. You are right that these PCBs are expensive. I checked some quotes at PCBWay and a 50mm x 50mm 4 layer 0.4mm PCB with 3/3 mil traces is about $300. If you have never ordered from PCBWay you can get a 2 layer 0.4mm PCB and order them for $5. Then you can stack them with something like thin Mylar spacers. This wouldn't be a production piece, but perhaps good enough for testing.
Well, a 2 layer 0.2 mm PCB can be made at PCBWay for about $100...
For one of our instruments we did a flyback DCDC with a planar transformer - it was a 12 layer PCB (polyamide) so it would be possible in this case as well.
@@slawomirjanicki9580 @Marcin Anforowicz
Look at JLCPCB, they have 8layers for just $~5. In my experience they are way easier and cheaper to work with in comparison to PCB way. It sounds to good too true until you use them.
How about a thin double sided PCB? You could print 2 slightly different coils to either give you more power or more precision. Come to think about it, how about 4-, 6-, 8-, 10-layer PCBs up to the manufacturing limit? You could simulate weight vs magnetic flux gain. Or, you could just show I have silly ideas 😀
I checked PCBWay and they offer a 0.4mm thick 2-layer PCB that is still not a flex PCB
In this video I used a 4-layer PCB as a price compromise. In a future prototype I might add more layers to increase magnetic moment given a fixed resistance. Layers don't impact weight too much since the manufacturer decreases inner substrate thickness to keep total thickness constant.
@@marcinanforowicz8319 I think usually, layer 2 and 3 are thinner than 1 and 4 as well. To keep in mind when calculating trace resistance
Yes, very important to make sure the scripts are configured properly for the manufacturing settings. I use a config.ini file for this: github.com/manforowicz/Magnetorquer-Calc
could you try making it neutrally bouyant and waterproof to test it in a near microgravity environment?
Interesting idea! Though I think micro convection currents and drifting toward the edges would be hard to control for. Also, if its buoyancy isn't evenly distributed, the water's density gradient would apply a torque.
@@marcinanforowicz8319 what about something with a bit more viscosity like mineral oil?
For now we'll focus our efforts on other things related to building a CubeSat, but nice idea!
If they all spin at the same time, it's anti-gravity Propulsion
1000th like 👍
Can you make a 4 or even 6 layer pcb?
Yes, the prototype in this video actually has 4 layers as a compromise between cost and efficiency.
Intuitively I just can't see how this could possibly be comparable to a wire winding. You can just get so much more closely spaced wires. And a square flat winding that's basically the same form factor of a pcb should not be to hard to achieve either.
You're completely right, physical wire can be crammed in more tightly. The PCB approach is lower-cost, more reproducible, more structurally robust, and easier to fit into a tight cubesat. But of course, it has many downsides too.
@@marcinanforowicz8319 Don't you need a special more expensive kind of PCB for space applications though?
How much does reproducibility matter here? As long as you can wind a coil that hits your minimum required effic/force output, then you could have software fine tune down to desired output.
I love the cognitive dissonance, "photons have no mass, but have momentum", how can something "not exist =
(i.e.= no mass)" but at the same time have an "effect"? Ubiquitous, but anisotropic in the presence of a mass body, the Aether shall set us free. Rotate the Michelson Interferometer vertically to find out.
❤🌐😭
😂 i hung a copper coil near my rest area pondering the future and magnetism...such an awesome fabrication ❤ unreal application
Have you considered 3d printing a copper coil media using a sweat tip 😅