FulanoDetail
The previous project log got so full I had to split it up to this one.
I added more information to the bot:
“To determine whether the characteristics of a brushless motor with a specific magnetic flux density are considered good or bad, you can evaluate several factors that influence motor performance. Here's how you can approach this:
Magnetic Flux Density Values: For brushless motors, the magnetic flux density (B) typically ranges from 1 to 1.7 teslas (T) for efficient designs. Your calculation of 1 to 2.5 T for an 80 kW motor at 4000 RPM and 300 Nm of torque indicates that you might be targeting higher performance levels. However, while NdFeB magnets can achieve values above 1.0 T, exceeding the recommended limits can lead to issues such as irreversible demagnetization, especially under high thermal stress or external demagnetizing fields [1][5].
Torque and Current Relationships: The torque produced by the motor is directly related to the magnetic fields in the air gap, which are influenced by the current flowing through the windings. When designing your motor, you should consider the maximum current that your motor can handle without overheating. For instance, typical current densities for brushless motors can be around 16 A/mm² [1]. If your design requires higher current densities, it may affect the thermal management of the motor and ultimately its performance.
Torque Ripple and Performance Consistency: Low torque ripple is desirable in brushless motors. A torque ripple of about 2% is considered very low and indicates a smooth operation of the motor [4]. You should aim for designs that minimize fluctuations in torque to improve overall performance.
Thermal Considerations: The performance of magnets, particularly the residual flux density and coercivity, is temperature-dependent. The temperature coefficients for NdFeB magnets indicate that the residual flux density can decrease by approximately 0.10% per °C [6]. Therefore, it is critical to account for the operating temperature range of your motor to ensure that it can maintain performance without demagnetizing.
Design Innovations: Consider innovative rotor designs that can help minimize leakage flux and maximize torque density. Structures like flux barriers can help concentrate the magnetic flux and reduce unwanted leakage, thus enhancing the overall output power density [5].
The maximum magnetic flux for neodymium magnets, particularly the NdFeB (neodymium-iron-boron) type, is a critical factor in the design and performance of brushless motors. The residual flux density (Br) for these magnets is typically around 1.12 T (teslas) at room temperature (20 °C) [1].
Flux barriers in the stator yoke of an electric machine are typically made from non-magnetic materials or air gaps. The primary purpose of these barriers is to increase the magnetic resistance (reluctance) in specific areas, thereby controlling the magnetic flux path and reducing unwanted harmonics.
Here are some common materials and methods used to create flux barriers:
Air Gaps: Simple air gaps are often used as flux barriers. By cutting specific shapes or slots into the stator yoke, designers can create regions where magnetic flux is impeded or redirected.
Non-Magnetic Materials: Materials such as plastics, ceramics, or composites with low magnetic permeability can be inserted into the stator yoke to serve as flux barriers. These materials do not conduct magnetic flux, effectively creating a high reluctance path.
Non-Magnetic Metals: In some cases, non-magnetic metals like stainless steel or aluminum may be used. These materials are structurally strong but have low magnetic permeability, making them suitable for certain designs.
The choice of material depends on factors such as the specific application, cost, structural requirements, and the desired magnetic properties.”
So, magnetic flux around 1 tesla on each slot, the flux barriers are already part of the design since I plan to make everything out of plastic. Of course, the 1 tesla per slot can change if you use an electromagnet on the rotor…
But I’m feeling like this won’t be that simple to achieve, especially because of the shape of the coil.
Too thin and long or too wide and short and the magnetic flux density will suffer…
But assuming that I make a coil/slot too large, is that really bad? The density will be lower, but that would only be an issue if you were using conventional heavy materials. Here it would only be a matter of size.
“Increasing the size of an existing motor while maintaining its power input would indeed lead to a significant reduction in magnetic flux density. This is because magnetic flux density (B) is defined as the magnetic flux (Φ) per unit area (A). If you increase the area (by enlarging the motor) while keeping the magnetic flux the same (since power input remains unchanged), the magnetic flux density will decrease accordingly. Specifically, doubling the area will halve the flux density, as indicated by the formula B = Φ/A [2].
Now, regarding the implications of this reduction in magnetic flux density, there are a few key points to consider:
Output Power: If the motor's power input remains constant, theoretically, the output power could stay the same. Power is the product of torque and speed (P = T × ω). However, increasing the motor size might affect the torque and speed characteristics due to the change in design and operational parameters. This means that while you may maintain the same input power, the efficiency of converting that power into useful work could change due to alterations in the torque-speed relationship [3].
Lower Losses: The reduction in magnetic flux density could potentially lead to lower losses, especially if the motor operates in a regime where higher flux densities cause increased iron losses (core losses). As mentioned, with a lower flux density, the chances of saturation in the iron core decrease, which could lead to improved efficiency [4][5]. Additionally, a larger motor may allow for better heat dissipation, reducing thermal losses.
Impact of Coil Weight: If the weight of the coils remains constant, but the overall size of the motor is increased, the coil's current density may decrease, leading to lower copper losses. This is beneficial since lower current density can mean less heat generation in the coils, which is advantageous for the motor's longevity and performance [1]. However, it is essential to consider that the overall design must still effectively utilize the available magnetic flux.
Magnetic Path Considerations: Increasing the size of the motor can also affect the magnetic path length and the design of the stator and rotor. A longer magnetic path can lead to increased reluctance, which may require higher magnetizing current to maintain the same level of performance. This could negate some of the benefits gained from reduced losses due to lower flux density”
I can attempt to 3D model it on my computer, but I have a feeling that I would only know it for sure by making a mockup version in real life…
For example, I assumed that the wires were 22mm thick and so on, but I also said on the “Manufacturing” section of this project log (way below) that I would change the production of aluminum wires to be as thin as possible for the sake of spreading them around the slots as efficiently as possible.
I’m not so sure how I could do that in a 3D model.
In a way or another, the best I can do is at least try to do it, right?
I will try first the way I know: making the slot geometry and then slowly making the coils as spreaded as possible.
Boi, I’m taking so long to make these I sure hope I can miniaturize the electric motors to fit inside the 35 joints… I’m still in doubt if it will be worth it since all of them are supposed to be at -70ºC.
I’m really procrastinating tho? I want to do something, something that I enjoy, and I can’t pull myself to do it. I just stare blankly at the computer and the music keeps going on. I heard something about “Executive Dysfunction” or something like that.
Which would explain why this project never goes forward…
Look at these cool 80’s/90’s mech animes:
(not mech, but cool anyway)
Damn, that was really cool.
I wish I wasn’t this stupid so I could finish this project already.
Am I really stupid or this project is just difficult?
Anyway, first 3D model attempt:
From smaller to bigger electric motors:
- 18 slots and 12 poles
- 24 slots and 20 poles
- 30 slots and 26 poles
- 36 slots and 30 poles
- 99 slots and 66 poles
Now modeling the coils on each combination and calculating its weight:
23 wires with 1mm of thickness, allows for 1000 amps in total, but you would need around half of it for the 500 amps 600 volts 300 kilowatts motor.
Just pretend every hexagon is a 1mm thick wire, obviously, it is not a wire that is complying with the shape of the coil, but this is an approximation.
1 - 18 slots 12 poles:
I’m only doing this one because the evolito’s brushless axial flux motor 3d model/image has this number of slots and poles.
I made the 3D model of the coil and it resulted in 30.48743cm³ of volume area with 110mm of height, around 30mm of width and 40mm of length, 24 turns of wire in total.
30487.4058 x 18 = 548773.3044 mm³
Inserting on the density to weight calculator :
1.852 kilograms in total for aluminum wire if you add 1.7x (or 2x for simplicity) more aluminum wire in order to compensate for its lower resistivity at room temperature: 3.1484 kilograms.
6.146 kilograms in total for copper wire.
So it seems like I was actually overthinking the weight issue…?
Now calculating the magnetic flux density, inputting these dimensions, number of turns etc into a magnetic field calculator, it said the magnetic flux density at the air gap between the permanent magnet and coils would be:
57,3356 milliteslas, or 0.0573 teslas.
A common magnetic flux density for brushless motors is around 0.7 teslas, so I don’t know if this would mean the motor is too inefficient or the like.
You still need to take into consideration that ferrite magnets have 0.5 teslas of magnetic field flux, not to mention that you can still use ferrite soft cores for the coils.
So maybe all of this stuff will add up?
In the end, if you need to add 10 times more turns, you would also weigh a little more than 10 times, but it would solve the issue. I think.
If I were to risk, I’d risk it with the 99 slot motor.
2 - 24 slots 20 poles:
For some reason, now the coil has 83mm of height, 30mm of width and 82mm of length. Also 24 turns.
45632 x 24 = 12343582.2384 mm³
Inserting on the volume weight calculator:
2.957 kg for aluminum, double it for compensation at room temperature.
9.813 kg for copper.
Magnetic flux density calculation:
65,2469 millitesla or 0.065 tesla.
3 - 30 slots 26 poles:
Coil has 94.3357 mm of height, 31.4mm of width and 43.4 of length. 24 turns.
26093.9144 x 30 = 782817.432 mm³
2.1136kg for aluminium, double for compensation.
7.014kg for copper.
60,6145 millitesla or 0.065 tesla.
4 - 36 slots 30 poles:
Coil has 264.596mm of height, 23.4mm of width and 9.48mm of length. 24 turns
21921.6888 x 36 = 789180.7968 mm³
2.131 kg for aluminum, double for compensation.
7.071 kg copper.
19,8961 millitesla or 0.019 tesla.
5 - 99 slots 66 poles:
Since I could only fit a single turn until I had the coil to go really tall, its height ended up being 848.3112mm of height, 7.97mm of width and 18.4 length.
22013.0712 x 99 = 2179294.0488 mm³
5.884 kg for aluminum, double for compensation.
19.526 kg for copper.
3,9578 millitesla or 0.003 tesla.
Remembering that these 23 wires with 1mm of thickness allows for 1000 amps, if you cut it by half for 500 amps, the weight is reduced by half.
Also, the amount of tesla increases to 0.009. lol
There is still the 3D model for the ferrite soft core for the coils; according to ChatGPT, the magnetic flux density is so small it is unlikely to go into saturation.
I’m also kinda done with this Project Log in specific, so any further 3D models will be done when I actually start building this thing.
Which… Yeah. From now on, I need to actually start building stuff.
Gearboxes Talk:
I didn’t want to work with gearboxes… But there is a possibility that this may be actually useful for someone that is not interested in the direct drive part, and also because I want to take this subject out of my mind. I won’t stop thinking about it until I write this stuff somewhere.
Yes, I went through all that trouble of making the coilgun actuators, the REB 90 3D model and the countless, countless gearboxes I came up with and all that stuff… BUT, I’m more knowledgeable right now, and I’d like to give a try at finding a good reduction box for a DIY project.
I saw this “explosive” electric actuator for robots, and it is essentially a planetary gearbox with an extra motor outside of it rotating the outer ring of the first planetary gear.
https://ars.els-cdn.com/content/image/1-s2.0-S2095809921005282-gr1.jpg
I think these videos can be easier to understand:
I need to figure out how much each motor changes the output, it would be nice to divide the power input between two equally sized motors tho.
ChatGPT explained that you would need the primary motor and secondary motor to have a 2:1 power ratio, if that is accurate I’m not sure.
This is essentially a continuously variable transmission (CVT) and there are many, many designs out there that wouldn’t need a secondary motor with equal or similar output power as the first one.
While looking for articles on the subject, I finally found the goddang inception-drive, the news articles that talked about it never linked the original article:
If I understood it correctly, the inner gear-belt has a tooth with a certain radius and the outer gear-belt is like a big gear, so when the outer gear-belt contacts the further part of the triangle, the bigger is diameter, and thus, bigger the gear ratio.
No, I definitely didn’t grasp how the hell it works…
https://journals.sagepub.com/doi/full/10.1177/17298806231193891
He uses a conventional v-belt CVT drive as an example, where the output/input pulleys are like a rotating arm, so I just need to imagine it as one in order to understand it. Right?
https://i.imgur.com/SlVzO74.png
As you can see in this crappy drawing I’ve made, you just need one input shaft and one output shaft, while the pulleys can slide around the v-belt. Just like in the conventional CVT v-belt.
Now, why in tarnation do you need four pulleys and all that weird stuff?
I think I finally understood how it works.
If you keep both pulleys fixed, even if they are bigger or smaller than the other, they will rotate in the same direction.
Since only one pulley is fixed and the other is rotating, if you change their sizes, the rotating pulley will rotate in an opposite direction from the fixed pulley.
From what I could understand, it is because if the rotating pulley is at a distance from the fixed pulley, it is like the rotating pulley is attached to a bigger wheel.
So, if you fix one of the three, the bigger pulley, the smaller pulley or the imaginary wheel that is rotating one of the pulleys, then the output of the other two will change.
Actually, I remember seeing something that works kinda like this, it is an orbital drive:
I guess it could work with a variable diameter pulley for an orbital drive CVT.
Like these two.
The simplest continuously CVT that I could find was this one here:
Essentially, the idea is to have two parallel disks with ropes attached, the attachments in both disks can move, essentially changing the size of the “virtual disks”.
If the attachments on the driving disks are closer to the center, while the driven disk has its attachments far away from the center, they are working like if a small gear is driving a bigger gear. Exchanging speed for torque.
If the reverse is done, then it will be like a bigger gear is driving the smaller gear, exchanging torque for speed.
But this thing is limited by the size difference between the two disks…
You could add more disks to increase the difference tho…
I think you would either need springs or linear actuators to actively change the gear ratio of these disks, but if you are thinking of something as insane as a 3 ton lifting actuator for a mech, I don’t think these two options would be very practical.
They aren’t infinitely variable, but they would be easier to 3D model and build.
Just found this one that seems quite simple in principle, but I think it would be kinda hard to fit the clutches into the system.
Bearings, couplings and screws:
You know what? On top of all of that, I will make the bearings electromagnetic and instead of screws/bolts I will have electromagnetic couplings/locks. I was thinking of making the metal core slightly porous so it can be embedded into the plastic structure. This is me on the next day, the electromagnetic locks replacing screws sounds kinda stupid to be honest. Electromagnetic locks are heavy and need iron cores.
Nah, it is stupid, a 14 ton holding electromagnet that I found on Aliexpress has 60cm of diameter and uses 1.2 kilowatts to work continuously. Even if I somehow made it smaller, it would still consume as much and be as heavy. Although I could have sensors and the like to proportionally input energy based on the load it suffers.
By the way, what garbage is pumped into my brain guts, but what made me think that electromagnetic bearings wouldn’t face the same problem of size and power input?
If the entire weight of the mech is applied to a electromagnetic bearing, then the bearing MUST be able to make that much weight to float.
Although I could use electromagnetic bearings and couplings on the connection between the electric motor and the hydraulic pump.
Interesting video, related to the subject.
Electro-hydraulic actuators:
I intended on making this part because I thought the weight of the electric motors would be too much, but in the end it wasn’t that big of a deal.
Between DIY electric motors and DIY hydraulic actuators, I think I prefer electric motors.
Although I talked about how not very useful it would be to make the hydraulic mech, I do want to check one last time this idea. Especially now that I’m aware that multiple independent actuators would be too heavy.
… Now that I look at it, the output values don’t seem that absurd to be honest.
On Project Log 81 I reached the conclusion that even a hydraulic mech with 1.5 or 1.2 times more force for every consecutive actuator would need around 4000 to 5000 liters per minute or 260 horsepower to 350 horsepower.
A brushless motor with 80 kilowatts/106 horsepower weighs 20 kilograms, so around 60 to 80 kilograms for the driving motor.
About the Hydraulic Vane pumps:
I could use a Vane Pump 3D model to be as thin as possible, but one interesting thing I thought of: you could hire laser cutting of metal plates based on how thin these things are.
This website shows some graphs on the efficiency proportional to pressure and flow of vane pumps.
The highest efficiency was 87% at 1800 RPM and 1000 PSI/68 bar according to the final graph, but for the highest efficiency, you would maintain the pressure at 500 PSI/34 bar and around 1000 RPM. … But I don’t know the dimensions of the vane pump in question.
According to this article, the most efficient would be around maximum 20 bar, 1600 RPM and 6 liters per minute of fluid flow. But I don’t know how reliable that is, because the article discusses a unique pump integrated into a bldc motor.
One thing that I absolutely hate about hydraulic pumps is that the components touch each other. Piston pumps (the most efficient), gear pumps, vane pumps, internal gear pumps, all of them have to be in direct contact with the walls, O-rings, seals and bearings.
Everything is beginning to overheat and deteriorate over time.
The only pumps that don't have that problem are centrifugal/impeller pumps, but those are the worst for high pressure applications.
I searched for hours and found nothing about a centrifugal pump with either high pressure nor good efficiency, the same is applied to axial-flow pumps.
I could only think of an archimedes screw pump, but those are used to transport water, not pump hydraulics.
Ok, I think I found a solution: diaphragm pumps.
https://d3i71xaburhd42.cloudfront.net/001140ff5c269ac326f1b9d8afdaa358b020601f/2-Figure2-1.png
The image actually shows a diaphragm actuator, but using it as a pump would work just the same.
It is at least 95% efficient and it won’t care if I use oil or water, which makes things easier.
I could also make the actuators like that.
By the way:
“In a water hydraulic system, to reduce the creation of air bubbles and prevent corrosion, you can consider the following additives and treatments:
Antifoaming Agents (Defoamers)
Purpose: These reduce the formation of air bubbles and foam within the hydraulic fluid.
Examples: Silicone-based antifoams or polyalkylene glycols (PAGs).
Corrosion Inhibitors
Purpose: These protect the metal components of the system from corrosion by forming a protective layer on the metal surfaces.
Examples:
Sodium Nitrite: Commonly used in water systems, it forms a passive oxide layer on metal surfaces.
Sodium Benzoate: Acts as a corrosion inhibitor and is less toxic than some other inhibitors.
Boron-based additives: Effective in preventing corrosion in ferrous metals.
Biocides
Purpose: Prevent the growth of microorganisms that can lead to biofilm formation and subsequent corrosion or blockages.
Examples:
Isothiazolinones: Commonly used in water treatment to control microbial growth.
Glutaraldehyde: An effective biocide for hydraulic systems.
Demineralized or Deionized Water
Purpose: Using purified water helps reduce the presence of dissolved gases and minerals that can contribute to corrosion and bubble formation.
Dissolved Air Removal
Purpose: Remove dissolved air from the water before it enters the hydraulic system.
Methods:
Vacuum Degassing: Exposing the water to a vacuum to release dissolved gases.
Deaeration Tanks: Allowing water to sit in a tank where air can escape before being used in the system.”
It would also be better to make BLDC servo valves instead of solenoid valves, this way the system wouldn’t need to be as heavy (and expensive) as the actual solenoid valves used on hydraulic systems.
https://en.wikipedia.org/wiki/Rotary_valve
This would allow you to simulate spring action on the hydraulic actuators by opening and closing the valves accordingly. You can program that with conventional bldc motors, but this time you would need encoders and other sensors.
Another detail is that I don’t think I will keep up with the idea of using Stewart platforms for the limbs, it would indeed be stronger and studier, but it would also require even more fluid flow, even more actuators and even more weight.
By the way, I was wondering if it would be a good idea to have the hydraulic actuators be hydraulic motors.
More specifically, radial piston hydraulic motors.
They seem to be very torque dense (and ChatGPT agrees), although I wouldn’t make it a fully rotating motor precisely because of the way the fluid is supplied. It is normally fed through the axis of rotation, which adds friction losses.
Although I fail to see that much difference between a radial piston engine/pump.
Why does the radial piston hydraulic motor have moving pistons pointing inwards and pushing a wobbly surface instead of having fixed pistons pushing a wobbly rotor?
Guess that answers the question: the rotor would be too small.
Still, I think I could keep the inner part where the pistons are fixed and let the wobbly surface move.
Looking at gifs and videos on how it works, I have no idea if I will be able to pull it off.
I don’t know how I would time correctly which piston would receive fluid.
I could only think of having various sets of pistons in different positions compared to the curves of the walls and simply activate each set/disk when it is at the best position for maximum torque.
Almost like a 3 phase BLDC motor.
Not like what I said, but pretty close.
… But that would require a LOT of fluid flow to compensate, no? I would need to make a valve system into the motor so that the exhaust of the previous “ring phase” helps power up the next “ring phase”. A regenerative system, like a turbocompressor of some-sort.
That could also be applied to the electric system of a motor, no?
I was just not paying attention to the details, I rewatched the first video I linked here about radial piston motors and essentially, the ports for each individual piston are holes on the rotor structure, and the feeding and exhaust of the working fluid are fixed holes too instead of a sliding opening like in the axial piston pumps/motors.
You need to get the position of each port very carefully aligned with the wobbly surface so every time the pistons expand or compress, the whole thing will rotate.
Some radial piston hydraulic motors work more like vane pumps, where you can change the distance between the walls and the rotor by moving a disk. Just like swash disk axial pistons.
This is about a radial piston pump, but it can be used as a motor anyway.
If I’m going to make linear actuators/pump pistons with membranes for as low friction as possible, then why not use that idea of flexible hydraulic cylinder artificial muscles that I suggested?
Although not as incredible as the axial piston hydraulic motor, these would be easier to manufacture.
Anyway, I need to figure out the linear speed of the pump, the cylinder diameter, pressure, fluid flow speed and thus, port size, cavitation production etc.
Should I change the actuator?
Pneumatic/hydraulic Artificial muscles:
Out of curiosity, I don’t remember posting these two artificial muscles, but these two redesigns of a McKibben artificial muscle achieved 45% and 65% of efficiency while working with Pneumatics.
https://www.mdpi.com/2076-0825/10/5/100
https://www.science.org/doi/10.1126/sciadv.adi7133
I think that is relevant because it could be a good option for the hydrolysis driven pneumatic system or a hydraulic system altogether.
Of course, I don’t see how I could mass-produce those (at the moment), but boi.
OH BOI, all this electric motor hydraulic motor talk is driving me CRAZY.
I’m willing to bet that if I simply stuck with the simple approach, with the pros and cons, I would've finished this project already.
Homemade Shape Memory Alloy:
By the way, I also saw this new video from NileRed where he shows an arc-melting furnace that is meant for melting highly reactive metals. It is simply a chamber with inert gas (like argon) with a conventional arc-tig setup where the arc is formed through the target material, instantly melting it. He uses it to make metallic glass, actually.
The interesting part, however, is that you could literally make Nitinol wire in bulk with this thing by constantly injecting material and melting it through a tube.
Even though Nitinol is not the most efficient, it is certainly the simpler and highest performance one. I would only need a dozen kilograms for an entire mech instead of hundreds of kilograms and complex systems without the risks of electrocution, frostbite and instant combustion.
Yes, both nickel and titanium are expensive as hell, the other shape memory alloys that use cheaper materials are less efficient and less energetically dense than Nitinol. But that could be so much easier to handle...
These actuators are less than 10% efficient, so the weight they are saving on themselves is compensated by the extra weight they will need to actuate, defeating the point all together.
I try to check for shape memory alloys with higher efficiency, but they never outright measure the actual efficiency. They only say that the efficiency is 10 to 15% and never try to check the energy input, contraction speed, distance, force and the time all of those things took in place.
So, if you input X amount of joules of heat for 1 second and then the whole thing takes 10 seconds (or even a minute) to fully contract with a given force, speed that in total would be 60% of the initial energy, then is it really 10 to 15% efficient?
This article claims “energy savings of up to 80% (80% of 90% is 72% +10% to 15% = 82% to 87% efficiency):
Nickel/Titanium free SMAs:
https://www.intechopen.com/chapters/67075
https://link.springer.com/article/10.1557/s43578-021-00444-7
https://www.sciencedirect.com/science/article/abs/pii/S0141029618305030
https://www.mdpi.com/1996-1944/15/22/8089
I don’t know if this one is a rare-metal free shape memory alloy, but it showed a method to make it actuate 35 per second (35 hz): https://www.nature.com/articles/srep21118
(maybe I just missed it, but they didn’t write which alloy was used)
Also, a detail about the SMA’s: you need to allow them to fully return to their original state (or “contract”, if you will) or else they will degrade rapidly.
I did think on making the wires as coiled as possible, so even though they try to contract, they would never fully restore their initial shape, making absurdly strong contractions. But that extra tension would also increase the degradation of the SMA through fatigue and deformation.
I also thought of braiding and/or twisting the SMA wires in order to increase the force output and resistance, just like real ropes. And just like real ropes, if you always use these actuators with the lowest force possible with a safety factor, you can extend their life cycle. But I’m not very sure how much that would extend its lifetime.
I would copy-paste a paragraph from ChatGPT explaining all of that, but it would be too long. Just know that for now, the average life-cycle of SMA’s are around:
“The values of the displacement drift generated after 1, 1000, 2000, 3000, 4000, and 5000 such cycles as measured using a Talbot interferometer are 0, 0.875, 1.275, 1.459, 1.720, and 1.859 mm, respectively.” Source: https://opg.optica.org/ao/abstract.cfm?uri=ao-57-20-5779
So I’d assume that you would lose 4mm of contraction every 10,000 cycles. 4cm after 100,000 cycles and 40cm after 1 million.
In this other article it is said that the faster the cooling of the SMA’s, the better. Longer cooling cycles increase fatigue. Although they are talking about nickel based SMA’s, maybe that would be true to other shape memory alloy types.
Why does every freaking option sounds good and viable?
Everything looks so simple yet so complex, so practical yet so impractical…
Mandatory Armored Core reference:
This project log got so full I had to split it up to the next project log. (holy crap, I'm sorry)
Discussions
Become a Hackaday.io Member
Create an account to leave a comment. Already have an account? Log In.