Design log

I’ve seen a lot of build logs here but I’m not there yet with my project, so I figured that I’d make a design log

I was impressed by @Halloween2go ’s build and documentation and was initially considering something largely inspired by his but the more I thought about it the less appropriate it seemed for me. For one, he seems to have a lot more space than I do so ideally my sim would be shorter. So back to the drawing board.

I looked at SimCalc. It seemed like a good place to start daydreaming about geometries but after it raised some questions I ended up writing my own similar tool, https://www.xsimulator.net/community/threads/motionvisualizer-software.16079/ I checked it compared to results I get from modeling various geometries in cad software. I adjusted the cad geometry by little bits to find the numerical derivative of the pitch and roll with respect to motor movement. These numerical derivatives equal the gear ratio between the motor and the rocker (sorry if there’s already a different term for the top portion of a sim that actually does the pitching and rolling) so, if for some set of geometry, moving both the ctcs by 1.25 degrees in the same direction changes the pitch by 0.05 degrees then:

pitch torque = 2 * motor torque * 1.25 degrees / 0.05 degrees (since there are 2 motors)

and

pitch speed = motor speed * 0.05 degrees * 1.25 degrees

and if moving the ctcs in opposite directions 1.25 degrees changes the roll by 0.1 degrees then:

roll torque = 2 * motor torque * 1.25 degrees / 0.1 degrees

and

roll speed = motor speed * 0.1 degrees * 1.25 degrees

I also did similar calculations for linear actuators but then I adjusted the actuator length by a percent of the pitch and said that the motor rotated by that same percent of 360 degrees.

From there, I still didn’t know what my target values should be so I plugged in my best estimates for Halloween2go’s. I believe he uses pgsaw motor so I entered their torque and RPM values. Around then I was also daydreaming about a diy linear actuator using a motor that’s about the same power so if I got the motion ratios right then I’d get comparable performance, but I couldn’t get anything close. That led me to notice that using the rated speed and torque for the pgsaw motor you get about double the rated power so I can only compare to designs that use pgsaw motors if I also use one (I’m not honestly very familiar with motor rating standards but some others that I looked at have rated speeds and torques that yield their rated powers)

So, now it’s time to come up with a less rule-of-thumb based approach for estimating how much torque and angular speed I want. And I know I’m making this unnecessarily complicated, but I enjoy these sorts of low-level deep dives

Yeah, It does take up some room for sure.. But I'm having a great time for the last year on this rig! DMAX

After thinking a bit about how much pitch and roll angular acceleration to target, it crossed my mind to have my head's linear acceleration be equal to what it'd be in a fun but not particularly high performance car. After looking around a bit 0.8g seemed reasonable for an initial hand wavy target, but then while trying to guestimate what the angular speed should be, it crossed my mind that if I have the max acceleration and the max distance it can go, then I can calculate the max achievable speed.

torque = mass moment of inertia * alpha

theta = 1/2 * alpha * time ^ 2

omega = alpha * time

so:

omega = (torque / mass moment of inertia) * (2 * theta * mass moment of inertia / torque) ^ 0.5

where:

torque: pitch or roll torque

theta: 2 * pitch or roll (assuming they're measured from neutral to limit)

omega: angular speed, also pitch or roll max speed given max motor speed and effective gear ratio

alpha: angular acceleration

Now, my new plan's to design geometry that maximizes acceleration while still allowing the max speed, as calculated above, to be lower than the max motor speed given their max RPM. This will make it run out of torque before it runs out of speed, but only by a little. I figure that like with DD wheels, even if I don't "need" all the torque, it'll allow me to have more subtlety than if I'm demanding more from the motors. On the other hand, having extra speed seems pointless other than to account for errors/uncertainties in the design. This is why I'm expecting to compare to the full range of motion, that seems unnecessary (since slowing down takes time/distance too, not just speeding up/coasting) but it'll build some room for error into the calculations.

This still all depends on how much pitch and roll I have and I don't have a good way to to estimate what would be good so my plan's to look around here as well as at mid-range commercial motion sims.

All this is dependent on knowing the mass moment of inertias in the pitch and roll directions so I made this.

Clearly it's missing a lot of parts but the person's right about my size and I set all the mass properties so I could get initial estimated inertias. I'm encouraged that it shows the center of mass (the pink coordinate thing) above about where many pivots seem to be mounted.

I'd been thinking for a while that I'd want symmetric torques between pitch and roll but my model's showing noticeably more pitch inertia (116881 pounds * square inches) than roll inertia (85852 pounds * square inches) which makes sense, so now I'm thinking that I'll shoot for asymmetric torque and ranges of motion so that I can have symmetric accelerations.

I updated my previously mentioned software so now when I enter the pitch and roll inertias it'll do all the above calculations for both pitch and roll as well as some other things.

While there's going to have to be a lot of "sanity checking" as things come together, I'm starting to feel like I have some direction!

Sounds interesting. You may be on to something.. I'm just a little different and just GO FOR IT Best of luck building the rig my friend.. (personally, I thought about it for about a month.. then got a motor working, then just made it..) the rest is history.. It's just a blast.. all the groundwork has already been done my friend. I'm an engineer so maybe that's why I built it the way I did. You probably are too (but I try to not overthink the physics when, in the end - you only have so many years to live ) My advice - Build it - change it if you need to- have fun in the process.. DMAX

Good to know about seat height vs sensation of motion. I've been wondering in general about cockpit geometry (which was another reason for modeling myself) and I'll definitely keep that in mind.

I'm curious to see how much torque and speed my approach will yield relative to some other existing designs which are clearly solid designs.

Back in my engineering days, I always thought that figuring out how to solve the problems that I was tasked with was the most interesting part of the job. For me, getting a welder, a grinder, and sticking things together until it works the way I want based on pictures would be taking some of the fun out of it I'm presumably going to spend most of the next 6 months hiking thousands of miles and interacting with thousands of tourists; to me, right now, having some quiet, physically relaxing computer work to look forward to sounds great! And as a side bonus, if the design's largely done before the fab begins, the time that the garage will be a disaster will be shortened, which would be better for my marriage

Things got busy but I made some time!

Initially I was considering DIY linear actuators but quickly realized that getting the gearing right was going to be more complicated than it'd be worth, so I decided that gear motors were better for me. I then played around with some layouts using them but realized that my initial design philosophy was giving less movement than I wanted so I did a bunch of reading here about degrees of movement, cast aside my previous approach, aimed for about +-8 degrees of pitch and +-10 degrees of roll, and figured I'd worry about torque later.

Those larger angles will require torquier motors than I was initially expecting to need so I'm expecting to get some 350W motors from aliexpress. With shipping, they're about the cost of pgsaw motors. They do operate at 24V which'll drive up the power supply cost, but at least according to the spec sheet for the main chip on IBT-2s, they should be able to handle the voltage and current. I expect the IBT-2s to be flakey, but I get the impression that they either work as expected or are DOA so I'm assuming that troubleshooting will be a moderate but not major chore.

It clearly still needs some work and polishing, but I'm not seeing any major shortcomings or open potentially show-stopping questions.

And as a nice validation, to several decimal places, the program I wrote matches the CAD model results as I adjust the CTC angles!