I'm using 12v/960rpm motors, so ChatGPT told me that it's too fast for a UGV also it has a lower torque with 100ma.
So ChatGPT suggested me using pulleys to lower speed and increase torque.
The problem for me now is how to tune the belt if I would follow ChatGPT's pullets setup to have the driver pulley at the center and it drives front/rear pulleys.
For your project I would look at one of the 12V 37mm metal gearmotors. They are readily available in the 330 to 360 RPM range.
Before selecting the motor we really need to know the approximate weight of the robot and the diameter of the drive wheels. Those two numbers will give us a much better idea of how much torque is required.
Also remember to size the motor driver for the motor stall current, not just its normal running current.
You can always adjust the component positions so you do not need a cardan.
Read about how traction differentials work... and stop using AI slop.
I'm not surprised you want to build one.
The cost of the identical device to the one in your first post is from just over $6.5k up to $14k and more.
That implies quite a lot of engineering/hardware. Why does it need to be identical?
But the problem is the motor shafts are 6mm, and ChatGPT suggested going 8mm for the rest of the mechanical parts, and the way to do that is with cardan that is 6mm to 8mm.
ChatGPT suggested going 8mm because it's more common and more powerful to handle the tough terrains or experimenting the UGV outdoors.
Not identical, I just want to follow this route of robots.
Even the next project I want to do after this one is a custom quad using STM32 balckpill board I found on Github:
https://github.com/rizacelik/STM32F411CEU6_INAV_Firmware
I love these kinds of projects, because it get me into deeper levels of understanding instead of purchasing a pre-made one. That's not the goal, I like to solder and connect the parts myself and do the testing and the rest of the project management.
Yep, first I went to AI, then came here. I guess the next station is doing more research and studying.
So what does "I want to build the exact robot like this one" mean?
Just leave the "exact" bit out.
If I'm not mistaken, your image is taken from Dr Robot Jaguar, an expensive piece of kit.
There are many robot chassis like the Jaguar available at a fraction of the price, some cheap enough just to try out.
The tricky part is going to be engineering it all, getting everything lined up.
Don't follow the AI stuff, you can see what a hash it can make.
You can do quite a lot using aluminium extrusions. Ball bearings can make it all run smoothly.
If you haven't already said, what do you want to do with it?
I'd quite like to have something along those lines to help with tasks around the house, moving storage units for example.
Good luck with it.
I mean this type of UGVs, not for racing or off-road climbing cars. The ones used for surveillance and exploration.
Yeah, I copied it from their website. It's similar to what I want to do in case of size and shape.
I know that, I have contacted a chinese shope that design big drones with wings and they told me they want $3000, so I have an idea of what these kits could cost. Even this drone considered a budget one for a working drone with software and hardware support. But I prefer to do these projects myself, it's a hobby for me.
That's the absolutely the most important part for me right now. But I'm not in a huge rush. It's not a graduation project or homework. It's a side work beside my main job.
Yeah, but still I learned a lot of things from Gemini and ChatGPT. Even Gemini is the reason my robot arm is working in coordination and joint modes, I'm sure without Gemini I couldn't have figured out the kinematics functions.
And it's also a lesson for me to know the limits of these AIs in case of providing very accurate schematics and diagrams. Even I enrolled in ChatGPT pro subscribtion, it still make errors. So here I leanred that I can't fully rely on the AI. But it still it can help me with theoritical infomation.
That's what I want to do, I found a laser shop in the city I live in. I'm planning to visit them but after I buy all the parts and do some measurements in home.
Experimenting, I learn a lot doing projects. Even for the quadcopters, not planning to buy a pre-made one, no joy in that for me.
I'm not really seeing a lot of use for AI in this.
You already have the motor, can you get the appropriate sized pulley to mount directly to the motor shaft, or do you need to change shaft diameter from 6mm to 8mm? Will the motor bearings be sufficient to withstand the side load of the drive belt, or is an externally supported shaft needed? Does the motor size require it to be mounted in a certain orientation - either front-to-back or side-to-side - or will either orientation work? Will you be able to adjust belt tension by moving the motor mount, or would a movable or spring-loaded idler pulley be needed?
There are solutions for this other than a Cardan coupling (AKA "universal coupling"):
https://www.ebay.com/sch/i.html?_nkw=6mm+8mm+shaft+coupling
I'm not sure what led you to believe a Cardan joint was the only option. I recommend instructing the AI to provide information in text form when possible. It seems like the use of diagrams is the cause of the hallucinatory garbage information you are receiving from it. Diagrams should only be requested when absolutely required.
As others have said, it seems you may be relying on AI a little too heavily. AI is a useful tool, but so is a good old Google search (ignoring the AI summary and going straight to the results of trusted human-created content).
Learn to make stuff yourself! The difference in shaft diameter just needs an adapter NOT a cardan/universal joint!
Is the chatGPT that you used to produce this fascinating mechanical arrangement the same, no-cost sort of AI that basically everyone trusts for advice about their innermost fears, love lives, job hassles, medical and psychological problems?
If not, does the quality of advice scale with the monthly subscription fee?
It's too late, I already bought 12V/960rpm. And that's why I want to be so sure about the rest of the parts, especially the mechanical ones because any issues with the tuning and dimensions means I have to buy more parts.
Even at this point, I can buy new motors like the ones you suggested, but I thought why not try with these current motors and speed reduction mechanism using 1:3 pulleys set and test if this one would provide enough torque to move the robot around, not to carry to transport stuff. Just cruising and exploration and I think I should work, because the motors I have itself are geared so they should provide reasonable torque.
Size: My planning is a robot L25cm x W25cm x H10cm
I aksed ChatGPT to give approximate dimensions:
Weight: Well ! this part is what changed the situation, and thank you for your question because I was confident that the motors will drive the whole robot weight.
I did some calculation with ChatGPT and it turned out the motors are very weak ! I had a little doubt that they have really weak torque, but I thought with the pulleys, the problem will be solved !
Here's ChatGPT answers, after I provided the motors the specs picture:
This specification changes the answer quite a lot. Your motors are much weaker in torque than I had assumed.
That means the motors would probably make the robot move on a smooth flat floor, but I would not use them for the terrain robot you are designing.
---------------------------------------------------------------------------------
That is a completely different drivetrain compared with your current 0.43 kg·cm per side.
So I would keep the overall mechanical concept we've developed — motor → Cardan → 16T → belt → two 48T pulleys — but I would seriously consider changing the motors before purchasing all the drivetrain components.
Your existing 960 rpm / 0.17 kg·cm rated motors are better suited to a light, fast robot than a ~4 kg terrain skid-steer vehicle.
As you recommended earlier, I think I better buy more powerful motors and start again.
Yes, I'm using this ESC:
According to Gemini:
Yes, this ESC will easily run two of these motors.
Motor Stall Current: 1.2 A at 12 V per motor (from image 2).
ESC Capacity: 5 A continuous per channel (5Ax2 dual-channel ESC).
Safety Factor: Even if both motors fully stall simultaneously, the combined draw per channel is only 1.2 A—well under the ESC's 5 A limit (less than 25% of its rated capacity).
Did you tell AI that you use 6S2P battery?
I just realized that wolfrose is an AI training exercise and we are supplying the training!
How did you realize it?
Yes
