It would be nice to have some feedback from each motor. I don't think the camera will be able to see the wheels. I have two current sensors but they're for batteries. For now I want to see some sand action.
The RC controller outputs a value from -512 to 512, with 0 in the middle. Same for both throttle and steer joysticks. That has to be scaled to 0-100 range, then depending on the sign toggle some pins, then remove the sign, and either add or subtract it to the left or right wheels and sent as a PWM value. It was a lot more complicated than I imagined lol. You can't send the PWM object a negative number, and you have to set enable and direction pins in the process. It works great, except the same motor that was giving me an issue last night. Very strange. I'll work on it some more later today. Right now I can go forward and backwards and turn left, but not right... The PST debug info is great I can test it all without having to spin motors.
Minecraft! It has been forever. I want to do that serial port output to control some LEDs when you flip a switch in-game.
I managed to get the RC 4 PWM motor skid steering code working. I took the rover off the stands, and for the first time the rover has too much traction. You could also say it doesn't have enough power. There is no way it can turn in place with the rubber bike tires on. I figured that would be the case, and moved onto larger turn radius.
I wrote another program that wouldn't allow the opposite motors to change direction. To turn left it would max out the right side motors RPM and advance the left at low speed. There's a small arc, it's not going to be enough. The wheelbase and power available aren't made for it.
At the same time, the thing is a tank. It will run over all kinds of stuff. The lackluster low end torque is workable now. It can move very slowly, which is what I want. AWD with front wheel steering is going to be awesome, and worth the effort. I hope to put the pillow blocks on a metal plate that can pivot. My actuator is only 2" (after I re-modify it). It has 200lbs of force, so it doesn't have to be too far from the pivot point. It would bring the turning radius close to where it was. It's crude, but easy. If that doesn't work out, it's goodbye PW chassis.
You only have a couple of choices, either pivot the whole front axle, or dump the solid front axle, and pivot each wheel separately. Lot of work on both, but less weight with the latter option. Skid steering would be less desirable for maneuverability. So I hope you can figure out how to turn them wheels separately, or together on a turntable. More money more metal and more work, you'll get it, It's so close.
BTW: Like what you have done with the electronics, can't wait to see it in action.
Thanks Mike. Here's a shot of the 2nd motor controller. With heat sinks, they can push a combined 120 A.
I will pivot the axle. I have to. It's going to suck putting it back in the garage. The living room is much better with a rover in it. Right now it's my foot rest. 70 Ah of battery, no need to turn it off. I do pull the motor controller fuses, I can't leave my RC transmitter on for days. I don't need a Ford F-150 shaped hole in the wall
The 4 wheels, gearboxes, motors and two axles could be used in a lot of ways. It's odd these parts aren't more popular for robots. For $60 I have 4 wheels, hubs, axles, gearboxes and motors. That seems worthwhile regardless of the chassis.
Worth noting, there used to be toy RC cars which had 2 motor differential steering on a free-pivoting front axle. Speeding up/slowing down the motors would move the axle. A smart guy like Jon could make a 4WD chassis with a pivoting center (with angle-measuring encoder) so changing motor speeds would articulate the chassis for improved steering. I'm not so articulate myself, so sketch attached.
Here is something achievable with some welding on your axle:
IMHO: I think your traction will suffer if you pivot the axle in the center.
I've been thinking about finding someone with a good welder to do some work. The issue becomes not wanting to leave it in the desert. We need to think of it like hitchbot. Crappy, with a limited degree of functionality.
Worth noting, there used to be toy RC cars which had 2 motor differential steering on a free-pivoting front axle. Speeding up/slowing down the motors would move the axle. A smart guy like Jon could make a 4WD chassis with a pivoting center (with angle-measuring encoder) so changing motor speeds would articulate the chassis for improved steering. I'm not so articulate myself, so sketch attached.
Thanks. I would have never thought of that. Mike you should try that with your crawler. It seems like it would work well and be a fun programming exercise. Once I have my pivot installed I will try it. I hope to get the steering "working" Saturday.
Progress! It was fun smashing those pillow blocks off, putting them in the dumpster, and never having to see them again. I should have just measured them and returned them.
I devised a way to extend and increase the diameter of the front axle. I had to cut a little to get the motor to fit. Its works nicely, but looks horrible.
Pic #1 - Original steering, spacer and piece of the front axle.
Pic #2 - Axle "extender" installed.
Pic #3 - New front end, put together wrong twice.
Pic #4 - This is how far the wheel sticks out. Not gonna happen, it looks bad and I don't think it would fit in my car.
So, I have to move everything inward. That will give me something to think about. Maybe I'll use some wood.
Today went well. To make up for the width of the motors I moved everything inward. I used a 2x4 as a spacer, thick metal bars and a bunch of nuts and bolts to get the steering working with AWD. It's very sturdy except for where I mounted the steering actuator to the rover. I also need more flexible wire for the front motors.
The only way to stop the gearbox from moving around is to attach it to the steering knuckle arm (white nylon). I opened the gearbox and put two bolts through the back into the knuckle. All of this feels a lot stronger than the original front end, but some of the attachment points are weak and not finished.
Once the wiring is done, and the bolts going into the rover body are reinforced, test drive!
1st pic - Total fail. I put the actuator right in the way of the motor.
2nd pic - I thought it looked cool with the open gearbox.
3rd pic - AWD steering finished, for now
The actuator is installed and out of the way now. I must have not taken a final pic.
I tell you what, you should get an award for reworking the original parts. I see that grey plastic cross member, it looks like you reinforced it every which way but loose.
When you think the dimensions are the way you want, make a second version completely out of metal. Use some bronze or nylon bushings, and you have a steering unit you can bolt under that body, or to a complete chassis made out of metal.
Go for the gusto, but don't leave it in the desert alone.
Thanks! Yeah the original arm was used because I couldn't make the connector in the green circle, attached. There will be more angle aluminum across the open part on the top, and some bolts to hold it to the opposite side. I'm using scrap so there's a lot of holes and lines that have nothing to do with anything. I will paint it if everything holds up.
Here's a video of the steering in action. For some reason the front motors won't turn. I can't wait to find out why, some other day
I did some troubleshooting, the front motors wouldn't turn because there was some grease and dirt knocked loose in the gearboxes. I guess the best option would be to clean these and repack them. I can wipe them out with paper towel. I'm not sure what grease to use. It appears that even a small trace of lithium grease on the gears adds too much friction.
They have an excellent seal. These are the original motors with a few years on them, and very little dirt inside.
I think I'll go with heavy grease where the gears mate with the gearbox, there are no bearings in there. Keep the rest dry, adding anything to the teeth adds a lot of friction. The teeth of the gears do not mesh tightly, so the excess grease doesn't get pushed out.
After cleaning out the gearbox, drying of the teeth, they spin friction free. Not overunity, around 98-99% efficient Is that enough?
Nope, motors still wouldn't turn. A week ago, I was AWD testing skid steering. It was a tank. Now it's dead. What happened?
The only difference is the front wheels turn on another axis. Same motors. It didn't make sense until this morning when I came to the realization that it can't stay in the house much longer. Not because it takes up too much space, which is does, more that it was born for the desert.
Grease in the teeth was an issue. My connection to the batteries turned out to be the biggest issue. I was using some cheap 12" jumpers. I swapped out the (light duty) clip leads and bolted the power wires to the battery terminals, and everything worked. The clips are much easier to disconnect than the bolts. Pulling the main fuse is easy, and that is what I have been doing. I don't know why I started using the clip leads. I thought they would get hot or smell funny if there was an issue. I did think about it when I was doing it. Now I can laugh about it.
We're closer than ever to the 2nd beach test. I have high hopes. It's back to tank status!
Lol, I don't know how many times I have been led down the garden path by clip leads! Way too many to count. I wished I had half of the hours wasted back.
Jim
You too? I threw them out, I can't believe how tiny the conductor was compared to the outside diameter of the silicon insulator. Very deceptive.
Now that the ball is rolling again, steering feedback will be next. I REALLY like these flex sensors. I've never owned one before. It seems a much easier choice than a potentiometer. I'll have to find a good mounting spot where nothing can interfere with it. For now I hot glued it so I can play around with code.
I've wasted several minutes of my life with alligator clip leads which didn't conduct at all. Some crimping failure at the clip. That's a pretty serious fail for such a simple device.
I've commented previously that clip leads are surprisingly expensive for what they are. We can put a man on the moon, but we can't make cheap, reliable clip leads!
Everything is connected and seems to be working. I have to add some more supports before the next sand test. I drove it around the yard and it did well. It has a tighter turning radius than before. I had to cut the fenders a little to get clearance.
Pic 1 - Larger rivets to hold the tread on, I ran out, I'll add some more before sand. They worked fine in the back yard on grass and dirt.
Pic 2 - I used zipties to do strain relief on the front motors. That seems to be working well.
Comments
The RC controller outputs a value from -512 to 512, with 0 in the middle. Same for both throttle and steer joysticks. That has to be scaled to 0-100 range, then depending on the sign toggle some pins, then remove the sign, and either add or subtract it to the left or right wheels and sent as a PWM value. It was a lot more complicated than I imagined lol. You can't send the PWM object a negative number, and you have to set enable and direction pins in the process. It works great, except the same motor that was giving me an issue last night. Very strange. I'll work on it some more later today. Right now I can go forward and backwards and turn left, but not right... The PST debug info is great I can test it all without having to spin motors.
Or you could bag the motors and all the hardware entirely and just create a UGV Simulator in Minecraft...
I managed to get the RC 4 PWM motor skid steering code working. I took the rover off the stands, and for the first time the rover has too much traction. You could also say it doesn't have enough power. There is no way it can turn in place with the rubber bike tires on. I figured that would be the case, and moved onto larger turn radius.
I wrote another program that wouldn't allow the opposite motors to change direction. To turn left it would max out the right side motors RPM and advance the left at low speed. There's a small arc, it's not going to be enough. The wheelbase and power available aren't made for it.
At the same time, the thing is a tank. It will run over all kinds of stuff. The lackluster low end torque is workable now. It can move very slowly, which is what I want. AWD with front wheel steering is going to be awesome, and worth the effort. I hope to put the pillow blocks on a metal plate that can pivot. My actuator is only 2" (after I re-modify it). It has 200lbs of force, so it doesn't have to be too far from the pivot point. It would bring the turning radius close to where it was. It's crude, but easy. If that doesn't work out, it's goodbye PW chassis.
Any thoughts on adding steering to this thing?
You only have a couple of choices, either pivot the whole front axle, or dump the solid front axle, and pivot each wheel separately. Lot of work on both, but less weight with the latter option. Skid steering would be less desirable for maneuverability. So I hope you can figure out how to turn them wheels separately, or together on a turntable. More money more metal and more work, you'll get it, It's so close.
BTW: Like what you have done with the electronics, can't wait to see it in action.
I will pivot the axle. I have to. It's going to suck putting it back in the garage. The living room is much better with a rover in it. Right now it's my foot rest. 70 Ah of battery, no need to turn it off. I do pull the motor controller fuses, I can't leave my RC transmitter on for days. I don't need a Ford F-150 shaped hole in the wall
The 4 wheels, gearboxes, motors and two axles could be used in a lot of ways. It's odd these parts aren't more popular for robots. For $60 I have 4 wheels, hubs, axles, gearboxes and motors. That seems worthwhile regardless of the chassis.
All that matters is what you do with it.
Mike
Put a sign on it warning of radiation exposure, and a little box with fins and wires coming out of it.
Who's going to argue, and they may just back off.
IMHO: I think your traction will suffer if you pivot the axle in the center.
Yes, mount some motors and ship it over
I've been thinking about finding someone with a good welder to do some work. The issue becomes not wanting to leave it in the desert. We need to think of it like hitchbot. Crappy, with a limited degree of functionality.
Thanks. I would have never thought of that. Mike you should try that with your crawler. It seems like it would work well and be a fun programming exercise. Once I have my pivot installed I will try it. I hope to get the steering "working" Saturday.
Do you folks in CA have to have a license for everything?
You would be surprised at what you can do with one of these:
https://m.harborfreight.com/90-amp-flux-welder-61849.html
A license to weld?
Probably, nothing surprises me anymore.
I devised a way to extend and increase the diameter of the front axle. I had to cut a little to get the motor to fit. Its works nicely, but looks horrible.
Pic #1 - Original steering, spacer and piece of the front axle.
Pic #2 - Axle "extender" installed.
Pic #3 - New front end, put together wrong twice.
Pic #4 - This is how far the wheel sticks out. Not gonna happen, it looks bad and I don't think it would fit in my car.
So, I have to move everything inward. That will give me something to think about. Maybe I'll use some wood.
Get another car.
The rover's Pi hit my shared data limit today
The only way to stop the gearbox from moving around is to attach it to the steering knuckle arm (white nylon). I opened the gearbox and put two bolts through the back into the knuckle. All of this feels a lot stronger than the original front end, but some of the attachment points are weak and not finished.
Once the wiring is done, and the bolts going into the rover body are reinforced, test drive!
1st pic - Total fail. I put the actuator right in the way of the motor.
2nd pic - I thought it looked cool with the open gearbox.
3rd pic - AWD steering finished, for now
The actuator is installed and out of the way now. I must have not taken a final pic.
When you think the dimensions are the way you want, make a second version completely out of metal. Use some bronze or nylon bushings, and you have a steering unit you can bolt under that body, or to a complete chassis made out of metal.
Go for the gusto, but don't leave it in the desert alone.
Here's a video of the steering in action. For some reason the front motors won't turn. I can't wait to find out why, some other day
Any thoughts on that?
What about something with graphite in it, keep it dry. Does the gearbox seal up pretty well?
I think I'll go with heavy grease where the gears mate with the gearbox, there are no bearings in there. Keep the rest dry, adding anything to the teeth adds a lot of friction. The teeth of the gears do not mesh tightly, so the excess grease doesn't get pushed out.
Nope, motors still wouldn't turn. A week ago, I was AWD testing skid steering. It was a tank. Now it's dead. What happened?
The only difference is the front wheels turn on another axis. Same motors. It didn't make sense until this morning when I came to the realization that it can't stay in the house much longer. Not because it takes up too much space, which is does, more that it was born for the desert.
Grease in the teeth was an issue. My connection to the batteries turned out to be the biggest issue. I was using some cheap 12" jumpers. I swapped out the (light duty) clip leads and bolted the power wires to the battery terminals, and everything worked. The clips are much easier to disconnect than the bolts. Pulling the main fuse is easy, and that is what I have been doing. I don't know why I started using the clip leads. I thought they would get hot or smell funny if there was an issue. I did think about it when I was doing it. Now I can laugh about it.
We're closer than ever to the 2nd beach test. I have high hopes. It's back to tank status!
Jim
Now that the ball is rolling again, steering feedback will be next. I REALLY like these flex sensors. I've never owned one before. It seems a much easier choice than a potentiometer. I'll have to find a good mounting spot where nothing can interfere with it. For now I hot glued it so I can play around with code.
I've commented previously that clip leads are surprisingly expensive for what they are. We can put a man on the moon, but we can't make cheap, reliable clip leads!
Everything is connected and seems to be working. I have to add some more supports before the next sand test. I drove it around the yard and it did well. It has a tighter turning radius than before. I had to cut the fenders a little to get clearance.
Pic 1 - Larger rivets to hold the tread on, I ran out, I'll add some more before sand. They worked fine in the back yard on grass and dirt.
Pic 2 - I used zipties to do strain relief on the front motors. That seems to be working well.
Nice solution to a difficult problem.
Mike