@evanh said:
Cool. Yeah, I've only seen hydraulic benders.
Well funnily enough, I am a huge fan of hydraulics. I actually refused to sell my electric machines to CAT. Why? Because they expected better quality product from the machine.
No. Unlike conventional hydraulic machines with dumb valves and cylinders, I closed the loop on absolutely everything hydraulic and electric. My machines were always far more repeatable than the actual material. This is why I'm interested in what Ray is doing here. It's all wrong 😂 theoretically (like hydraulics) but when you are able to do whatever with the control loop, it's amazing what can be achieved.
We don't have resolution and that's the problem but we might be able to get this thing to behave, somewhat 😁
Car exhaust?
I'll have you know that they are now "emission control systems" and no BS because the tolerances are mind blowing.
I insulted Boeing when they were puzzled that I wasn't overly concerned about their tolerance spec's
"Man you guys are almost as tight as the exhaust pipe guys" 🤣 😂
And it was true. Yeah, no longer a sewer system for the ICE but a very critical tuning component.
I don't actually know but they may have just been aftermarket mods. ie: what the hoons wanted made for maximum noise emissions. Presumably parts importers and official service centres were also bringing in complete exhaust sets for most every car on the road.
Actually, I don't know the bender was exclusively used for cars either.
Cars aren't my thing, I never asked questions about the business. EDIT: I remember .. The business name originally was Magoo Mufflers. Then they named themselves Autobend, same as the bending machine.
Hi
Ray-
Is there anything you haven't attempted to do with the prop?
invariably you succeed!
This is looking good
Have a look at this guy- a high achiever- all his stuff is public domain I believe..
And boy- does he do balancing....
@"Christof Eb." Well, suppose it's not horrible. But, the idea is for this to have a camera and be a way to remotely view and area.
Don't think this thing is steady enough yet for a camera...
Fortunately, bought two pairs of motors and some caster wheels from Parallax. So, have a fallback plan to set in motion...
Did move the battery up higher and it's interesting. It is steadier when upright, but it is also more unstable and can't correct itself much if pushed.
It should work better. Have to take a more detailed look at what is going on ...
Problem with motor up high is that there's no P value that both maintains a steady upright position and can recover when pushed.
Thinking about making motor drive non-linear so gives more force when error is high...
I meant: If you try to rotate a wheel by hand with moderate force left and right, can you feel it moving easily for a small angle before it stops. If you can feel such backlash, then this is also the travel, that the motor has to do before it's counter movement has any effect, if the direction changes.
Thinking about the application again, maybe could just have it turn the motors off and park instead of standing still upright. That would be just as good, maybe better...
If only turned on motors when wanted to move, might behave well enough...
Mechanical lost motion notwithstanding, the low resolution of 360/144 = 2.5 degrees of deadband (aka: backlash). The loop won't respond until ít sees the next encoder count.
Wound up moving the battery back down the way it was. Just need to manage expectations.
Do want to try something with the gear play though. Thinking that can monitor for a change in direction if when that happens, give it a jolt of full power to overcome the slip...
Also, need to figure out how to get it to park itself gently when doesn't need to move...
I wonder if it would be acceptable and work to let the machine always spin slowly with the vertical axis. So the play would be irrelevant. Perhaps some gyro effect could even add stability. Going straight would be more difficult....
Mickster keeps mentioning the encoder. Do you use it here?
Comments
Cool. Yeah, I've only seen hydraulic benders.
PS: I even repaired a British build one ... 20-25 years back ... I think Autobend was the name.
PPS: It was used only for making car exhaust pipes.
Well funnily enough, I am a huge fan of hydraulics. I actually refused to sell my electric machines to CAT. Why? Because they expected better quality product from the machine.
No. Unlike conventional hydraulic machines with dumb valves and cylinders, I closed the loop on absolutely everything hydraulic and electric. My machines were always far more repeatable than the actual material. This is why I'm interested in what Ray is doing here. It's all wrong 😂 theoretically (like hydraulics) but when you are able to do whatever with the control loop, it's amazing what can be achieved.
We don't have resolution and that's the problem but we might be able to get this thing to behave, somewhat 😁
Car exhaust?
I'll have you know that they are now "emission control systems" and no BS because the tolerances are mind blowing.
I insulted Boeing when they were puzzled that I wasn't overly concerned about their tolerance spec's
"Man you guys are almost as tight as the exhaust pipe guys" 🤣 😂
And it was true. Yeah, no longer a sewer system for the ICE but a very critical tuning component.
I don't actually know but they may have just been aftermarket mods. ie: what the hoons wanted made for maximum noise emissions. Presumably parts importers and official service centres were also bringing in complete exhaust sets for most every car on the road.
Actually, I don't know the bender was exclusively used for cars either.
Cars aren't my thing, I never asked questions about the business. EDIT: I remember .. The business name originally was Magoo Mufflers. Then they named themselves Autobend, same as the bending machine.
Tried a few things today, nothing really improved things.
Also implemented wireless keyboard to move back and forth. This is not impressive...
Might try a lighter battery, mounted up higher, before moving on to 3 wheel bot...
Actually, maybe will try with this battery moved up higher and then give up...
Hi
I had a look at your video and actually thought that it was not too bad.
What is the tolerance that you would consider good?
Agreed. I have checked-out a few others where a lot more time has been invested and they still have the jitters. Not bad at all
Hi
Ray-
Is there anything you haven't attempted to do with the prop?
invariably you succeed!
This is looking good
Have a look at this guy- a high achiever- all his stuff is public domain I believe..
And boy- does he do balancing....
https://youtube.com/c/jamesbruton/videos
Dave
We need to get Bruton turned-on to the Prop. Seem to remember informing him some time ago.
@"Christof Eb." Well, suppose it's not horrible. But, the idea is for this to have a camera and be a way to remotely view and area.
Don't think this thing is steady enough yet for a camera...
Fortunately, bought two pairs of motors and some caster wheels from Parallax. So, have a fallback plan to set in motion...
Did move the battery up higher and it's interesting. It is steadier when upright, but it is also more unstable and can't correct itself much if pushed.
It should work better. Have to take a more detailed look at what is going on ...
Can you feel backlash / play in the gear train of these motors?
@"Christof Eb." IDK, not a motor person...
Problem with motor up high is that there's no P value that both maintains a steady upright position and can recover when pushed.
Thinking about making motor drive non-linear so gives more force when error is high...
I meant: If you try to rotate a wheel by hand with moderate force left and right, can you feel it moving easily for a small angle before it stops. If you can feel such backlash, then this is also the travel, that the motor has to do before it's counter movement has any effect, if the direction changes.
@"Christof Eb." Yes, there is some play. Amounts to about 1/2" of travel there. Any way to overcome that?
I did see that it is pretty steady when moving. Maybe that's the answer, just don't let it completely stop...
Thinking about the application again, maybe could just have it turn the motors off and park instead of standing still upright. That would be just as good, maybe better...
If only turned on motors when wanted to move, might behave well enough...
Mechanical lost motion notwithstanding, the low resolution of 360/144 = 2.5 degrees of deadband (aka: backlash). The loop won't respond until ít sees the next encoder count.
Wound up moving the battery back down the way it was. Just need to manage expectations.
Do want to try something with the gear play though. Thinking that can monitor for a change in direction if when that happens, give it a jolt of full power to overcome the slip...
Also, need to figure out how to get it to park itself gently when doesn't need to move...
I wonder if it would be acceptable and work to let the machine always spin slowly with the vertical axis. So the play would be irrelevant. Perhaps some gyro effect could even add stability. Going straight would be more difficult....
Mickster keeps mentioning the encoder. Do you use it here?
Spinning should work too. Have to try that. There are encoders but not using them yet..
Tried a couple more things that didn't really work out...
Spinning looks like it would improve stability, but not so practical while tethered with USB cable.
Am able to make it park itself the way wanted like this:
loop++ if loop == 98 pidt:=-1100 if loop ==100 m1.stop() m2.stop()So, this looks like the solution. Move to where it's needed and then park...