The problem with this is that you can’t hit the microswitch if you don’t know
where you are, and you won’t know where you are until you hit the switches
there is no machine ever made that is always returned to it’s parking position
when you pause (especially if you are stopping because the cut isn’t what you
want)
With the maslow:
If you tell a motor to pull on the sled, you could rip the sled apart, make it
go off the top, or pull the sled into the motor.
If you tell a motor to relax the pull on the sled, you could have it hit the
ground, swing off the side of the machine (so it’s unsupported), or run the
chain off the motor so the sled swings freely.
Even if you had limit switches to keep you from pulling the sled into the motor,
or running the chain off the motor, until you know where you are, you can’t know
if it’s safe to pull or relax the pull on the sled.
The only thing you can do is tell the user to manually reset the
position.
This is either the full calibration of the manual calibration, which is turn the
motors manually until a point on the sprocket is pointing straight up and then
put a known amount of chain past that point (by moving a pre-marked link to that
point)
now, it is always possible that we have all missed something, there’s no such
thing as a dumb question, so think about how the machine moves, and where you
would try to position your limit switches and see if you’ve spotted something
we’ve missed so far.
It is hard to know what won’t work when I haven’t completely grasped what does work. I’ll have to get my hands on one and spend some time thinking about it. My specialty is electronic engineering, not mechanical engineering So I’m a little out of my league.
About the near-vertical loss of acceleration/force: Yes this always going to be a problem. Oversizing the chain anchor points for the given work-area is an obvious approach here, but of course there is a practical limit to how much you want to do this. I can see a tradeoff being made. We could make the tradeoff manually by setting low acceleration settings when using a large workspace and high settings in a small space.
Obviously having the planner dynamically figure this out would be the best, but my point is it may be workable before we have that.
About limit switches: How about limit lasers, like the ones you have on garage doors and race finish-lines?
Though I think worm gear is better overall, then you don’t have to deal with the sled dropping to the ground when you power down.
It’s the out of the comfort zone that makes us grow. I grew up around home grown creative manufacturing. I’m as comfortable with manual tools as digital ones. I’ve been an electronics hobbyist all my life too. When I built the early Makerbot Cnc 3D Printer kit I had no knowledge of microcontrollers. It’s the little projects out of the comfort zone that lead to bigger ones and suddenly you have a whole new skill set. I enjoy this space because the people in it are forcing me to stretch my skills. Keep being you.
Fairly high power, fast, brushless DC motors would be the goal, but if the Maslow motors are paid for and on hand, there is no real reason they couldn’t be used for development. Once the ODrive fork is proven to work, then research into a faster motor/gearbox set could follow.
Right now we really only do stall detection during calibration, but another wild idea…
Reduce the motor power to some (experimentally determined) low value. Suck the chain in in the reference side until it stops moving as seen by the encoder. This puts you near the top with relatively tight chain, maybe constrained by the second chain, maybe not. If you hit the limit switch then move to homing. If not start paying out the other chain while moving until you do. The thought is that this will help you from dropping the second chain on the floor after running it off the pulley.
Now tighten up both chains, still reduced power do you don’t pull things apart. When you have both chains stalled and the home switch or switches triggered then you’re at a consistent known point. These home switches might be hitting the sled, or maybe the chain is at the right angle to trigger them, again some experimenting. The chain on the opposite side should be at it’s steepest angle if i:'m visualizing this right, not the so-called homing side.
It’ll take some experimenting to find the safe power level, and and to figure out how to keep it on the backer (table, whatever), plus not busting stuff. This part is left as an exercise
Just a thought. Moose who live in swamps are all wet a lot
I tend to read ever faster now the amount of interesting posts is growing… I hope i can keep up!
Some questions come to mind:
Do brushless motors have some kind of holding torque?
Can you put DC on one a winding to hold it in place without burning outp a driver?
Are there affordable worm gears available for this kind of setup? Say 1/10 gear ratio
Is there a Odrive video that can give us an idea of Odrive vs gravity? Odrve lifting the full weight of a router up and down and still have enough power to do the woodworking. The speed of the Odrive is impressive though i’m still puzzled about the effective power in the direction against gravity.
I have more questions, but i’ll first try Duck an Google for those to keep this thread more sane
I tend to read ever faster now the amount of interesting posts is growing… I hope i can keep up!
Some questions come to mind:
Do brushless motors have some kind of holding torque?
not any more than bruched
Can you put DC on one a winding to hold it in place without burning outp a driver?
yes, as long as you keep it below the driver current (on odrive, 100A)
Are there affordable worm gears available for this kind of setup? Say 1/10 gear ratio
that’s the interesting question
Is there a Odrive video that can give us an idea of Odrive vs gravity? Odrve lifting the full weight of a router up and down and still have enough power to do the woodworking. The speed of the Odrive is impressive though i’m still puzzled about the effective power in the direction against gravity.
you can run 2Kw motors on the ODrive (up to 24v, 100A, to be bumped to 48v in
the next version), so these motors have a LOT of power, but most are designed to
spin very fast, so trading speed for power with a gear drive would be a very
good thing.
The stock maslow has a 291:1 worm drive, starting with a small BLDC motor you
could probably go to 900:1 or so and still fastly outperform the maslow motors.
Unfortunately I don’t think it will be very easy to make ODrive much cheaper, there is a fair bit of cost coming from just the processing required and so on.
Basically consider the minimum price for an ODrive controller (which is good for 2 motors) and a pair of motor/encoders to be $200.
This then only leaves the need for a Z-axis.
I’m thinking a little more long term here, and this BLDC really makes sense to me. I wish i had a budget to dive into this. At this point i can only say please keep at it. This stuff is awesome.
So I’m a little out of my league.
If we can only figure out a way to make it work without overspending.