One of my friends told me I should adapt it to decorate cakes. I thought that was an interesting idea & quite doable. The wheels would just have to have more elevation around the cake. Not sure I will try it anytime soon though.
1 Like
Tool changer with a rake…
3 Likes
🧙♂
1 Like
Speaking of automagic litter boxes. An ex-girlfriend and I bought one for her cat. The cat didn’t like it, so it would hold it all day until the middle of the night. Then it would pee right where the rake would shift down into the litter. Clumping little makes little nuggets where they pee, unless you wait all day, then it makes a chunk about the size of a brick. Then we’d get to listen to the damn thing slip it’s clutch for a couple minutes until it gave up. I’d call her a stupid cat, but that’s actually kinda genius if you don’t like your new litter box. This is the same cat that would leave partial mouse bodies in my pants. Rhiannon said it’ was because she loved me and was giving me gifts. I said it was the equivalent of a mafia horse head.
7 Likes
I have some TMC2209 drivers that could make it quiet litter box cleaner. Think I am going to change the Z-axis to use some small stepper as that servo motor is just too jerky & I have to modify gcode. I was going to try the inkcut plugin for inkscape. After installing 6.5gb of Visual Studio & compiling it, it almost worked. I am getting closer & closer to switching to Linux.
I have a 28BYJ-48 5v stepper motor. According to this page .05v Vref would be a good starting point for it with a DRV8825 & see if it gets too warm & is strong enough. https://arduino.stackexchange.com/questions/72731/can-i-control-a-28byj-48-5v-stepper-with-a-dvr8825-driver
Are there any other suggestions for a small inexpensive stepper motor. The nice thing about this is that it doesn’t need much weight. Guess I could use a bigger motor just in case I add something heavier to it.
2 Likes
I used a little 28BYJ-48 stepper motor with a printed linear slide to fashion a light-weight Z-axis for a laser engraver I was once playing with. It employed a simple bipolar hack that greatly increases the torque of the little stepper and enables it to be used with A4988/DRV8825 stepstick drivers (please forgive… I’m hard of hearing and forgot to turn down the loud TV)…
Check out the videos from the Thingiverse page for the linear slider. Converted to bipolar, the motor is capable of far more torque than unipolar and can be driven from a Ramps with A4988 or DRV8825 drivers, just like our beloved NEMA17’s. To set it up, simply take off all the micro-stepping jumpers (the motor is geared way down) and turn down the Vrefs… the motor can only handle 150 ma or so. Then go into the firmware and set the steps/mm setting to a value that moves it the distance required (I use this procedure to determine the proper steps/mm setting). Please recognize precision is NOT the little motor’s forte… but it’s more than good enough for applications like the above.
2 Likes
Even with the geared stepper i have plenty sitting around that is cool
3 Likes
Thanks for the links. That linear slider might be just what I need.
2 Likes
I found this design on Grabcad that uses a nema11 stepper. I have a spare pancake nema 17 that weighs only 135 grams (4.6oz) that I might try a design with this as a starting template. This pancake is lighter than a nema 11 that I found on amazon.
https://grabcad.com/library/cnc2040-spring-loaded-zaxis-nema11-1
1 Like
I redesigned the wheel drive to use double helical drive. Initially I made a 33T drive gear, but it was too wobbly when moving the gears around. I then tried a 16T drive gear & changed the wheel spacing to fit, but then it was not going to be possible to adjust the motor mounts screws after the wheels were on. The next option was to add another gear between the drive gear & bearings & think that will work. I also had the bright idea to add another o-ring to outer part of drive gear & o-ring grooves on the in between gear to keep them in alignment. Looks like it should work. I am also trying to figure out the
DEFAULT_AXIS_STEPS_PER_UNIT in Marlin for a geared drive. Does anyone know the proper method to calculate that? I found https://www.matterhackers.com/news/3d-printer-firmware-settings-stepper-motor-configuration this, but it is for extruder motor. I might just start with the settings for a 12T drive gear with GT2 belt & adjust it until correct if I can’t find a better starting point. I am using 12T drive & in between bearing gears & 22T wheel gear with a gear ratio of 1.83:1 Here is a current screen shot of the wheel drive design.
3 Likes
I’m loving where you’re going with this…
You really don’t have to know anything about the drive method. If it consistently moves a given distance when told to move X distance… simply tell it to move some convenient amount (i.e. 100mm, 50mm, etc… the bigger the better) and then measure how far it actually moved. Then simply multiply the currently stored value for steps/mm by (commanded / actual) and use that as your new steps/mm. For example, tell it to move 100mm and it actually moved 160… if the current steps/mm is 250, multiply it by 0.625 (i.e. 100/160) and your new steps/mm becomes 156.25. Rerun the test… command 100mm move and verify it moves 100mm. Do it as many times as needed to home in on the proper value.
I describe this in this recent post and also in my R&P MPCNC thread and verify by printing/engraving accurate rulers.
1 Like
Let N1 be the number of teeth on the gear on your motor.
Let N2 be the number of teeth on your wheels.
Let D be the outer diameter of the o-rings on your wheels.
Let M be the microstepping factor, most likely 16 or 32.
Let S be the number of full steps per revolution of your motor, generally 200.
M is microsteps / full step
S is full steps / motor revolution
M*S is microsteps / motor revolution
N1 is gear teeth / motor revolution
M*S/N1 is then microsteps / gear tooth
N2 is gear teeth / wheel revolution
(M*S/N1)*N2 is microsteps / wheel revolution
(the intermediate gear doesn’t matter)
Let C be the circumference of the wheel, which is pi*D
The circumference of the wheel is linear movement / wheel revolution
(M*S/N1)*N2/C is microsteps / linear movement
or
(M*S/N1)*N2/(pi*D)
or
(M*S*N2)/(N1*pi*D)
This should get you close, and then if you want accuracy you should use @dkj4linux’s method to dial in the rest. Or you could do that from the start if you have an aversion to the math.
2 Likes
While I deeply appreciate those who can pencil-whip problems… math has never been my strong point. Even when I was doing this stuff professionally, I was more a seat-of-the-pants engineer… probably because I was a technician before I was an engineer 
The astounding thing [to me] about the method I described is that while needing to know virtually nothing about the drive method, the components involved, or the source(s) of any repeatable error… the method rapidly “converges” to the proper steps/mm regardless of the magnitude of the error. Say you inadvertently left off a micro-stepping jumper and it’s off my a factor or 2, or 4, or 16… or used a 20-tooth pulley where it should have been a 16-tooth pulley… it still works
3 Likes
I will start with Jamie’s formula 1st & use dkj4linx method after that. I was going to use the method used for calculating e-steps after getting a starting point which seems like dkj4linx method.
The numbers I get for that formula are:
N1 12
N2 22
D 59.5
M 32
S 200
Using the 1st formula I get 62.77 & the 2nd formula I get 18.675
I will start with the 62.77 unless someone comes up with a better idea. I should be ready to test this again tomorrow. I still have to print 4 gears & another side plate. I cheated on not having to reprint 1 sideplate & made a plastic drill jig to drill the 2 slotted holes for the new bearing gears.
1 Like
Thanks for the help. I will post a short video when I test it again.
1 Like
I tested it today & having that extra gear in between is too difficult to get a good mesh. I used 62 for the steps per unit & drew a 10mm line to start with & it was 7.5mm, so that 1st formula seems to be a good starting point. The gears didn’t go well, so that number might be closer to 10mm if the gears meshed well. I am going back to the single 33T drive gear. I have a couple of ideas to make that large gear not wobble. In any event, it is not nice meshing of the gears with just that one gear to adjust. That 62 will be about 23 with this larger drive gear.
1 Like
It appears that with the 33T drive gear there is not enough torque to turn the wheels. I adjusted the vref from my initial .47v on the DRV8825 to .65v & then .91v & the wheels still will not turn. The drive gear will turn when the wheels are not engaged. I also tried it with the wheels off the ground to see if they would turn. I either need to go to a smaller tooth drive gear or go back to my original belt driven method. I might shorten the wheel offset & see if I can have motor mount access holes placed in the wheels. It is funny that my 1st idea worked the best. I have some 105zz 5x10x4 bearings I might try for making a smaller idler pulley with the belt design.
1 Like
33T would have about half the torque of a 16T. Can you make the wheels bigger?
1 Like
That wouldn’t help. He’d have more torque turning the wheels, but the wheels would need more torque to move the machine. There’s currently almost a 1:1 ratio between the motor pinion teeth and the distance the machine moves. All the gears in the train are meshed, so they don’t change anything.
How about moving the motor up (or down) and add another small pinion, driving the currently driven gear? Or one central gear that meshes with both wheels, driven by a small pinion?
2 Likes
Oh yeah. Unless the drive gear would get smaller because the wheels get bigger, but you’re definitely right.
1 Like