RepRap using DC Motors!
Posted by lucasmun09
|
Re: RepRap using DC Motors! July 22, 2014 10:09PM |
I have been myself thinking about DC motors instead of steppers, with some kind of encoder.
I always want to push faster, and stepper motor is not well suited for higher speeds, the question is still having the accuracy as well.
A very cheap DC motor could run much faster than a stepper, with way higher torque (==more power!).
I really hope you can make this succeed!
I always want to push faster, and stepper motor is not well suited for higher speeds, the question is still having the accuracy as well.
A very cheap DC motor could run much faster than a stepper, with way higher torque (==more power!).
I really hope you can make this succeed!
|
Re: RepRap using DC Motors! July 22, 2014 10:17PM |
WHY all this negativity?
Every time someone present a idea not conforming the standard people are naysaying!
As for PWM for DC motors -> that's actually the way most things control speed for DC motors!
DC motors can take a huge punishment, and last quite a long time, just look at RC cars - well during the time DC motors were the standard still! It's a choice between brush hardness and brush tension, probably lowish tension and semihard brushes could be good for this, long lasting and not ruining the collector as fast.
Every time someone present a idea not conforming the standard people are naysaying!
As for PWM for DC motors -> that's actually the way most things control speed for DC motors!
DC motors can take a huge punishment, and last quite a long time, just look at RC cars - well during the time DC motors were the standard still!
It's a choice between brush hardness and brush tension, probably lowish tension and semihard brushes could be good for this, long lasting and not ruining the collector as fast.
|
Re: RepRap using DC Motors! July 23, 2014 06:02PM |
Quote
PulsedMedia
WHY all this negativity?
Every time someone present a idea not conforming the standard people are naysaying!
Whats being said is that it will take days of work to get this working, when for $11.80 usd at robotdigg.com you can have yourself a working stepper motor + driver.
But, once this is made, a junkpunk bot will be a lot more feasable, if this design proves effective.
I was thinking about this topic too as I have inkjet parts lying around.. as do many people. But, is it worth it?
Perhaps. You will learn a lot, that's for sure, but be prepared in case you fail.
That being said we learn most from our failures. I'll keep an eye on this thread
|
Re: RepRap using DC Motors! July 23, 2014 06:22PM |
... I have the guts from several inkjet-printers lying around somewhere (salvaged them for this sort of tinkering) - but got/found then all sorts of high quality gear around steppers, servos, encoders and such ...
Snd so I've refurbished a CNC-mill with servos+encoders for laser-engraving -- it's much more accurate and fast now, compared to stepper motors ... but compared in costs, this would be nearly twice the price
Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]
Call for the project "garbage-free seas" - [reprap.org]
Snd so I've refurbished a CNC-mill with servos+encoders for laser-engraving -- it's much more accurate and fast now, compared to stepper motors ... but compared in costs, this would be nearly twice the price
Viktor
--------
Aufruf zum Projekt "Müll-freie Meere" - [reprap.org] -- Deutsche Facebook-Gruppe - [www.facebook.com]
Call for the project "garbage-free seas" - [reprap.org]
|
Re: RepRap using DC Motors! August 02, 2014 04:21PM |
Quote
VDX
... I have the guts from several inkjet-printers lying around somewhere (salvaged them for this sort of tinkering) - but got/found then all sorts of high quality gear around steppers, servos, encoders and such ...
Snd so I've refurbished a CNC-mill with servos+encoders for laser-engraving -- it's much more accurate and fast now, compared to stepper motors ... but compared in costs, this would be nearly twice the price
I checked out some of your recent video clips and found it looks good. Are you using a dedicated driver (and microprocessor/DSP) for each motor? What is the max input rate accepted by the driver (assume it is pulse/dir driven)? What is the motor's gear ratio and encoder count (assume encoder is mounted directly on the motor shaft)? I thought that the DC motor and driver would cost much more than twice the price of comparable stepper mechanism.
Edited 4 time(s). Last edit at 08/02/2014 04:25PM by vn.
|
Re: RepRap using DC Motors! November 18, 2014 09:15PM |
I may be stating the obvious but you guys ARE talking about PID control of a DC motor, right? The important thing is the encoder, not the DC motor. You can use the encoder with the stepper too because with the encoder you have FEEDBACK. A magic word All it takes is one mis-step of the stepper and it never knows it is now always one-step off. Feedback can keep that from happening. (or more accurately: fix it on the fly)
All it takes is one mis-step of the stepper and it never knows it is now always one-step off. Feedback can keep that from happening. (or more accurately: fix it on the fly)
|
Re: RepRap using DC Motors! December 10, 2014 02:37PM |
Hi man!
I don't know how far you've gone with your project (since I didn't read all of the posts...) but I'm tryin to do the same thing using a different microcontroller than arduino (from ST microelectronics) 
I'm a young control engineer and I have already worked with "handmade" servo before...here's my post on the italian reprap forum:
[forums.reprap.org]
I'm still workin at it (and I think It'll take quite a long time...), but we can share knowledge and help each other maybe. If you make any progress let me know about, if you have any questions don't hesitate to ask.
Cheers
A.
EDIT: @zavashier, I'm afraid I have to tell you're absolutely wrong. A well-controller DC motor can be as fast and precise as no stepping motor can be, the strip in the printer is part of a linear encoder, which is a position sensor that can be used to close one or more control feedback loops.
Edited 1 time(s). Last edit at 12/10/2014 02:41PM by ApBrandy.
I don't know how far you've gone with your project (since I didn't read all of the posts...
) but I'm tryin to do the same thing using a different microcontroller than arduino (from ST microelectronics) I'm a young control engineer and I have already worked with "handmade" servo before...here's my post on the italian reprap forum:
[forums.reprap.org]
I'm still workin at it (and I think It'll take quite a long time...), but we can share knowledge and help each other maybe. If you make any progress let me know about, if you have any questions don't hesitate to ask.
Cheers
A.
EDIT: @zavashier, I'm afraid I have to tell you're absolutely wrong. A well-controller DC motor can be as fast and precise as no stepping motor can be, the strip in the printer is part of a linear encoder, which is a position sensor that can be used to close one or more control feedback loops.
Edited 1 time(s). Last edit at 12/10/2014 02:41PM by ApBrandy.
|
Re: RepRap using DC Motors! December 10, 2014 05:20PM |
Yep, and you can call that a servomotor 
Collective intelligence emerges when a group of people work together effectively. Prusa i3 Folger (A lot of the parts are wrong, boring !)
Collective intelligence emerges when a group of people work together effectively. Prusa i3 Folger (A lot of the parts are wrong, boring !)
|
Re: RepRap using DC Motors! December 12, 2014 05:14PM |
Encoders are good things, and I like them.
DC motors are problematic, even with encoders as the accelleration is difficult to get right.
As far as CNC goes, all of the ones that I've seen (I work with several machine shops) either use high feedback servo style motors, or steppers. Practically all DIY CNC machines use steppers for the same reason that we do.
2D printers do some things a little different, which makes them better suited to high speeds using DC motors:
They only move 1 axis at any given time. The Internal logic is something like this: "Advance the page (Y axis) until it approaches 1 print head's worth of distance. Verify and adjust. This step happens while the print head (X axis) is reversing direction. For the X axis, measure speed, at [lead time] before you get to the point where you need a dot, trigger the print head to send a droplet of ink." Printers NEVER deliver ink on a diagonal path, and because ink is a lot less viscous than melted plastic, it can deliver tiny little droplets which it knows a very precise lead time to hit the surface of the paper, and it can do so in a rapid-fire sequence of droplets, instead of a contiguous stream.
I am not saying that it can't be done, but I am saying that it would take a very specialised control engine to do this, and it would most likely work best if you were to re-write the gcode to allow the printer to move only one axis at a time with fixed steps in the other direction. (We already do this to limit movement on the Z axis.) I think that the extruder would need to be fine-tuned to be able to deliver very precise amounts of material as well.
As is, with the stepper motors, the print head can be moved in a very defined way to have a "line" of material extruded at almost any X/Y ratio. With DC motors, you can approximate this with PWM control, but it would take a lot of R&D time to measure acceleration rates and calibration time to get this working, even if you only allow movement on the 90 degree planes and 45 degree angles, particularly since movement on the Y axis will vary in available acceleration because of the increasing mass due to the project being built on it over time. By comparison, the mass of a sheet of paper printed vs unprinted does not change much and is still rarely a significant portion of the moving mass of the paper feeding rollers and motors, whereas a 40% fill of ABS plastic in an object up to 150mm cubed probably DOES weigh in at a significant amount as compared to the print bed motors and belt of a typical RapRep.
So these are the hurdles that I can foresee you needing to overcome to make this work. I you combine the opto-encoders and use PWM output to the DC motors then you should be able to get enough accuracy to be able to get reasonable results. With enough R&D work, you might even get excellent results, with such things as the ability to pause/resume which could come in very handy for such things as multi-colour printing. You will also need other control electronics so that you can deal with the encoders, switch polarities as well as deal with the PWM output in order to get the control that you need. These are basically a handful of power transistors, which shouldn't cost much.
If you are working for $0/hour, and you have adequate electronics available to you that you can get this all done, it may be worthwhile.
By comparison though, I just bought a set of 5 76 oz-in stepper motors off of eBay for $37. Even if I were to value my own time at 1/2 minimum wage, I'm pretty sure that I'd come out ahead from this deal when compared to the development time needed to make the DC motors work okay.
Once the 3D printer is working, I'd print a couple of RC car chassis and use the free DC motors in those instead, but that's me.
DC motors are problematic, even with encoders as the accelleration is difficult to get right.
As far as CNC goes, all of the ones that I've seen (I work with several machine shops) either use high feedback servo style motors, or steppers. Practically all DIY CNC machines use steppers for the same reason that we do.
2D printers do some things a little different, which makes them better suited to high speeds using DC motors:
They only move 1 axis at any given time. The Internal logic is something like this: "Advance the page (Y axis) until it approaches 1 print head's worth of distance. Verify and adjust. This step happens while the print head (X axis) is reversing direction. For the X axis, measure speed, at [lead time] before you get to the point where you need a dot, trigger the print head to send a droplet of ink." Printers NEVER deliver ink on a diagonal path, and because ink is a lot less viscous than melted plastic, it can deliver tiny little droplets which it knows a very precise lead time to hit the surface of the paper, and it can do so in a rapid-fire sequence of droplets, instead of a contiguous stream.
I am not saying that it can't be done, but I am saying that it would take a very specialised control engine to do this, and it would most likely work best if you were to re-write the gcode to allow the printer to move only one axis at a time with fixed steps in the other direction. (We already do this to limit movement on the Z axis.) I think that the extruder would need to be fine-tuned to be able to deliver very precise amounts of material as well.
As is, with the stepper motors, the print head can be moved in a very defined way to have a "line" of material extruded at almost any X/Y ratio. With DC motors, you can approximate this with PWM control, but it would take a lot of R&D time to measure acceleration rates and calibration time to get this working, even if you only allow movement on the 90 degree planes and 45 degree angles, particularly since movement on the Y axis will vary in available acceleration because of the increasing mass due to the project being built on it over time. By comparison, the mass of a sheet of paper printed vs unprinted does not change much and is still rarely a significant portion of the moving mass of the paper feeding rollers and motors, whereas a 40% fill of ABS plastic in an object up to 150mm cubed probably DOES weigh in at a significant amount as compared to the print bed motors and belt of a typical RapRep.
So these are the hurdles that I can foresee you needing to overcome to make this work. I you combine the opto-encoders and use PWM output to the DC motors then you should be able to get enough accuracy to be able to get reasonable results. With enough R&D work, you might even get excellent results, with such things as the ability to pause/resume which could come in very handy for such things as multi-colour printing. You will also need other control electronics so that you can deal with the encoders, switch polarities as well as deal with the PWM output in order to get the control that you need. These are basically a handful of power transistors, which shouldn't cost much.
If you are working for $0/hour, and you have adequate electronics available to you that you can get this all done, it may be worthwhile.
By comparison though, I just bought a set of 5 76 oz-in stepper motors off of eBay for $37. Even if I were to value my own time at 1/2 minimum wage, I'm pretty sure that I'd come out ahead from this deal when compared to the development time needed to make the DC motors work okay.
Once the 3D printer is working, I'd print a couple of RC car chassis and use the free DC motors in those instead, but that's me.
Sorry, only registered users may post in this forum.