Komputer Numerically Kontrolled (KNK) Router
Moderators: Head Monkey, kelvin, bigKam, skidesmond, chrismp
Komputer Numerically Kontrolled (KNK) Router
DONE!
I worked 62.5 hours straight without sleep --- milling, lathing, drilling, welding, gluing, bolting, and writing custom software for my new custom-built CNC machine. It weighs 2000 lbs, runs at a max. linear speed of 18.23 ft./sec, consumes 6000 W of power, positions with +/- 10^-6 inch accuracy over 12 ft., eats live chickens, and is day-glow orange in color.
OK, so I'm kidding. Made you read, though.
The truth is I am actually going to build a machine. Whether it eats live chickens will depend on what parts I pull together.
I have decided to document my little adventure, with the hopes of sharing as much information as possible. My diary may do any one of the following: inspire, motivate, or discourage. Hopefully at the very least the common denominator will be that the information will make an impact.
To be honest I haven't put much thought into this project yet. A while back, a friend of a friend donated parts to an old ShopBot to Kelvin and I. Kelvin put some work into it and created the SB CNC machine which some of you have seen. It works beautifully. Now I want my own.
At this point, there's no set design or 'plan' that I'm following. It's going to develop as it goes, so it will be interesting.
Let's get started.
Last weekend I visited my family and cousin Kam (littleKam) in So. California. While at my brother Pat's house, I picked up four Ametek servomotors. Last year Pat converted a standard manual milling machine into a CNC machine and he had some motors laying around not in use. I was glad to take them off his hands, and here they are:
They are Ametek 965922-101 servomotors, rated at 12 amps (peak), nominal 38 VDC, and I think capable of delivering a whopping 600 oz-inches of torque. The 5/8" diameter output shaft seems to suggest that they can deliver some umph! Each motor weighs a fair amount, too. On the bottom of the motor is an optical encoder. I don't know the counts/revolution, but will figure it out at some point.
Some CNC's use stepper motors, DC motors that are designed to move in steps, for example, 200 steps per 360-degrees. Steppers are easy to work with and can be operated without a sensor feeding back information about what position the shaft is in. In this case, the motor can be run in open-loop. Steppers are controlled by applying pulses to step the motor from on position to the next. It's a simple setup and great for CNC-type machines where a tool has to be positioning relative to a workpiece in 3D.
Other CNC's use servomotors, a motor with a sensor attached to the motor's shaft. A controller is needed to sense the shaft's position and then adjust the input voltage/current appropriately to control the (angular) position of the shaft. These motor/sensor/controller systems are a bit more complicated, and if the sensor was to ever fail with the controller active, the motor can spin out of control to infinity and beyond. Of course there are ways to minimize this, but in general they are slightly more complicated.
There are advantages and disadvantages of using either steppers or servos. Everybody seems to have their own opinion, so I'm not going to contribute to the debate. However, I've been intrigued by servos and have built similar setups at my day job, so I figured since the motors were free I'd make the best of them. As for the controller and power supply, it depends on my budget and avaiable time, but I'll try to build as much from scratch as possible without nuking myself --- bzzz, bzzzz... I always keep in mind that electricity is extremely dangerous and always double, triple check my connections before turning on power. It doesn't take much current to stop one's heart!
Anyway, in addition to the motors, I also picked up some modular aluminum framing to create the skeleton of the CNC machine. I haven't decided on the size of the machine yet. For now I'm just collecting parts and dreaming.
I will do my best to update this thread as I go and hope that it materializes into something worthwhile. Feel free to ask or inquire about anything or better yet, make suggestions and point me in the right direction. Until next time.
I worked 62.5 hours straight without sleep --- milling, lathing, drilling, welding, gluing, bolting, and writing custom software for my new custom-built CNC machine. It weighs 2000 lbs, runs at a max. linear speed of 18.23 ft./sec, consumes 6000 W of power, positions with +/- 10^-6 inch accuracy over 12 ft., eats live chickens, and is day-glow orange in color.
OK, so I'm kidding. Made you read, though.
The truth is I am actually going to build a machine. Whether it eats live chickens will depend on what parts I pull together.
I have decided to document my little adventure, with the hopes of sharing as much information as possible. My diary may do any one of the following: inspire, motivate, or discourage. Hopefully at the very least the common denominator will be that the information will make an impact.
To be honest I haven't put much thought into this project yet. A while back, a friend of a friend donated parts to an old ShopBot to Kelvin and I. Kelvin put some work into it and created the SB CNC machine which some of you have seen. It works beautifully. Now I want my own.
At this point, there's no set design or 'plan' that I'm following. It's going to develop as it goes, so it will be interesting.
Let's get started.
Last weekend I visited my family and cousin Kam (littleKam) in So. California. While at my brother Pat's house, I picked up four Ametek servomotors. Last year Pat converted a standard manual milling machine into a CNC machine and he had some motors laying around not in use. I was glad to take them off his hands, and here they are:
They are Ametek 965922-101 servomotors, rated at 12 amps (peak), nominal 38 VDC, and I think capable of delivering a whopping 600 oz-inches of torque. The 5/8" diameter output shaft seems to suggest that they can deliver some umph! Each motor weighs a fair amount, too. On the bottom of the motor is an optical encoder. I don't know the counts/revolution, but will figure it out at some point.
Some CNC's use stepper motors, DC motors that are designed to move in steps, for example, 200 steps per 360-degrees. Steppers are easy to work with and can be operated without a sensor feeding back information about what position the shaft is in. In this case, the motor can be run in open-loop. Steppers are controlled by applying pulses to step the motor from on position to the next. It's a simple setup and great for CNC-type machines where a tool has to be positioning relative to a workpiece in 3D.
Other CNC's use servomotors, a motor with a sensor attached to the motor's shaft. A controller is needed to sense the shaft's position and then adjust the input voltage/current appropriately to control the (angular) position of the shaft. These motor/sensor/controller systems are a bit more complicated, and if the sensor was to ever fail with the controller active, the motor can spin out of control to infinity and beyond. Of course there are ways to minimize this, but in general they are slightly more complicated.
There are advantages and disadvantages of using either steppers or servos. Everybody seems to have their own opinion, so I'm not going to contribute to the debate. However, I've been intrigued by servos and have built similar setups at my day job, so I figured since the motors were free I'd make the best of them. As for the controller and power supply, it depends on my budget and avaiable time, but I'll try to build as much from scratch as possible without nuking myself --- bzzz, bzzzz... I always keep in mind that electricity is extremely dangerous and always double, triple check my connections before turning on power. It doesn't take much current to stop one's heart!
Anyway, in addition to the motors, I also picked up some modular aluminum framing to create the skeleton of the CNC machine. I haven't decided on the size of the machine yet. For now I'm just collecting parts and dreaming.
I will do my best to update this thread as I go and hope that it materializes into something worthwhile. Feel free to ask or inquire about anything or better yet, make suggestions and point me in the right direction. Until next time.
Last edited by bigKam on Tue Mar 31, 2009 12:20 am, edited 2 times in total.
bigKam,
He,he... I actually believed every word of that first paragraph when I read it... till I got to the "eats live chickens" part.
This is certain to be a really exciting venture. How large of a power supply are you considering in terms of DC voltage out and amp rating? Do you have a preference for linear vs. regulated? I have a lot of parts sources filed away for reference.
While you're in the dream stage, here's something to consider regarding your linear drive system. Most long table systems use rack and pinion drives because ACME screws or ball screws tend to whip and cause resonance problems when the screw length gets much over 4 or 5 feet. One mitigation to the whipping problem associated with a long lead screw is to use a relatively large diameter screw, which has the down-side of being a huge inertia sink when changing directions of travel. I'm assuming (duh) that you're going to want a long table system so that you can cut ski parts. I worked out a very simple drive system that uses a 104" long ACME lead screw (1/2 inch diameter... nice and light) that doesn't whip at speed (and has zero backlash). The drive screw cost about $60 vs. many hundreds of dollars for a rack and pinion system (or thousands for a large diameter ball screw). So, if you're interested, I'll share that drive screw design with you.
Nice summer project.
G-man
He,he... I actually believed every word of that first paragraph when I read it... till I got to the "eats live chickens" part.
This is certain to be a really exciting venture. How large of a power supply are you considering in terms of DC voltage out and amp rating? Do you have a preference for linear vs. regulated? I have a lot of parts sources filed away for reference.
While you're in the dream stage, here's something to consider regarding your linear drive system. Most long table systems use rack and pinion drives because ACME screws or ball screws tend to whip and cause resonance problems when the screw length gets much over 4 or 5 feet. One mitigation to the whipping problem associated with a long lead screw is to use a relatively large diameter screw, which has the down-side of being a huge inertia sink when changing directions of travel. I'm assuming (duh) that you're going to want a long table system so that you can cut ski parts. I worked out a very simple drive system that uses a 104" long ACME lead screw (1/2 inch diameter... nice and light) that doesn't whip at speed (and has zero backlash). The drive screw cost about $60 vs. many hundreds of dollars for a rack and pinion system (or thousands for a large diameter ball screw). So, if you're interested, I'll share that drive screw design with you.
Nice summer project.
G-man
Hey G-man,
Do you mean switch-mode vs. linear regulated? I'm not sure yet since I haven't sat down to crunch the numbers in terms of the amount of power I need. I've built several switch-mode and linear power supplies before, but not at the power levels that I'll need for the CNC. I will talk to you about the parts, though.
I'm very interested in your $60 setup. Wow, sounds like a sweet deal. I have been looking at purchasing a long drive screw, but did not consider the resonance issue. This is interesting. Let's chat more about this...
Boy, this is so much fun...
Do you mean switch-mode vs. linear regulated? I'm not sure yet since I haven't sat down to crunch the numbers in terms of the amount of power I need. I've built several switch-mode and linear power supplies before, but not at the power levels that I'll need for the CNC. I will talk to you about the parts, though.
I'm very interested in your $60 setup. Wow, sounds like a sweet deal. I have been looking at purchasing a long drive screw, but did not consider the resonance issue. This is interesting. Let's chat more about this...
Boy, this is so much fun...
Yes, that is actually what I meant. On the CNC forums, we most often use the term 'regulated' when talking about regulated switching supplies, and use 'linear' when talking about linear unregulated supplies (admittedly not correct... we've gotten a bit lazy in our use of terms). Most of the folks on the hobbycnc.com forum site are using unregulated linear power supplies. Not many builders add the regulating circuitry to their unregulated linear supplies, for some reason. As you know, it's pretty easy to add. My supply is an unregulated linear supply that puts out 36v DC, rated at 10 amps. My driver board chips can handle up to 42v, but, I only feed the board the usual 36v in case I get any supply grid voltage spikes. It's a very simple supply design, which is probably why it is as popular as it is... heavy, though. One of the main reasons I like the unregulated linear supply is it's simplicity. If it goes bad, I just have to troubleshoot 3 parts, replace the bad one, and I'm back in business. One of the best deals I've seen on regulated switch mode supplies is located here:
http://www.mpja.com/prodinfo.asp?number=17429+PS
I don't know if that's enough juice for your application, but it's sure a good deal. A while back, they had 36v/12.5 amp supplies for 25 bucks. I sure wish I'd picked up a few of those.
I'm sure that resonance is a bigger deal when using stepper motors, than with servo motors, because the resonance causes the steppers to lose steps, ruining the cut. I tried to run my lead screw without my adaptation, and it wobbled wildly (like 2 to 3 inches out of center) at very low rpm (driven with a drill motor rather than a stepper). That kind of wobble wouldn't cut it, no matter what kind of motor was used.
Yes, this is indeed a lot of fun and I'm betting we'll be talking a bunch..
G-man
http://www.mpja.com/prodinfo.asp?number=17429+PS
I don't know if that's enough juice for your application, but it's sure a good deal. A while back, they had 36v/12.5 amp supplies for 25 bucks. I sure wish I'd picked up a few of those.
I'm sure that resonance is a bigger deal when using stepper motors, than with servo motors, because the resonance causes the steppers to lose steps, ruining the cut. I tried to run my lead screw without my adaptation, and it wobbled wildly (like 2 to 3 inches out of center) at very low rpm (driven with a drill motor rather than a stepper). That kind of wobble wouldn't cut it, no matter what kind of motor was used.
Yes, this is indeed a lot of fun and I'm betting we'll be talking a bunch..
G-man
I worked briefly for a company that converted machines over - they built them for a place in Corona,Ca that built airplane actuators for trim tabs. I soldered the boards together while another hooked up hardware.
It really worked well and it looks like your on the same track
My bat cave is still full of computers
It really worked well and it looks like your on the same track
My bat cave is still full of computers
I used to be a lifeguard, but some blue kid got me fired.
I set this project aside back in July, but picked it back up in early December. Made good progress, but the slowest part is waiting for parts to arrive in the mail. Here's what I envision it will look like:
Here's a sneak peek at an x (long) axis rail -- the rails will be mounted to the vertical plane of the aluminum structure:
Several parts I farmed out to my brother for him to make since he has a CNC mill in his garage. He built the mill several years ago and loves to make parts. The rest of the parts I will do my best with tools I have on hand. The more complicated work such as machining the ball screws I handed over to a local machine shop. It's not trivial turning down a 100-inch long screw. As for the electronics, I'm using Gecko servo motor drivers, but built the power supply from discrete components.
The router will have over 85" of travel along x; 45" along y; and roughly 5" in z. Based on my calculations, the gantry should be able to run at 200 inches-per-minute before any whipping will occur. This number is somewhat conservative. I will follow up with more details, so stay tuned....
Here's a sneak peek at an x (long) axis rail -- the rails will be mounted to the vertical plane of the aluminum structure:
Several parts I farmed out to my brother for him to make since he has a CNC mill in his garage. He built the mill several years ago and loves to make parts. The rest of the parts I will do my best with tools I have on hand. The more complicated work such as machining the ball screws I handed over to a local machine shop. It's not trivial turning down a 100-inch long screw. As for the electronics, I'm using Gecko servo motor drivers, but built the power supply from discrete components.
The router will have over 85" of travel along x; 45" along y; and roughly 5" in z. Based on my calculations, the gantry should be able to run at 200 inches-per-minute before any whipping will occur. This number is somewhat conservative. I will follow up with more details, so stay tuned....
Put a few more hours into this project after work tonight. Assembled part of the gantry (y-axis and z-axis). Used a hand drill to test the travel of the z-axis -- got 4.75" of travel. Not bad. Y travel is more than 41". Overall, these two axes feel stiff, just what I wanted. My brother Pat machined some of the components for the z-axis and he did a great job.
Here's a quick video of the z-axis in motion and sliding along y [click here (4MB)].
Some photos:
More to come...
Here's a quick video of the z-axis in motion and sliding along y [click here (4MB)].
Some photos:
More to come...
- Head Monkey
- Posts: 310
- Joined: Sun Jul 17, 2005 4:53 pm
- Location: Carnation, WA
- Contact:
Very nice work man.
Everything I know about snowboard building, almost: MonkeyWiki, a guide to snowboard construction
Free open source ski and snowboard CADCAM: MonkeyCAM, snoCAD-X
Free open source ski and snowboard CADCAM: MonkeyCAM, snoCAD-X
-
- Posts: 10
- Joined: Wed Mar 11, 2009 6:34 pm
- Location: Eastern US