Text Size

New Build- Blacktoe Clone with Belt Reduction

New Build- Blacktoe Clone with Belt Reduction

Postby u2fletch » Wed Jul 11, 2012 3:26 am

After spending many hours scouring the site I launched into my own build.

Image

Image

Bed utilizes a torsion box that conveniently consumed a bunch of leftover birch ply, tedious to build, but well worth it for stability.

Image

Initially I built the machine as a direct drive, using 1/2" drive shafts. Because of the NEMA shaft size I was limited to a 14 tooth drive sprocket, which gives a 3.5" per revolution. The Gecko drives are native 10 microstep so doing the math that equates to 571 steps per inch. A bit marginal it turned out when mated to the Pico USC. Had to do some serious PID loop tuning to get a set of values that would work, but Jon Elson from Pico did a great job providing support. Plan was to install a belt reduction unit later after machine was up and running.

Steppers are NEMA 34, 650 oz-in purchased from BuildYourCNC. Gecko203V drivers on each axis controlled by a Pico Systems USC. (http://pico-systems.com/univstep.html This controller simulates a servo system and handles all the step timing via hardware, which frees up the computer a bit. Running EMC2 under Ubuntu. Computer is an Intel D525 Atom with a 60Gb SSD.

Timing belts are 9mm Gt2, 3mm pitch, purchased from http://www.sdp-si.com. Timing pulleys were initially only available in 60tooth and 80 tooth, which would not give much of a reduction, but I later found they have a 120 tooth pulley which gives a nice 2:1 ratio using the 60tooth on the stepper output shaft.


Had to notch out the back of the gantry to make room for the belt drive. Minimum available belt size and size of 120tooth pulley prevented getting everything internal to existing gantry.

Image

Made a modification to the Z-axis mount to allow a modular type system. Occasionally, I will need to run a deep reach ball nose bit to shape guitar necks, so I needed a way to offset the entire spindle up or down since I only have 6" of Z axis travel available as is.
This will also allow me to mount a smaller spindle set 90 degrees to do guitar fret cutting, with a special fret cutting blade, kind of like a radial arm saw configuration. If you are curious what that will be used for you can check out this thread of an earlier build of mine.
http://www.talkbass.com/forum/f57/first-build-s-two-basses-once-337390/


Image

I modified a standard PC case and power supply to house the electronics. The Gecko drives require a heatsink, so I found 3 heatsink/fan combos designed for an AMD CPU for cheap. They are mounted to a slab of aluminum bolted to the chassis. Works great.


Image

Here's everything installed in the case. I used the standard Home Depot 4 conductor shielded security cable for the drives and terminated them with 4 pin Molex connectors designed for floppy drives near the motors. Super cheap from Fry's Electronics.

Image

As far as software goes I am running Rhino-3D for CAD duties, and sending that to CAMBAM to produce the G-Code.

I also installed a Logitech USB Playstation controller as a jog pendant. Buttons are programmed to select various jog rates, home, and touch off duties.

Next phase of the project is to install limit switches, probably just one on each axis to double as home/limit.

Another modification in the works is to install the belt drive on the Y axis as well. When I make the new Y axis plate I am going to switch the adjustable v-guide bearings to the bottom, and have fixed ones up top, like the x axis bearings. Will be much easier to do periodic tensioning that way.

Overall this has been an awesome project and I am inspired by all the pioneering work that has gone before by other members of this forum.

Jeff
u2fletch
 
Posts: 14
Joined: Wed Jun 06, 2012 2:37 pm
CNC Machine: BlackTooth

Re: New Build- Blacktoe Clone with Belt Reduction

Postby fastesthouse » Wed Jul 11, 2012 7:55 am

Nice work. I had a look at the Bass you made and it is a thing of beauty.
fastesthouse
 
Posts: 1
Joined: Tue Jul 10, 2012 4:39 pm
CNC Machine: None

Re: New Build- Blacktoe Clone with Belt Reduction

Postby ipmcc » Wed Jul 11, 2012 2:07 pm

Ah, someone else who used Molex HDD power connectors for stepper hookups. Had one pull out on you? I have, and I see you've added zip-tie strain relief. I'm migrating to screw-on CB mic connectors soon.

So I'll bite: why belt drive over direct gear reduction or chain reduction? Belt seems like it would have the most backlash... Is it just parts availability? Or was there something else?
ipmcc
 
Posts: 94
Joined: Fri May 04, 2012 11:17 am
CNC Machine: BlackToe v4 2x4

Re: New Build- Blacktoe Clone with Belt Reduction

Postby u2fletch » Wed Jul 11, 2012 5:12 pm

Hah! Yes, I had one pull out on me today. The Y axis snagged on the new outcropping created by the X axis belt drive extension. So now I also put zip ties through the connectors.

To answer your question about the belt reduction: The X and Y axis are chain drive just like the Blacktoe machine. Since I am using NEMA 34 with 1/2 inch output shaft, I was limited to a 14 tooth, instead of a 9 tooth sprocket that most people can use when using a 1/4" drive shaft. Because of this, I was limited in resolution. Jon from Pico Systems gave me the analogy of driving around in your car using only 5th gear.

The timing belt is only going from the motor to the existing chain drive shaft. Lots of people run complete timing belt systems, although most I don't think go over 4 feet or so because of concerns about belt stretch.

For this application I think it works very well. Belts do not have backlash issues like a hard gear set, which I found hard to locate as well. The performance difference is dramatic now that the stepper motor can "open it up" in RPM.

Jeff
u2fletch
 
Posts: 14
Joined: Wed Jun 06, 2012 2:37 pm
CNC Machine: BlackTooth

Re: New Build- Blacktoe Clone with Belt Reduction

Postby Awesomeness » Tue Jul 17, 2012 7:12 pm

Cool build!

It is worth noting, since people are asking questions, that belt reduction is not a "magic bullet". There is no assured performance increase, and you are just as likely to decrease performance. Choosing to use a reduction system can have benefits, as long as you are using drive electronics that are capable of good torque production at higher RPMs. For a lot of people in DIY, it seems, this ends up NOT being the case, because they are buying the cheapest components available, not higher end equipment that has power to spare.
If you don’t have time to do it right, when will you have time to do it over?
Image
User avatar
Awesomeness
 
Posts: 2878
Joined: Wed Feb 24, 2010 7:46 am
Location: Firestone, CO
CNC Machine: BlackFoot v4 4x8

Re: New Build- Blacktoe Clone with Belt Reduction

Postby u2fletch » Fri Jul 20, 2012 1:11 am

Excellent point Awesomeness. Adding extra complexity to a system can be more trouble than it is worth. In my case we discovered an issue within LinuxCNC and it's servo loop handling that did not deal well with low scale factors. Because the effective resolution was so small, the PID loop was having abnormally large errors attempting to track commanded position properly without performance robbing oscillation.

I have a set of Mesa 5125/7176 cards that I want to hook up and compare performance to see what issues are hardware and what are software. I know from experience that if the PID values are not right in a servo-feedback loop system, the machine will barely function, which has nothing to do with the drive mechanics. But, I also suspect that using the Mesa cards would be a lot more forgiving in setup and i probably would never have felt the need to resort to belt reduction to increase RPM's of the stepper motors.
u2fletch
 
Posts: 14
Joined: Wed Jun 06, 2012 2:37 pm
CNC Machine: BlackTooth

Re: New Build- Blacktoe Clone with Belt Reduction

Postby u2fletch » Sun Jul 22, 2012 12:56 pm

One last thought on the decision to implement belt reduction, or any form of gear reduction for that matter.

If you look at the torque versus RPM graph for stepper motors you will see that you need to get up to around 100 rpm to achieve maximum torque. Significant torque drop off generally does not occur until you are over 500 rpm or so. Now this is a huge generalization, all motors are different, and some will drop off faster than others. But for the rpm range that motors are running on our DIY chain drive machines the torque curve is pretty flat. In other words, if you double the rpm, you do not get half the torque, you might get 75% of the torque, but because of your new effective torque from mechanical gear reduction it is worth the change.

As an example with my 14 tooth .25 pitch sprocket (3.5" per rev) cutting at 140Ipm the motor is only turning at 40rpm. Doing a 500ipm rapid still only gets up just over 140. Doubling those values with a 2:1 reduction still gets nowhere near where the motor is going to struggle.

It is easy to see though why machines running threaded rod with, say 10tpi or more get maxed out. Try and do 70 ipm and you are up to 700rpm. Now we really are getting towards significant torque loss and the problem of lost steps.

There is no perfect solution, but half the fun of doing this is is trying to see what works to get the most bang for our buck performance wise.

Cheers!
u2fletch
 
Posts: 14
Joined: Wed Jun 06, 2012 2:37 pm
CNC Machine: BlackTooth


Return to DIY

Who is online

Users browsing this forum: larry104 and 3 guests

Who is online

In total there are 4 users online :: 1 registered, 0 hidden and 3 guests (based on users active over the past 5 minutes)
Most users ever online was 158 on Thu Mar 08, 2012 8:54 pm

Users browsing this forum: larry104 and 3 guests