Daniel's Work Thread: Toward a Hackerspace Scanner
Moderator: peterZ
Re: Daniel's Work Thread: Toward a Hackerspace Scanner
THAT WAS THE MOST EXCITING THING IVE SEEN
- rob
- Posts: 773
- Joined: 03 Jun 2009, 13:50
- E-book readers owned: iRex iLiad, Kindle 2
- Number of books owned: 4000
- Country: United States
- Location: Maryland, United States
- Contact:
Re: Daniel's Work Thread: Toward a Hackerspace Scanner
Hall effect sensors are like black magic to me. It works, but I don't trust it! What made you decide on magnetic sensors instead of a switch or some other electrical contact? Or are they just something you had lying around or found at the surplus store (totally valid reason).
The Singularity is Near. ~ http://halfbakedmaker.org ~ Follow me as I build the world's first all-mechanical steam-powered computer.
- daniel_reetz
- Posts: 2812
- Joined: 03 Jun 2009, 13:56
- E-book readers owned: Used to have a PRS-500
- Number of books owned: 600
- Country: United States
- Contact:
Re: Daniel's Work Thread: Toward a Hackerspace Scanner
Hey Rob,
I paid a lot of money for these stupid sensors! Around $15 each with shipping. I read most everything I could find about the design of home switches. Essentially you can use
1. A switch
2. Something optical
3. Hall effect stuff
Switches are easy to break and jam up with stuff, and they don't have high mechanical precision. Optical stuff tends to get sawdusty, and is also not particularly accurate, at least with the little guys you get out of printers and stuff. Hall effect is the most durable/sealed, and as long as the magnet doesn't change over time it will come back to EXACTLY the same spot (the sensors are designed really well, in that respect). So, hall effect.
I paid a lot of money for these stupid sensors! Around $15 each with shipping. I read most everything I could find about the design of home switches. Essentially you can use
1. A switch
2. Something optical
3. Hall effect stuff
Switches are easy to break and jam up with stuff, and they don't have high mechanical precision. Optical stuff tends to get sawdusty, and is also not particularly accurate, at least with the little guys you get out of printers and stuff. Hall effect is the most durable/sealed, and as long as the magnet doesn't change over time it will come back to EXACTLY the same spot (the sensors are designed really well, in that respect). So, hall effect.
- rob
- Posts: 773
- Joined: 03 Jun 2009, 13:50
- E-book readers owned: iRex iLiad, Kindle 2
- Number of books owned: 4000
- Country: United States
- Location: Maryland, United States
- Contact:
Re: Daniel's Work Thread: Toward a Hackerspace Scanner
Okay, okay! You've convinced me! I just imagine some magnetic field whomping through the room and triggering the sensor. Then again, I suppose the sensor requires such a high magnetic field that nothing except an EMP would set it off accidentally...
The Singularity is Near. ~ http://halfbakedmaker.org ~ Follow me as I build the world's first all-mechanical steam-powered computer.
Re: Daniel's Work Thread: Toward a Hackerspace Scanner
Maybe Daniel's working on a DIY-MRI machine? If only liquid helium weren't such a chore to keep a lid on ...rob wrote:Okay, okay! You've convinced me! I just imagine some magnetic field whomping through the room and triggering the sensor. Then again, I suppose the sensor requires such a high magnetic field that nothing except an EMP would set it off accidentally...
- daniel_reetz
- Posts: 2812
- Joined: 03 Jun 2009, 13:56
- E-book readers owned: Used to have a PRS-500
- Number of books owned: 600
- Country: United States
- Contact:
Re: Daniel's Work Thread: Toward a Hackerspace Scanner
heheheehGado wrote:
Maybe Daniel's working on a DIY-MRI machine? If only liquid helium weren't such a chore to keep a lid on ...
Here's what I did over Friday and Saturday:
- daniel_reetz
- Posts: 2812
- Joined: 03 Jun 2009, 13:56
- E-book readers owned: Used to have a PRS-500
- Number of books owned: 600
- Country: United States
- Contact:
Re: Daniel's Work Thread: Toward a Hackerspace Scanner
Well, after a full day of programming the cuts, I finally got back to cutting wood tonight. However I still have some bugs to work out.
The good news:
Using the feed and speed settings I found with G-Wizard Calculator, I was able to cut through the plywood in just three passes! So that's really great, it means cutting can be fast. Here's a picture of the edge finish with three passes, .8" stickout, 220IPM, 12000RPM (a bit slower than the actual calcs):
The bad news:
I'm still having angular errors that depend on which way the machine is traveling around corners. Here's a pic. There are about a million things that influence this behavior in Mach3, and I need to systematically chase them down. I thought I'd done that, though - my acceleration is set extremely fast, my speed is about half what the machine is capable of, and the G-code is fairly clean. I guess I need to do EVEN MORE READING. When I cut the first scanner, I was able to cheat on stuff like this - I made cuts with extra lines or cut very slow or whatever because I'd just built the machine and I only needed one. However now that I am doing bulk work, I need to nail this problem to the ground.
The good news:
Using the feed and speed settings I found with G-Wizard Calculator, I was able to cut through the plywood in just three passes! So that's really great, it means cutting can be fast. Here's a picture of the edge finish with three passes, .8" stickout, 220IPM, 12000RPM (a bit slower than the actual calcs):
The bad news:
I'm still having angular errors that depend on which way the machine is traveling around corners. Here's a pic. There are about a million things that influence this behavior in Mach3, and I need to systematically chase them down. I thought I'd done that, though - my acceleration is set extremely fast, my speed is about half what the machine is capable of, and the G-code is fairly clean. I guess I need to do EVEN MORE READING. When I cut the first scanner, I was able to cheat on stuff like this - I made cuts with extra lines or cut very slow or whatever because I'd just built the machine and I only needed one. However now that I am doing bulk work, I need to nail this problem to the ground.
- daniel_reetz
- Posts: 2812
- Joined: 03 Jun 2009, 13:56
- E-book readers owned: Used to have a PRS-500
- Number of books owned: 600
- Country: United States
- Contact:
Notes to self
Alright, notes to myself for tonight's machine research:
Here's Art (who programmed the original Mach3) talking about how to use CV and the problems with cutting square corners.
Here's a PDF guide to CV settings.
Here's a veteran description of the problem citing total tool velocity and acceleration as contributing factors.
Another post identifying CV settings which contribute to non-square corners.
CV Dist Tolerance Settings
All in all, I can see a few trends here:
1. Using true arcs in the design is not going to go well. I've already edited the rounded pieces to be polygonal.
2. It's possible to enable Exact Stop mode via GCode. A best-case version of this code would only disable CV when it needs to.
3. The cuts are just going to have to be slower, or the artwork more elaborate, I am not sure which.
4. From cutting last night, my tabs were insufficient (again!!!!) gonna need a sander or a vac table at some point.
Here's Art (who programmed the original Mach3) talking about how to use CV and the problems with cutting square corners.
Here's a PDF guide to CV settings.
Here's a veteran description of the problem citing total tool velocity and acceleration as contributing factors.
Another post identifying CV settings which contribute to non-square corners.
CV Dist Tolerance Settings
All in all, I can see a few trends here:
1. Using true arcs in the design is not going to go well. I've already edited the rounded pieces to be polygonal.
2. It's possible to enable Exact Stop mode via GCode. A best-case version of this code would only disable CV when it needs to.
3. The cuts are just going to have to be slower, or the artwork more elaborate, I am not sure which.
4. From cutting last night, my tabs were insufficient (again!!!!) gonna need a sander or a vac table at some point.
- daniel_reetz
- Posts: 2812
- Joined: 03 Jun 2009, 13:56
- E-book readers owned: Used to have a PRS-500
- Number of books owned: 600
- Country: United States
- Contact:
Re: Daniel's Work Thread: Toward a Hackerspace Scanner
looks like it will be possible to do the fillets in Vcarve instead of in the artwork. I may go down this path.
- Attachments
-
- Fillets.JPG (48.23 KiB) Viewed 8488 times
- rob
- Posts: 773
- Joined: 03 Jun 2009, 13:50
- E-book readers owned: iRex iLiad, Kindle 2
- Number of books owned: 4000
- Country: United States
- Location: Maryland, United States
- Contact:
Re: Daniel's Work Thread: Toward a Hackerspace Scanner
Yes. The fillets in VCarve are awesome... as long as you're not planning to change the artwork afterwards, because then you have to go in and add the fillets again. I've had the occasional problem with VCarve just not wanting to put a fillet in. But it sure is a time-saver!
BTW, often you don't need to attach the clip to the router, since the bit could be grounded. You can check this by just touching the touchplate to the bit and watching your switch lights. If the light goes on when you touch the bit with the touchplate, you don't need the clip, which can be an enormous pain
BTW, often you don't need to attach the clip to the router, since the bit could be grounded. You can check this by just touching the touchplate to the bit and watching your switch lights. If the light goes on when you touch the bit with the touchplate, you don't need the clip, which can be an enormous pain
The Singularity is Near. ~ http://halfbakedmaker.org ~ Follow me as I build the world's first all-mechanical steam-powered computer.