i like that you took the time to make the big easy thumbwheels instead of having a nut or a setscrew to fiddle with .. little extra time at the beginning to make life easier later on is a good thing
just one thing I don't think that calipers have Absolute Encoders because the same model was teardown by AvE and the moment battery is removed it "forget" and needs re-zeroing best way to test is just remove the battery and put it back in (do wait for caps discharging ) if it remembers then its Absolute
Awesome. I dig the encoder setup. Question, is there something you are going to use the motor outputs on the back of the motors for? I remember in a previous video you had mentioned you purchased those specifically for motion measurement. Is the current encoder method different out of necessity because you have a gearbox?
I purchased them so I could have options because I wasn't quite sure what I was going to use yet. I could have attached an encoder to the back shaft but I would need to use incremental encoders and limit switches. Due to the gearbox, the back shaft will make more than one full revolution but the output shaft will not even move half of one revolution. The absolute encoder will allow me to avoid limit switches and never have to home the machine. It will also help with back lash that is only seen on the output shaft.
The hardware used in the is FHS-256-12, SOS-256-8, and F-256-2. I used more hardware in this video: ruclips.net/video/RyHtSL4euIk/видео.html - Anthony :)
Where did you bought the eyebolts that are attached to the rods ? is your delta robot for comercial applications ? is your robot open source ? where can I find more information about your bot?
Good question! I would need to use incremental encoders if I put it on the back because I'm using a 5:1 gearbox. The back shaft will make more than a full revolution but the output shaft will not even move half of one revolution. Using it on the back shaft would force me to also use limit switches to zero the machine. The absolute encoder will allow me to never have to home the machine.
The assembled sensors on PCBs were only $10. A potentiometer would be ok for small projects that do not move frequently and don't need the accuracy like this. There are two main problems with using them for my robot. Most potentiometers are not as linear as you would think so the positional accuracy would not be that great. The other issue is that they would not last long on a robot like this that is rapidly moving all the time. The wiper in the potentiometer rubs against a resistive ring which eventually scrapes off material and changes the resistance. The sensors I'm using are non contact so will not be subjected to wear and tear.
Holy shit! You and I need to talk, this is the first time I'm stumbling across your videos looking for a deeper understanding of quadrature encoders and already I need to know what you're doing, why, and what you're using. I have a ton of questions. I have built a Delta/Kossel 3D printer and I want to make a 6 DOF motion platform and, from my understanding, it is not too dissimilar from these Delta robots. The main difference is, of course, we aim for precision by keeping the "effector" (at least that's what we call it with the 3D printers) parallel to the print surface or work table. In these systems, you try to eliminate "effector tilt" as much as possible. I'm a motion system, like a full motion flight simulator, you want that effector tilt...and it's upside down....And much heavier duty. What is that? A pick and place machine? It's gotta be, why else would you be using machine vision and that geometry machine? You and I should talk because I am extremely interested in the motor control and I don't fully understand encoders, more the programming of them than anything else, I have access to a lot of motors and a fully equipped machine shop....also...I am proficient in Autodesk Inventor HSM and am a CNC Machinist and have access to a bunch of hardware, electrical, and software engineers...but they don't really do robotics, which is so close to being crazy helpful 😆 liking the Mitutoyo and that thing looks awesome, you use Misumi for the extrusion? Those are some beefy motors, not cheap. Whatever you buildin.....it certainly looks like you know what you're doing, damn.
I am not 100% sure, but I think the newer Mitutoyos still use incremental encoders. It is just that they never turn off. That is how it knows the position when you "turn it on". One way to test is to remove the battery and see if the caliper still knows the position.
I am 100% sure that they use absolute encoders as it is in the patent and their website www.mitutoyo.com/wp-content/uploads/2014/02/2129-AOS-Calipers.pdf The calibration value is stored in memory and is wiped when the battery is removed which is why there is a need to re-zero after the battery has been replaced.
Some addons.. quadrature encoders are called that, due to the pulse pattern they send out, quadrature pulses The "absolute" caliper, is only turning the LCD off.. not the encoder interface
That's my understanding as well. It's still powered on, hence why it keeps the count. Removing the battery will reset it, so the actual sensing track isn't absolute. They are cheating a bit there.
I have heard that Mitutoyo does that from other people but not ever seen any confirmation. Do you have confimation of that being the case? Patent US5886519A states, "A pair of push button switches 134 and 136 are also mounted in the cover 139. The switch 134 turns on and off a signal processing IC and display electronic circuit 166 of the slider assembly 120."
That is not confirmation as they are storing the initial calibration value in memory. Removing the battery wipes the memory. I would like to see confirmation from Mitutoyo, as I have not seen it mentioned anywhere in patents or directly from the company that they read values while in the "off" state. In the new model in the video, Mitutoyo uses an electromagnetic induction absolute encoder.
I dont know about that brand, but I have seen multiple teardowns of simular, and one looked into the capacitive strip, and it was just a repeating pattern :) But hey, thats not what your video is about, the different of absolute and incremental was on point :)
Here are links to some of the equipment shown in the video:
ESP32-EVB - amzn.to/2LYepZP
Object Tracking Pixy2 Camera - amzn.to/2tgO5Cy
Braided Cable Sleeve - amzn.to/2LedCnb
Stainless Steel Cable Ties - amzn.to/2LevZsb
Strain Relief Cable Glands - amzn.to/2LdnuO8
Silicone Fusing Tape - amzn.to/2J5RPO8
Locking Casters - amzn.to/2kJ75WB
Aluminum Extrusion - amzn.to/2Lh14M7
Cable Clips - amzn.to/2xColG8
M5 T Slot Nuts - amzn.to/2HdnNWP
M5 Screws - amzn.to/2J4OSBs
LED Ring Light - amzn.to/2sqeIFs
Carbon Fiber Tubes - amzn.to/2J5xVqz
Suction Cup Nozzle - amzn.to/2LQFrDw
MA730 Magnetic Encoder - www.monolithicpower.com/en/ma730.html
i like that you took the time to make the big easy thumbwheels instead of having a nut or a setscrew to fiddle with .. little extra time at the beginning to make life easier later on is a good thing
just one thing I don't think that calipers have Absolute Encoders because the same model was teardown by AvE and the moment battery is removed it "forget" and needs re-zeroing best way to test is just remove the battery and put it back in (do wait for caps discharging ) if it remembers then its Absolute
Awesome. I dig the encoder setup. Question, is there something you are going to use the motor outputs on the back of the motors for? I remember in a previous video you had mentioned you purchased those specifically for motion measurement. Is the current encoder method different out of necessity because you have a gearbox?
I purchased them so I could have options because I wasn't quite sure what I was going to use yet. I could have attached an encoder to the back shaft but I would need to use incremental encoders and limit switches. Due to the gearbox, the back shaft will make more than one full revolution but the output shaft will not even move half of one revolution. The absolute encoder will allow me to avoid limit switches and never have to home the machine. It will also help with back lash that is only seen on the output shaft.
Hey Greg, looking great! Could you talk a bit about the press in hardware you are using?
The hardware used in the is FHS-256-12, SOS-256-8, and F-256-2. I used more hardware in this video: ruclips.net/video/RyHtSL4euIk/видео.html - Anthony :)
Thanks Anthony, apologies for getting your name wrong!
Pete Brubaker no worries Pete!
Awesome video ! Impressive workshop and equipment !
Thank you!
Where did you bought the eyebolts that are attached to the rods ? is your delta robot for comercial applications ? is your robot open source ? where can I find more information about your bot?
Why did you choose to use these kind of encoders and not the regular encoders that are attached to the back of the motors?
Good question! I would need to use incremental encoders if I put it on the back because I'm using a 5:1 gearbox. The back shaft will make more than a full revolution but the output shaft will not even move half of one revolution. Using it on the back shaft would force me to also use limit switches to zero the machine. The absolute encoder will allow me to never have to home the machine.
Proto G crystal clear! Where did you buy the encoders you're using?
Directly from MSP here: www.monolithicpower.com/en/tbma730-q-rd-01a.html
this might be a dumb question, but couldn't you just have used a potentiometer to see at what position it is? seems like a way cheaper option.
The assembled sensors on PCBs were only $10. A potentiometer would be ok for small projects that do not move frequently and don't need the accuracy like this. There are two main problems with using them for my robot. Most potentiometers are not as linear as you would think so the positional accuracy would not be that great. The other issue is that they would not last long on a robot like this that is rapidly moving all the time. The wiper in the potentiometer rubs against a resistive ring which eventually scrapes off material and changes the resistance. The sensors I'm using are non contact so will not be subjected to wear and tear.
That makes sense, Thanks a lot for the response!
What an excellent video 📹
Thank you! 🙏
Is there a video or section where you describe how you join the ball joint linkage to the stepper motor linkage?
Great stuff by the way. Keep up the awesome work!
Holy shit! You and I need to talk, this is the first time I'm stumbling across your videos looking for a deeper understanding of quadrature encoders and already I need to know what you're doing, why, and what you're using. I have a ton of questions. I have built a Delta/Kossel 3D printer and I want to make a 6 DOF motion platform and, from my understanding, it is not too dissimilar from these Delta robots. The main difference is, of course, we aim for precision by keeping the "effector" (at least that's what we call it with the 3D printers) parallel to the print surface or work table. In these systems, you try to eliminate "effector tilt" as much as possible. I'm a motion system, like a full motion flight simulator, you want that effector tilt...and it's upside down....And much heavier duty.
What is that? A pick and place machine? It's gotta be, why else would you be using machine vision and that geometry machine?
You and I should talk because I am extremely interested in the motor control and I don't fully understand encoders, more the programming of them than anything else, I have access to a lot of motors and a fully equipped machine shop....also...I am proficient in Autodesk Inventor HSM and am a CNC Machinist and have access to a bunch of hardware, electrical, and software engineers...but they don't really do robotics, which is so close to being crazy helpful 😆 liking the Mitutoyo and that thing looks awesome, you use Misumi for the extrusion? Those are some beefy motors, not cheap. Whatever you buildin.....it certainly looks like you know what you're doing, damn.
Which encoders are you using??
MA730 Magnetic Encoder - www.monolithicpower.com/en/ma730.html
Very nice work. I hear your intelligence.
I think it's fair to say that these motors (and you even geared them down, geez!) have a ton of power. Do you think it could bowl like the pros?
I am not 100% sure, but I think the newer Mitutoyos still use incremental encoders. It is just that they never turn off. That is how it knows the position when you "turn it on". One way to test is to remove the battery and see if the caliper still knows the position.
I am 100% sure that they use absolute encoders as it is in the patent and their website www.mitutoyo.com/wp-content/uploads/2014/02/2129-AOS-Calipers.pdf
The calibration value is stored in memory and is wiped when the battery is removed which is why there is a need to re-zero after the battery has been replaced.
good to know. Great stuff by the way. I like your content. Keep it up :)
Why do you have 3 separate power supplies ?
Because they are 60V 6A power supplies. 3 of those are cheaper, safer, and more readily available than one 60V 18A power supply.
Some addons.. quadrature encoders are called that, due to the pulse pattern they send out, quadrature pulses
The "absolute" caliper, is only turning the LCD off.. not the encoder interface
That's my understanding as well. It's still powered on, hence why it keeps the count. Removing the battery will reset it, so the actual sensing track isn't absolute. They are cheating a bit there.
I have heard that Mitutoyo does that from other people but not ever seen any confirmation. Do you have confimation of that being the case? Patent US5886519A states, "A pair of push button switches 134 and 136 are also mounted in the cover 139. The switch 134 turns on and off a signal processing IC and display electronic circuit 166 of the slider assembly 120."
Try popping out the battery for a minute and see what happens.
That is not confirmation as they are storing the initial calibration value in memory. Removing the battery wipes the memory. I would like to see confirmation from Mitutoyo, as I have not seen it mentioned anywhere in patents or directly from the company that they read values while in the "off" state. In the new model in the video, Mitutoyo uses an electromagnetic induction absolute encoder.
I dont know about that brand, but I have seen multiple teardowns of simular, and one looked into the capacitive strip, and it was just a repeating pattern :)
But hey, thats not what your video is about, the different of absolute and incremental was on point :)
Great video, and thankyou
More...thx
Mmmm single board computers are getting stupid sexy!