Auto tool changer (XATC Clone for LinuxCNC)
- kiwi.lost.in.melb
- Offline
- New Member
-
Less
More
- Posts: 2
- Thank you received: 1
15 Jul 2018 03:14 #114196
by kiwi.lost.in.melb
Auto tool changer (XATC Clone for LinuxCNC) was created by kiwi.lost.in.melb
Hi All,
I am currently building a LinuxCNC port of Frank Herrmann's great work.
plus.google.com/collection/QX47gB
I havent got far but will post updates as I finish something useful.
My info is here if you would like to take a look...
plus.google.com/collection/8T5gcE
Cheers.
I am currently building a LinuxCNC port of Frank Herrmann's great work.
plus.google.com/collection/QX47gB
I havent got far but will post updates as I finish something useful.
My info is here if you would like to take a look...
plus.google.com/collection/8T5gcE
Cheers.
The following user(s) said Thank You: tommylight
Please Log in or Create an account to join the conversation.
21 Nov 2018 19:37 #121163
by jmr
Replied by jmr on topic Auto tool changer (XATC Clone for LinuxCNC)
Hi.
What can You point as a most serious issue with this kind of spindle?
Does it work well? How much hands free operations does it get without something going wrong? I believe it could gain much speed with the upgrade of rotating wrench - maybe a small version of impact wrench with long hex or spline nut with some kind of torque limit. Do You use any torque sensing?
I had some ideas for a similar changer. Now I'm thinking about adding a tensometer (force->resistance converter) to worm/worm gear assembly
There is also one other modification which is really an add on spindle (with additional pair of bearings) with ball locked pull stud and propably air actuator. The topic is dead for a few years now. That is not too easy to make considering the speed of application, however that ATC should be very fast and reliable. Maybe it could be just a bit harder to make the spindle itself.
The concept with mechanized wrench has some drawbacks - possibly the wear on thread being the reliability factor, or the wear on the flats of shaft as the "fork" can be later adjusted or changed. Other is a problem of ER taper not being self releasing.
I was thinking more about having tools in toolchanger inside their ER collets with nut attached (yes possibly one nut for one tool). This is bad too, as not all tools block in the collet, but it has some advantages like not having to use one diameter for the tooling.
It would too require a tightening/loosening station and tool carousel/parking lot.
Considering the longevity of this spindle... If you needed the PCB done you'll change tools manually anyway causing similar wear as the ATC so the overall effect is possibly a lot of time saved so that one can clean the floor if things are going reliably.
What can You point as a most serious issue with this kind of spindle?
Does it work well? How much hands free operations does it get without something going wrong? I believe it could gain much speed with the upgrade of rotating wrench - maybe a small version of impact wrench with long hex or spline nut with some kind of torque limit. Do You use any torque sensing?
I had some ideas for a similar changer. Now I'm thinking about adding a tensometer (force->resistance converter) to worm/worm gear assembly
There is also one other modification which is really an add on spindle (with additional pair of bearings) with ball locked pull stud and propably air actuator. The topic is dead for a few years now. That is not too easy to make considering the speed of application, however that ATC should be very fast and reliable. Maybe it could be just a bit harder to make the spindle itself.
The concept with mechanized wrench has some drawbacks - possibly the wear on thread being the reliability factor, or the wear on the flats of shaft as the "fork" can be later adjusted or changed. Other is a problem of ER taper not being self releasing.
I was thinking more about having tools in toolchanger inside their ER collets with nut attached (yes possibly one nut for one tool). This is bad too, as not all tools block in the collet, but it has some advantages like not having to use one diameter for the tooling.
It would too require a tightening/loosening station and tool carousel/parking lot.
Considering the longevity of this spindle... If you needed the PCB done you'll change tools manually anyway causing similar wear as the ATC so the overall effect is possibly a lot of time saved so that one can clean the floor if things are going reliably.
Please Log in or Create an account to join the conversation.
- kiwi.lost.in.melb
- Offline
- New Member
-
Less
More
- Posts: 2
- Thank you received: 1
21 Nov 2018 20:53 #121177
by kiwi.lost.in.melb
Replied by kiwi.lost.in.melb on topic Auto tool changer (XATC Clone for LinuxCNC)
Hi Mate,
There are always drawbacks with any project.
Ok, so to answer some of your questions....
>>Does it work well? How much hands free operations does it get without something going wrong?
The major drawback is the tool sometimes doesnt release. I added a test to check for this - I.e. try to remove tool and then do a tool height check - it should fail if there is no tool. Using superglue on the plastic ring helped to reduce this a lot. Adding a test made sure that there were no breakages if it did fail.
I have had 1 failure since using this and had to manually remove the tool. It autoloads the next tool fine.
>> RE Torque
I looked for a long time for a suitable mechanical torque wrench. I could not find anything.
So the way I do it is:
- tighten hand tight by spinning spindle at low speed and the spindle will bind up - low RPM and low torque for the spindle so no damage. THEN
- move wrench a set number of degrees with the spindle locked.
This method is reasonably reliable for a repeatable torque. I measured it and it was +- about 5%.
Remember that I use this for PCB's not steel or ali - I dont need exact or even a high torque for this to work. I dont see why it wouldnt work for high torque - I used plastic for the lock. High torque will need a stronger material.
>> Torque sensing
I had a few ideas for torque sensing - measuring the back EMF of a stepper used for tightening is one way that could provide a reliable torque measurement. Do a search on coolstep for a way this could be done... www.trinamic.com/technology/adv-technologies/coolstep/
>> The concept with mechanized wrench has some drawbacks - possibly the wear on thread being the reliability factor, or the wear on the flats of shaft as the "fork" can be later adjusted or changed.
Yup sure does. Flats arent really a problem as it is a big enough nut to handle wear for a long while. The socket handles flats no problem. anyway - nuts are cheap.
>> I was thinking more about having tools in toolchanger inside their ER collets with nut attached
Hmm. I think you will have a LOT of problems with this approach. Removing the nut entirely means that you need to line it up exactly when putting it on - otherwise cross threading on the spindle. I never remove the nut so this is never a problem.
Remember nuts and collets are cheap. A spindle is not.
>> Stepper driven wrench
I have found a few suitable wrenches on alibaba. I just dont have the time to implement this change and wont until I need the bed space. I mill PCBs (automated change) and acrylic in manual mode as they normally only use 1 tool.
>> Collets with additional bearings
There are ER11 collets which have bearings against the nut. I bought one and tried it - didnt work as the solution NEEDS the friction on the collet against the nut for the removal process to work - I.e. the collet needs to move with the spindle when doing the release of the tool. I.e. spin the tool IN the collet. Sure there is some collet wear on the inside. They are cheap.
Good luck! Feel free to PM me for my email if you need some help.
There are always drawbacks with any project.
Ok, so to answer some of your questions....
>>Does it work well? How much hands free operations does it get without something going wrong?
The major drawback is the tool sometimes doesnt release. I added a test to check for this - I.e. try to remove tool and then do a tool height check - it should fail if there is no tool. Using superglue on the plastic ring helped to reduce this a lot. Adding a test made sure that there were no breakages if it did fail.
I have had 1 failure since using this and had to manually remove the tool. It autoloads the next tool fine.
>> RE Torque
I looked for a long time for a suitable mechanical torque wrench. I could not find anything.
So the way I do it is:
- tighten hand tight by spinning spindle at low speed and the spindle will bind up - low RPM and low torque for the spindle so no damage. THEN
- move wrench a set number of degrees with the spindle locked.
This method is reasonably reliable for a repeatable torque. I measured it and it was +- about 5%.
Remember that I use this for PCB's not steel or ali - I dont need exact or even a high torque for this to work. I dont see why it wouldnt work for high torque - I used plastic for the lock. High torque will need a stronger material.
>> Torque sensing
I had a few ideas for torque sensing - measuring the back EMF of a stepper used for tightening is one way that could provide a reliable torque measurement. Do a search on coolstep for a way this could be done... www.trinamic.com/technology/adv-technologies/coolstep/
>> The concept with mechanized wrench has some drawbacks - possibly the wear on thread being the reliability factor, or the wear on the flats of shaft as the "fork" can be later adjusted or changed.
Yup sure does. Flats arent really a problem as it is a big enough nut to handle wear for a long while. The socket handles flats no problem. anyway - nuts are cheap.
>> I was thinking more about having tools in toolchanger inside their ER collets with nut attached
Hmm. I think you will have a LOT of problems with this approach. Removing the nut entirely means that you need to line it up exactly when putting it on - otherwise cross threading on the spindle. I never remove the nut so this is never a problem.
Remember nuts and collets are cheap. A spindle is not.
>> Stepper driven wrench
I have found a few suitable wrenches on alibaba. I just dont have the time to implement this change and wont until I need the bed space. I mill PCBs (automated change) and acrylic in manual mode as they normally only use 1 tool.
>> Collets with additional bearings
There are ER11 collets which have bearings against the nut. I bought one and tried it - didnt work as the solution NEEDS the friction on the collet against the nut for the removal process to work - I.e. the collet needs to move with the spindle when doing the release of the tool. I.e. spin the tool IN the collet. Sure there is some collet wear on the inside. They are cheap.
Good luck! Feel free to PM me for my email if you need some help.
Please Log in or Create an account to join the conversation.
22 Nov 2018 14:14 #121220
by andypugh
Replied by andypugh on topic Auto tool changer (XATC Clone for LinuxCNC)
I was impressed by this idea when I saw it a few months ago.
(For those who haven't figured it out yet, the idea is to use the relatively powerful X/Y axes to do the final tightening of the nut by making a circular move around a pivoting spanner.)
I have just had an idea, though not really thought it through in detail...
Ring magnets are readily available, you could put a collar round the tools with a magnet in it to slot into a finger-style magazine.
The magnet would keep the tool stuck to the collet to allow it to be transported to a standalone tightening station, and should also give repeatable tool insertion distance.
(For those who haven't figured it out yet, the idea is to use the relatively powerful X/Y axes to do the final tightening of the nut by making a circular move around a pivoting spanner.)
I have just had an idea, though not really thought it through in detail...
Ring magnets are readily available, you could put a collar round the tools with a magnet in it to slot into a finger-style magazine.
The magnet would keep the tool stuck to the collet to allow it to be transported to a standalone tightening station, and should also give repeatable tool insertion distance.
Please Log in or Create an account to join the conversation.
Time to create page: 0.149 seconds