nicengineering

Try to look here: https://www.thingiverse.com/thing:1398215

Hi,That is something I could have described better I see !Looking in the documentation, theta3 + theta2 is the angle of the secondary gear while theta2 is angle of the primary gear and theta1 is the angle of the rotating gear. Then you have to apply gearing ratio to end up at the motor angle :) You can look in our code from line 770 to see how we did it:https://github.com/uStepper/robotControl/blob/mast...Best regards,Mogens

No problem - I just updated the Kinematic documentation on Git with a little more information so that it's hopefully a bit easier to follow :)

Hi! They are 228mm belts - pretty widely used for 3D printers, so you should be able to get hold of some pretty easily :)

Hi!Yes I moved to belt drives to reduce backlash and make movement more smooth and silent.You could argue that the mechanical stability of the rest of the system contributes more to repeatability error than the gears did. It is a chain of inaccuracies that impacts the repeatability and the drives do have some impact of course.Repeatability is comparable on rev 3 and 4 when the arm is fresh assembled and gears are adjusted to give minimum backlash. However over time the gears will wear, but can be readjusted of course. So, in reality there is no noticeable difference since the rest of the structure contributes more to the error (at least if loaded with weight). But on noise and smoothness there is a noticeable difference :)I have looked at counterweight systems but due to simplicity and th…

Hi!Yes I moved to belt drives to reduce backlash and make movement more smooth and silent.You could argue that the mechanical stability of the rest of the system contributes more to repeatability error than the gears did. It is a chain of inaccuracies that impacts the repeatability and the drives do have some impact of course.Repeatability is comparable on rev 3 and 4 when the arm is fresh assembled and gears are adjusted to give minimum backlash. However over time the gears will wear, but can be readjusted of course. So, in reality there is no noticeable difference since the rest of the structure contributes more to the error (at least if loaded with weight). But on noise and smoothness there is a noticeable difference :)I have looked at counterweight systems but due to simplicity and the limited impact I believe it would have in the applications I designed this arm for (even though I know I cant imagine all applications) I did not put much thought into adding it.But, you are free to adjust and adapt ! I have shared my source files so that people have an easy way of adjusting the design instead of just posting STL files :)Best regards,Mogens

Looks nice !Maybe a combination of this with my take on the same chicken challenge could take it to the next level :) You never had issues with the wire getting tangled up or snapping?I like the limit switch types you use - non-contact switches are the best solution in the long run :)See my take on it here:https://www.instructables.com/Chicken-Gate/Best regards,Mogens

Thanks :)

Thanks ! The controllers I use can detect if the limits are hit - so a homing procedure is done without extra switches 😉 Right now we are writing a complete library for it with that feature included...

Have a look at the latest iteration of the robot where I have also done kinematics for it: https://www.instructables.com/UStepper-Robot-Arm-4/

Thank you for the kind words. Yes it should be possible to build it without the aid of another CNC or 3D printer - but I know that in this particular design I depend on 3D printed parts though. I am glad that you find it useful!

Hi Robert,With stepper motors you can achieve a positioning of ~0.12 deg with regular microstep drivers. The repeatability is however dependent on that you do not end up loosing steps by putting to much load on the motors. With the uStepper boards, missed steps are corrected for so the repeatability is ~0.12 deg. As I remember it I achieved less than 1 mm error on positions. The main reason is mechanic tolerances.Best regards,Mogens

If you open the Fusion 360 link under step 2 (http://a360.co/2m8K6Dk) You should be able to download the assembly in multiple formats.

Sorry for the late reply... You can have a look at the code on Git, and take what you need to get it working. I simply do not have the time to do custom coding for people :)

It's not well suited for PCB routing - it requires more precision.

Based on the comments and likes, I don't think the majority is with you on this :) But with a little imagination you can do without 3D printing. I did not use a cnc to cut out any of the wood here and the first versions I built was without the use of a 3D printer :)

Sorry for the late reply... I am not sure what you mean. You want to cut out the machine using a laser?

Fusion 360 for drawing the machine and GRBL/gcodesender for running the machine.

Hi,I dont know since i havent tried that. But it can do 16mm MDF wood (in multiple passes) and 6mm plywood.Best regards,Mogens

Hi! I guess the problem will be that the y-axis rods will tend to bend a bit when they become to long (caused by the gantry weight). I don't have a number on how big you can go, but to overcome the previously mentioned problem, you could go up in tube dimension.

It is accurate, but yes, spacers should have been washers...

Have a look here:

It should be able to do that. You will have to do it in multiple passes though. But that is required on most small cnc routers. Best regards, Mogens

Hi! Thanks for the nice words. Two of the controllers are slaves and one is the master. Commenting in either of the two in the loop will configure the board as one of these. It just calls the appropriate function. Scroll up in the code, and you will see :) Best regards, Mogens

Hi!I use regular M8 threaded rod. With the long nuts, they barely have any backlash.Best regards,Mogens

Hi. I have not had that issue. The bearings sits nicely on the u-bolts. It is of course important to align them properly before tightening the assembly. I guess it could be caused by the u-bolt not being totally perpendicular to the tubing. I could imagine that if one of the u-bolts was misaligned, it could result in what you describe.

Hi!I can only offer you the help given in the instructable. Everything you need to know, should be there. I cannot offer you personal support in the entire process of building and configuring the machine.Best regards,Mogens

Good to hear! I have paid for Lazycam some years ago, so that is what I use for simple tasks. There is also something like Camotics and a bunch of other open source CAM solutions that I have not yet tried to work with. At the moment I don't have a CNC (I sold it a year ago or so), so I am not up to date on the newest open source CAM software out there.Best regards,Mogens

Thanks. I guess 10mm would be a bit much. But maybe it could be achieved with a good spindle and the right settings.

Hi, I do not have a diagram. But just Google "grbl wiring ". There are many examples to find, and they are all similar.

If you wan't to buy it, kits are available at www.ustepper.com If you wan't to build it yourself, you just have to get the parts listed in the instructable. The highest cost is the stepper and drivers.

Thank you for the nice words. I can see the problem with the imperical vs metric units in the US. The grooves I just made with a table saw, and yes, the aluminum angle is held in the groove by the pressure from the u-bolt and bearing. The pipe wall thickness is around 2.65 mm. The gantry does weigh a bit, so I don't think you should go much further down in tube dimension. You might risk that it will bend a bit on the middle if the dimension gets to small.Best regards,Mogens

Thnak you for the nice words. I can see the problem with the imperical vs metric units in the US. The grooves I just made with a table saw, and yes, the aluminum angle is held in the groove by the pressure from the u-bolt and bearing. The pipe wall thickness is around 2.65 mm. The gantry does weigh a bit, so I don't think you should go much further down in tube dimension. You might risk that it will bend a bit on the middle if the dimension gets to small.Best regards,Mogens

Thank you for the nice words!

Thanks!

Hi!Thanks for the kind words. I have just checked the link again, and it is still pointing to a nut cover (item number 25 on my list).

Thanks for the kind words. Your machine also looks nice - a bit bigger though :)

Hi,Yes, you could use any other stepper driver solution out there, be it GRBL shields, 3D printer controllers etc. Of course the sketch I wrote was done for the uStepper, but the principles would be the same as long as you have feedback. If you don't have feedback, then you will have to run it open loop (like a 3D printer og GRBL machine does).

Thank you for the kind words!

Hi Thanks for the kind words! No the uStepper boards actually copes very well with the heat generated by the driver chip. This is due to the design of the board, which has large ground planes (also internal) that takes the heat. If loaded excessive over longer periods, one might have to add some cooling - but I haven't had any issues so far.

Hi!Thanks for the comment - totally forgot the rods... I have just added them :)

Thanks for the kind words and the vote! Appreciate it!

Strange! I got the same error... I read on the Fusion forums that some versions had a flaw that causes this - and that I could try saving it with the newest Fusion release. I have just done that, and now it works once again!

Hi! I simply chose MDF because it was cheapest and it is easy to work with. I have done another machine in plywood, and that also works fine. The plywood is more sturdy, but this machine does not require it.

Yep. There are 6 alu angle pieces with 4 bearings each = 24. Then there is 8 u-bolts with one bearing each = 8. And finally there is a bearing in each end of the threaded rod used for lead screw = 6. So all in all 38 :)

I buy bearings at reprapworld.com They are not that costly. But ebay may be cheaper!

That shield should work just fine!

Yes that is correct. I use 3 limit switches, or microswitches - so 1 per axis :)

Hi! Yes it should work with any grbl shield having at least 3 drivers. I just found the cheapest on ebay. But that may also be why I had to modify it a bit. It wasn't an issue for me though...

No problem! I can see now, that of course not everyone is familiar with Fusion 360... I look forward to see your result!Best regards,Mogens

I paid around 200 USD in parts for this machine. That includes everything, also the dremel (not a genuine dremel), controller and PSU.

I have just added drawings showing the MDF parts with measurements. Hope that helps!

No, why should it be that? Based on the parts I have done with it, I have gotten an accuracy that is at least 1mm.

Hi,You should open the file in the provided link in fusion 360 (step 2). In there you can measure everything out :)Question 2 I don't understand?Software is just plain GRBL, and settings is shown in Step 4. There is not much programming to it.

Hi, The idea is that you should be able to build it yourself quite easily. It is not available for purchase :)

Hi!The bearings are not attached to the u-bolts. But they stay where they are, since the u-bolts sits tight against the pipes. The bearings are unable to move to the sides when everything is tightened.The alignment has not been a problem - everything runs smooth. I have built 10 of these machines in the current design revision, to ensure that the design was okay. Essentially it all comes down to how much effort is put into building the machine (ensuring that all measurements are kept within a small margin). The machine design here is a bit forgiving though. The MDF is a bit soft, so some of the misalignment will be removed through this.

Hi! I can't see any problem in just extending the different axes according to your needs. So just extend the tubes, threaded rods and MDF parts :)

Hi! Thanks for the support!You can use something called img2gcode (google it), for converting an image to gcode. It has some settings enabling you to cut out the dark areas.The machine should be able to handle a reasonable size trim router. If the machine gets wobbly, you may have to offer some speed. But I believe a larger tool than a dremel can be used without causing to much of an issue.

Thanks! - I have 1.25mm/rev on an M8 threaded rod, 200 steps is one revolution on a full step stepper, I use 1/16 microstepping giving: 200*16/1.25 = 2560 steps/mm

Hi again! Just got a friend of mine to open the STEP file in Rhinoceros, and it opened up fine. It's a bit heavy, but that's because it's a combined model. But if you have trouble, try out Fusion 360. Unfortunately I don't know any with a SolidWorks licens, so could not try that one out for you.

Hi! I am not the creator of the robot arm design here. Unfortunately I can't remember where I got it from (it's like 8 or 10 years since I found it). I do have a dxf file which I downloaded at that time. The design was open source, but I can't credit the author for the reasons just mentioned.

Hi!I have just added a couple of fotos, showing what the machine can do. Unfortunately the machine is sold, so I can't even make new examples to show. But I have been doing engraving in MDF, which it also did very well - just don't have any pictures of it.

Hi! I did some routing in 1mm aluminium. The important thing is to get the right spindle rpm and the right bit. The threaded rod is just regular 8mm from the hardware store. Best regards, Mogens