I have the following few queries regarding the pick and place machine .
1. Can you place the SMD components shown in the attached picture, as the pitch in your video is higher than the ICs I want to place on the board .What is the orientation angle correction possible, as in the video there were some orientation issues with the devices placed.
2. Our part is low precision (attached picture-have dimensions) and comes in tape and we can also place it in tray if required. What is the maximum precision pitch at which you could place the components .
3. Can I modify the software to change the location (x,y,z movement) or configure it for any other motion .Which software are you using, will the package contain the source code.
4. Can I manually move the tool arm using mouse or any joystick or any other way.
5. Can I pick a part from a tray and do the testing and place it back in a Bad tray or Good tray.
Low precision SMD placement from tape query
Low precision SMD placement from tape query
- Attachments
-
- dimensions.PNG (142.7 KiB) Viewed 3489 times
Re: Low precision SMD placement from tape query
1&2: Yes, that is the same pitch as the IC used in the “hello world” example. Please see http://www.liteplacer.com/running-a-job/ page, close to end. Also, see the project logs on http://hackaday.io/project/1755-litepla ... ce-machine. For a video showing IC placement, I will shoot some after I have shipped the already sold kits. I apologize, I am unable to spend time for video before that. And finally, please see the beginning of http://www.liteplacer.com/phpBB/viewtopic.php?f=12&t=15 for discussion about small pitch part improvements in planning.
I have no formal specification for the machine, partly because it is more accurate than my measurement capability, partly because it is hard to put a specification to something that the customers build. The examples show typical achievable results. The rotation resolution is 0.05 degrees, the x and y movement resolution is 0.0125mm. Adding all the other inaccuracies (camera noise on video processing, mechanical flexing of the machine, practical limits of user calibration and so on), I would estimate real accuracy to be some 5x worse than the resolution.
Much of the orientation issues you mention is due to the parts being loose in their pockets. Currently, there is no correction for that, as this error source is well below what reflow requires.
3: The software is custom written in C#. Source code is available at https://github.com/jkuusama/LitePlacer.
4: There is no joystick interface, but you can move the machine by arrow keys and clicking on camera image and with function keys, you can get precise movement up to 0.01mm resolution. But see above about true accuracy – in practice, I don’t see a difference for one 0.01mm move. I don’t know if the machine does not move or if the movement is just too small to detect with my methods. Doing two or three in a row, and it does indeed move.
5: Exactly this function is not in the software. However, all the building blocks for such are, so it shouldn’t be very big task to do. As noted, source code is available.
I have no formal specification for the machine, partly because it is more accurate than my measurement capability, partly because it is hard to put a specification to something that the customers build. The examples show typical achievable results. The rotation resolution is 0.05 degrees, the x and y movement resolution is 0.0125mm. Adding all the other inaccuracies (camera noise on video processing, mechanical flexing of the machine, practical limits of user calibration and so on), I would estimate real accuracy to be some 5x worse than the resolution.
Much of the orientation issues you mention is due to the parts being loose in their pockets. Currently, there is no correction for that, as this error source is well below what reflow requires.
3: The software is custom written in C#. Source code is available at https://github.com/jkuusama/LitePlacer.
4: There is no joystick interface, but you can move the machine by arrow keys and clicking on camera image and with function keys, you can get precise movement up to 0.01mm resolution. But see above about true accuracy – in practice, I don’t see a difference for one 0.01mm move. I don’t know if the machine does not move or if the movement is just too small to detect with my methods. Doing two or three in a row, and it does indeed move.
5: Exactly this function is not in the software. However, all the building blocks for such are, so it shouldn’t be very big task to do. As noted, source code is available.