My building Experiance.

[Covert]

Last night a friend arrived to help me start the build. I have been waiting for a time we are both free, even already have the kit for over two weeks. So far we have got "VII: Y-Bar, Right Roller Plate" step complete. We are moving along well. Having all the parts labeled and a great building instructions has made the process smooth.

Worse part was tapping all the extrusion ends. I have found some places offer this for a minimal cost when purchasing extrusion. I would recommend anyone to pay this little extra. The collect your tools page http://www.liteplacer.com/the-machine/a ... the-tools/" does not mention the M5 tap, this might be because the screws are selfs tapping ?

The biggest problem was a bad thread on the circular hook. My friend who is a fitter and tuner took the job over of betting it threaded. Took some time. The entire thread on the hook was bad my friend said. But since it was such mild steel he managed get the nuts working on it. http://www.liteplacer.com/the-machine/a ... ular-hook/

Back at the build tonight but during the day I'm going to look at printing some chain cable. I'm also going to explore using a 45deg mirror on the upwards facing camera to be able to mount the endoscope horizontal.

[JuKu]

> this might be because the screws are selfs tapping?

Indeed they are. Too late, but all you needed to do is turn the screws in. OF course, I would haev replaced the hook, but I'm clad you were able to solve it.

[Covert]

Thanks for the comments JuKu. I have decided to replace the Circular Hook. I'll just replace it with something from the local hardware store. The thread is still to tight to easy adjustment.

Day 2 last night of the build. It was mostly smooth. The 608ZZ flanged bearings where extremely tight. Had to use a vice to press them in. Looking back I should of just put the plates in the reflow over at 150dec C for a little.. I don't want anything protruding under the base board so I have decided to use a mirror on the upwards camera. I've attached a image of the prototype. Going to include a defuser holder today and a mounting solution on the bottom of the holder. In the assembly instructions 4 holes are shown for the cable chain mounting but only 2 holes are shown in the Plates Guide page. The plates I got in the kit match the plate guide. The 2 holes unfortunately did not match my cable chain ends (the missing holes would of been right). I 3D printed some cable chain ends to fix this. There is no mounting solution for the rear X cable chain that I can find. It's the problem I'm working on at the moment.

End of day 2 (night 2) resulted in all the mechanical items, motors, pullys, belts. All except the cable chain.

Whats left is the cable chain, Cameras (printing light and defuser holders), wiring.

JuKu have you considered replacing the X drive shaft and assembly to hold it with a second NEMA17 on the left side ? Depending on the specifications of the TinyG driver it might be able to drive 2 motors at once.

[mrandt]

Hi Covert,

you're making good progress there

Just my two cents regarding some points of your recent post:

The 608ZZ flanged bearings where extremely tight.

I have built my machine from scratch based on Juha's plans long time before the kits became available. I had all the steel plates laser cut and suffered small problems with too tight tolerances. In order not to damage any of the bearings, I carefully filed the holes larger until it was a snug fit. Maybe the same is true for the kit plates?

have you considered replacing the X drive shaft and assembly to hold it with a second NEMA17 on the left side ?

I believe Juha's design was inspired in part by ShapeOko - using a second, parallel stepper for the Y axis is a common solution with those machines. TinyG could deliver plenty of current, so that was never the limiting factor. I must say though, that I like Juha's solution with the horizontal shaft much better than two parallel motors.

It is very elegant and you can be sure that you have no synchronisation issues, no squareness error due to missed steps on either side, etc. Last not least, the machine stays square when moving the gantry manually with power switched off - a problem that becomes more obvious with larger frames. My machine is 1000mm wide and still works nice and fast with the shaft solution. If torque ever was a problem, I would just put a stronger motor. In the end, this solution is also cheaper than buying a second motor. But maybe Juha also wants to comment on this...

I have decided to use a mirror on the upwards camera.

I am interested to find out if and how that will perform... Please keep us posted. The upward camera is important for needle wobble correction and - in the future - part orientation corrections. I think it is crucial to have this one aligned well and reduce distortion. Given the focal length, I doubt that you will achieve a focussed and undistorted picture at PCB level - but you may prove me wrong... See my comment here as well:
http://liteplacer.com/phpBB/viewtopic.php?f=3&t=36

Regards
Malte

[Covert]

I am interested to find out if and how that will perform... Please keep us posted. The upward camera is important for needle wobble correction and - in the future - part orientation corrections. I think it is crucial to have this one aligned well and reduce distortion. Given the focal length, I doubt that you will achieve a focussed and undistorted picture at PCB level - but you may prove me wrong... See my comment here as well:
http://liteplacer.com/phpBB/viewtopic.php?f=3&t=36

Regards
Malte

Been reading about the C720 logitech camera in the other thread. I'll test one of them and see my results once I get up and running. It's does seem to save a lot of space. Your feedback is very welcome.

[mrandt]

For me, altering the cameras was never about saving space but getting a better camera picture. I hope to be able to place fine pitch IC and maybe even BGA at some point

Still, the flat camera board and lense with relatively short focal length seem to use less space underneath my table top than the kit endoscope cams - and my assembly is perfectly even with PCB surface level:

Upfacing camera assembly underneath table top

cam_underneath_table.jpg (36.8 KiB) Viewed 8029 times

Camera assembly requires less vertical space underneath table top than front aluminium extrusion

cam_table.jpg (34.35 KiB) Viewed 8029 times

[Covert]

Today I had time to do the wiring, Been short on time due to events outside my control, including days at a hospital. Nice to get back to the bench. I have been tinkering around getting the cable chains in place before now. Wiring is going smooth. It's all rather straight forward. I thought I didn't have enough wires at first but ended up with just enough. 3D Printing a TinyG case from Thinkiverse just to keep it neat. Might get to power up tomorrow. Will see.

.. Only issue today was the bracket for the top Z limit switch was mounted the wrong way and I could not get the switch to engage with the end of the bolt. Lined up perfectly once I removed the right angled bracket and flipped it around.

[Covert]

So finally I have the unit moving and I thought it would not take long to get it into building some PCB's. Still haven't got that far. Seems to go 2 steps forward and one step back.

The FET for the vacuum pump is stuck close. It's not fully closed so the pump just idles away then "rev's" up when the fet gets powered. I will replace it soon but for now it still works as needed.

After a half dozen success pick and place of parts I would get a circle not found error from the downward camera unable to find the circle in the paper tape. I noticed sometimes the camera setup screen would include the zoom setting from the main screen but sometimes it would not. I started to think the software might be zooming in too far with 2 zoom settings so removed the zoom setting I had in the main screen. Still get the problem aster a few parts. I replaced the downward facing camera with a logitec C270. It improved for some time then started getting circle not found. I added another function to the camera list (histogram ?) and it seems more stable. Yet to fully test.

I have to manually adjust the Camera to Needle distance. When I use the calibration procedure in the camera setup screen it seems to be about 0.4mm out. I use a piece on bluetack on the bench and drop the needle down into it. Then I jog the camera directly over the indent into the Bluetack. Having a camera to needle test would be nice to check this.

I also found I had my belts loose and giving me backlash. I found the Slack compensation option in the main screen helped a lot for this.

I might get to build my first boards tomorrow. Lets see...

[Covert]

Today I redid the tension on all the belts and did the squareness calibration again. Seems to have improved the accuracy.

Had a role of components in clear tape I don't need so there the test part for today. I can only seem to manage to get 5 or 10 parts down until something goes wrong. It's unable to find the hole in the tape from time to time. I have tried a piece of colored paper under the holes on the tape to see if that helps but still seems to eventually stop from not being able to find a hole. I go to the camera screen and check it view and find the threshold needs to be adjusted again. Nothing has changed not even the light in the room (it's dark outside now).

I'm going to try rmod to see if its any more stable.

[Covert]

Had some loose pullys on the machine as well. Once that was sorted out things became a lot better.

rmod made a huge difference. Manage to place all small parts in white paper tape and some parts on clear tape using part recognition by picture.

I'm going to start tweaking the source code to get it running smoother for my purpose. I hope it will help others.