Some doubts and problem with picking and placing components

[JuKu]

Next, you configure the cameras and set the correct Box sizes (determines px-mm ratio) for up and downlooking cam (Setup cameras tab) - and make sure to put your ruler, caliper or mm grid at PCB level (!). Exact measurement is important here as the measures px will be caculated to mm offsets when compensating needle wobble.

How can I place my caliper at the exact PCB level. If I use a caliper to set box size(that I did),it's height will obviously higher than that of PCB .Couldn't it affect when determining box size?????

Again, without testing, a gut feeling only: If we are talking about a mm or two, the error is likely insignificant, a cm or two would matter.

[android]

"PCB level" means the distance between needle and PCB surface when it barely touches it(one that you do in "needle height"), right?
So obviously if I place a scale on the board(consider both have same thickness),it'll be definitely in PCB level,right?
So why is it necessary to specify 'measure box sizes' in PCB level since the down camera does not move when needle moves. Even in setting up camera location, we click 'needle down' , here also needle will move down to the distance we already measured in "needle height"(PCB level).
Correct me if I'm misunderstanding the term "PCB level".

[JuKu]

> "PCB level" means the distance between needle and PCB surface when it barely touches it(one that you do in "needle height"), right?

No. It is the amount the Z needs to come down from zero (full up) for the needle to touch the PCB.

> So why is it necessary to specify 'measure box sizes' in PCB level

> So that the software can figure out how many mm's a pixel in the camera image is.

[android]

> "PCB level" means the distance between needle and PCB surface when it barely touches it(one that you do in "needle height"), right?

No. It is the amount the Z needs to come down from zero (full up) for the needle to touch the PCB.

> So why is it necessary to specify 'measure box sizes' in PCB level

> So that the software can figure out how many mm's a pixel in the camera image is.

I've done the calibration processes once again.
My PCB level is 34.16mm. I set Up Camera location at the exact centre of needle tip. When machine does needle calibration, needle only goes down to 33.16mm(1mm less than actual PCB level) . Is that normal?
And after calibration, when I click "needle to cam" and takes the needle down (34.16), tip is not at the exact centre.
This time there is no problem with component pick up but placement is wrong.

[JuKu]

Yes, there is a 1mm difference, on purpose. The idea is that optimally, the calibration is done at the level the needle is when placing parts. We are not doing the calibration for each component. The accuracy is most critical on small parts, and I'm guessing 1mm for average small part height.

Does slack compensation make a difference?

[android]

nop,
this is my wobble calibration info,
A: 0.000, X: 0.165, Y: -0.124
A: 22.500, X: -0.488, Y: 0.213
A: 45.000, X: -1.251, Y: -0.069
A: 67.500, X: -1.458, Y: -0.990
A: 90.000, X: -0.832, Y: -1.684
A: 112.500, X: -0.041, Y: -1.581
A: 135.000, X: 0.399, Y: -0.846
A: 157.500, X: 0.028, Y: 0.021
A: 180.000, X: -0.722, Y: 0.199
A: 202.500, X: -1.389, Y: -0.282
A: 225.000, X: -1.430, Y: -1.052
A: 247.500, X: -0.846, Y: -1.636
A: 270.000, X: -0.089, Y: -1.588
A: 292.500, X: 0.447, Y: -0.935
A: 315.000, X: 0.131, Y: -0.158
A: 337.500, X: -0.571, Y: 0.186
A: 360.000, X: -1.148, Y: 0.028
G0 X0 Y0 A-10


I am also uploading the pictures when setting up cam position

1.jpg (235.86 KiB) Viewed 3143 times

and after "needle to cam"

2.jpg (225.78 KiB) Viewed 3143 times

[android]

shouldn't the values at A=0 and A=360 be same????????

[JuKu]

It looks like your A axis has a mechanical problem:

> A: 0.000, X: 0.165, Y: -0.124
> ...
> A: 360.000, X: -1.148, Y: 0.028

When the needle does a full rotation during the calibration, the X position ended up more than a mm away from where it started. The 0 and 360 results should be the same or very close.

Please check that:
- Make a mark on the bigger pulley to see that it does a whole rotation during a calibration and when clicking the A test button on basic setup page.
- The tube rotates freely when you take the belt off from the pulley (You need to loosen the motor for this)
- The belt is reasonable tight. The bearings have little* slack, the belt tightens this slack in X direction
- The flat spring pushes the slack away in Y direction. This does not need very much force, but it needs some
- That the bearings or any other parts are not loose

* If the bearings have excessive slack (like the >1mm seen above) they are faulty and I'll replace them.

Also, your needle nominal offset values look a bit odd. They are much larger than in a standard machine - you have modified the camera or the needle mount, I suppose? This is fine, of course, but the odd thing is that they are so even.

[JuKu]

shouldn't the values at A=0 and A=360 be same????????

Yes. You posted this while I was writing the above message and double checking my own machine.

[RMM]

I now have both camera systems working excellent. The C270 as an up camera is awesome and the image it produces is excellent. The resolution on the needle and consistency of the circle detection is excellent. I can now pick up parts in any rotation, but the positioning still isn't right on and when I tried to place some 0805 components on boards there is no way it would work well enough for a reflow; 0603 components are currently out of the question.

Now that the up camera is taken care of, I realize that I have other mechanical issues with the machine. I have realized that I cannot get the machine to consistently return to the same up camera position. It gets close, but if you home the machine again then return it will almost always be a little ways off of what it was set to in the needle calibration. I spent about 4 hours yesterday on the mechanical side trying to figure this out, but I haven't been able to so far. Here is what I checked:

• Belt Tightness
• Wheel Tightness
• Smooth Rolling
• Rotation Compensation (distance per motor revolution)

The most frustrating part about this is that for someone without any experience with these things there is no way to know how tight the belt should be, or how tight the wheels should be. Everyone says "tight enough" or "not too tight", but if you don't have a point of reference that doesn't help. I feel like I'm running around in circles not knowing what to look for.

How tight should the wheels be? Should I be able to slip the wheels on the slide by hand? It seems like when you go too tight you can tell because the slide doesn't roll smoothly.

My A axis bearings do have quite a bit of slop in them, but I don't think that is the main problem I am experiencing. I will tighten the belt a bit more and try that again.

What information can I provide that would help figure out what is going on? I have spent about 40 hours on this project after the build completion and don't know where else to go from here.