LitePlacer.com forums

Knas wrote:Could someone do a quick rundown on the proper procedure of adding tapes? I'm having no luck so far, it seems i cannot set the orientation correctly?

Update:
I keep getting unhandled null pointer exceptions every time i try to measure the number of components on a tape, it does measure all the holes correctly but after that when it's trying to go to the first component it fails in MainForm.FindPositionAndMoveToClosest, i get the same error in another function when i try to reconnect the cameras - it seems it won't always find them so i have to restart the software a few times before it works..

Any help greatly appreciated!

-Karl

I would like to figure out how to get better debugging information from a crash - not sure how. can you provide a stack trace or something? Did the filters help fix your problem?

there are 2 ways to add - by holes or by image of part. if you are doing holes, then move camera over a hole - add entry and configure options - the right click and select hole based calibration.

Covert wrote:I had trouble with little parts on black tape like SOD323 and SOT23. I found the needle would push the component into the tape so hard it would stick to the double sided tape. I changed the PickupZ on the fly in the Tape table and found 1.3mm less was about perfect.

I checked the code and I see it adds 0.5mm.

MainForm.cs

Code: Select all

double Z = tape.PickupZ + 0.5 - 1.3; //not sure why the .5 is there - increased pressure? //covert back off by 1.3. seems to push parts into tape too hard.

quick bug fix to keep me running. I think this settings needs to be accessible in calibrations ?

Why the double-sided tape? Are the parts in a tape or just sitting on double-sided tape?

Not sure why the .5 - hence my comment. It makes more sense making this a per-component configuraiton if it's an issue others are having. easy to add an Z-offset that is added to the pickup height if others want.

Also - i'm not doing as much development - the code is fairly usable for me. I'm going to make a devel fork and a bugfix fork. Would like to get others to start contributing if they are interested. willing to try git again if it makes merging easier but i will need to learn how to use the cli as the gui is crap.

Covert wrote: I had trouble with little parts on black tape like SOD323 and SOT23. I found the needle would push the component into the tape so hard it would stick to the double sided tape.

We changed the Z-limit switch to a micro-switch to get less hysteresis, hence less pressure on the component, when the pickup height is determined.

thereza wrote: Not sure why the .5 - hence my comment. It makes more sense making this a per-component configuraiton if it's an issue others are having. easy to add an Z-offset that is added to the pickup height if others want.

The .5 seems to be arbitrary. I removed it from the code, since the additional 0.5mm activated the limit (micro) switch again. I don't think there is additional pressure needed.

thereza wrote: Also - i'm not doing as much development - the code is fairly usable for me. I'm going to make a devel fork and a bugfix fork. Would like to get others to start contributing if they are interested. willing to try git again if it makes merging easier but i will need to learn how to use the cli as the gui is crap.

First of all thanks for the alternative version. I would be willing to contribute. I made some small fixes locally, which improved stability/usability for me quite a bit.
If you would be willing to go back to git that would be great.

During development I noticed that sometimes, the machine tried to pick up parts without actually touching them. The table or the double side tape can be warped a bit. There is no Z calibration (needle tip to table surface variation across the whole working area) either. The 0.5mm is there to ensure that the needle will touch the part when picking it up.

Maybe it would be better to use half of the distance between the limit switch turning on and off, if that is needed. From my experience, this value is not required. But we are using a stronger vacuum pump...

I'm keen to contribute. I have a few ideas I want to add and I would like to share it back. I have also made a few changes locally for stability. I don't mind how it's done, SVN, git. Just need to work out a way for those of us actively working on it.

glad to hear that others want to contribute. i'm at a point where i think the code suites my needs so i'm happy to let others drive for a bit. I still have some work I want to do on component placement - part detection - etc. And support for more types of tapes (i.e. component tapes) as well as better optical detection to see if a part is present or not. But will not have time for a bit to work on it for a bit.

JuKu - do you want to manage the code moving forward? The only thing I want to be anal about is the reduction of redundant code. If functions can be combined, they should be. The mainform.cs code is finally getting to be a manageable number of lines. The code still needs to be reorganized further. I would recommend a bugfix fork and a new features fork. Also it would be cool if you acknowledged my contribution someplace.

With the latest updates, I've started adding right click context menus for the pictures, some bug fixes, added the ability to draw arbitrary arrows.

-reza

thereza wrote: the part is http://cache.freescale.com/files/sensor ... X4250A.pdf - perhaps not optimal but works best of the sensors I had laying around.

I grabbed the same part, hope you don't mind me freeloading

Where did you get the +5V for it from? The TinyG is all 3.3/12/24v and the pressure sensor datasheet claims it needs 5V +/- 3%.

Did you steal it from the TinyG's USB connector?

thereza wrote:JuKu - do you want to manage the code moving forward? The only thing I want to be anal about is the reduction of redundant code. If functions can be combined, they should be. The mainform.cs code is finally getting to be a manageable number of lines. The code still needs to be reorganized further. I would recommend a bugfix fork and a new features fork. Also it would be cool if you acknowledged my contribution someplace.

I assume once your version is as you want it, you want to move forward. It is way too good work to be abandoned, so yes, I want to take care of it. It is just a matter of practicalities. You hate Github, that I'm currently using. I don't have any opinion on what version control or what host to use. As for the contribution acknowledgement, of course. Please indicate how you would like to do that (with full name, with company name also, with nickname only, other points). I'm honored for the effort you have given to the small community and you deserve all praise for that.

WayOutWest wrote:

thereza wrote: the part is http://cache.freescale.com/files/sensor ... X4250A.pdf - perhaps not optimal but works best of the sensors I had laying around.

I grabbed the same part, hope you don't mind me freeloading

Where did you get the +5V for it from? The TinyG is all 3.3/12/24v and the pressure sensor datasheet claims it needs 5V +/- 3%.

Did you steal it from the TinyG's USB connector?

I just tested it with a 3.3v supply (whatever the tiny outputs) and it just dropped it's output voltage proportionally but it still was responsive -- tapped from one of the limit switches. So just use that. Don't forget to add a cap for noise filtering on the input pin, and you can get a T fitting from most hardare stores.