parts moving after placement?

[loki42]

I'm having a lot of trouble placing a long thin connector (DF40C-100DS from Hirose) all my other parts including 0.3mm pitch parts are placed very well, but this part is much longer than anything else and is often placed with rotation or position errors.

I'm using a custom nozzle as I thought it was moving after the nozzle picked it up, but now I'm thinking it could be moving after it's placed. The paste has basically no marks on it, and the parts can be easily moved with even the lightest tap on the board. How should I fix this? Increased placement force? Stickier paste? Placement is with a UIC GX11.

[Aspartame]

Did you check if there i a version with peg/post?

[jmelson]

I have had problems with large, heavy parts slipping on the nozzle during the vision gauging of the parts on my P&P machine.  When I detected that, I went and selected to use a larger diameter nozzle, and that prevented the issue.

Hmm, no marks in the paste?  That may indicate the machine has the wrong component height, and is not putting it all the way down to the board.  Try reducing the height that the machine has been given, then it should bring it down more onto the paste.

Also, a problem that has been driving me a bit crazy is very small, light parts bouncing after perfect placement.  The symptom I see is badly placed parts, wildly off position, with perfect dimples where the leads were pressed down into the solder paste.  This proves the part was initially correctly placed.  I think the issue is that the vacuum/blow off timing for these parts is not correct.  There's a page in the software where you can set these timings for each footprint, and I just have to find the ideal numbers.  This machine is a Quad QSA30A, basically a Samsung CP30 with Quad alignment cameras and Quad electronic feeders.
I'm pretty sure the issue is the nozzle lifts too soon after placing the part, before the vacuum has completely gone away, and that causes the part to lift off the board and the drop off the nozzle.

Not asking for help with this, I know I just have to try a few different settings.
Jon

[loki42]

Did you check if there i a version with peg/post?

yes, for the DF40 series there isn't.

I'm running a custom nozzle now (after testing all the nozzles on all my heads... ) and it still is doing it, so I'm thinking it's not the nozzle. If I inspect multiple times it's not moving around. I suspect the component height might be the issue.

Your idea about vacuum / blow off timing is interesting. I might be able to tell it to push in and then dwell.

[SMTech]

The other thing you can potentially check is how flat the PCB is and adjust underside support if that seems to vary. This is exaggerated on our system as we use a standalone pick and place that holds the PCB in a fixture rather than with rails. Adjustable supports are arranged under the PCB to help keep things flat, but all this is manual and done by eye/feel. Every time you change over a PCB you risk not quite securing the PCB 100% and over the course of a placement cycle something might shift or flex enough to throw out fine pitch placement. This is noticeable for this like long FPC connectors or large things such as 240+pin QFPs.

If I'm not mistaken you are fitting connectors for a Pi compute module and if memory serves the datasheet explicitly says you shouldn't use 2 precisely because some drift is pretty normal, but apparently the Pi foundation doesn't give a flying *** that this makes it difficult for someone that isn't Sony. Dwell time, nozzle choice, placement pressure (if you can set that), blow off & travel speed are definitely areas you could also fiddle with. You could also potentially dive into your machine diagnostics and double check your vacuum levels aren't marginal indicating a need for a filter change/service or adjustment.

[48X24X48X]

It looks like component height set wrongly. Set the height correctly and add a bit extra pushing it downwards like 0.2-0.3mm.

[Accu-Sembly]

Keep in mind that setting a component height in a PnP machine also considers the height of the nozzle.  You mentioned using a custom nozzle, so if the nozzle data itself is off, then it could cause mounting issues even if your component height and mount height are set correctly.

[SMTech]

It looks like component height set wrongly. Set the height correctly and add a bit extra pushing it downwards like 0.2-0.3mm.

It's possible that with a UIC GX11 @Loki42 does not need to rely on crude techniques like over travel in Z because the fitted head has placement force measurement. While a height issue somewhere looks quite likely I would assume it was probably checked and fiddled with before the thread started as part of initial troubleshooting.

[SMTech]

Keep in mind that setting a component height in a PnP machine also considers the height of the nozzle.  You mentioned using a custom nozzle, so if the nozzle data itself is off, then it could cause mounting issues even if your component height and mount height are set correctly.

With that in mind, can a GX11 measure its own tools for tolerance? This is something I can do on an Essemtec..

[loki42]

I haven't run the machine measures itself stuff since I changed to the custom nozzle,  but I was getting the same issue before. I do have a cal kit for the machine but I don't have the force sensor calibration tool.  I do have force sensing on the heads and the spindle force tests look sensible for all spindles but I'm unsure what kind of placement force I should choose and if stuff like dwell time after placing helps.  I might have selected something stupid somewhere. Also I have"improved" board support since my initial runs of this board as I thought that might be the issue so their might still be something going on there.

[loki42]

So I recalibrated both heads,  checked the board support for flatness,  tested a bunch of nozzles and checked the part height with my micrometer instead of just reading it off the data sheet.  I also tested increasing the placement force and dwell time.  After this,  it turns out the multi port nozzles on the 7 spindle head work better than the custom nozzle on the 4 spindle. The 3d print nylon nozzle was preforming a bit better than the 3d stainless steel.  I now get decent placement and minor rotational error. I think this might be as good as it gets with this nozzle set up.  They now seems to stay on a lot better after placement. I tried placing just the connectors and turning the board up side down and walking it a few times and didn't lose any from the panel.  I also double checked my printing so improvements there might have helped to. My new SPI should be arriving at the end of the month so hopefully that helps me tune the print process. 

[jmelson]

I tried placing just the connectors and turning the board up side down and walking it a few times and didn't lose any from the panel.

YIKES!  I would never have the confidence in the paste tackiness to try this!  Well, I guess that shows they are planted well in the paste.
Jon