Templates and varying Z height

[DavidYon47]

I got some Pi Hat templates recently and was playing a game of "let's see if I can print a 0.5mm FPC connector footprint". I'm pretty hopeful about the Pi template as it solves a lot of the issues I had the last time I tried doing a hat with the V One (mainly the stability of the connector, and having a lot of vias). For me, my OtherMill makes quick work of the holes and cutting the outline perfectly, but having real plated-through holes with breakout pads on both side of the board is really handy.

The FPC connector required careful calibration of the Z height and flow rate, and the resulting footprint looked promising, at least unbaked. But I destroyed a nozzle during the experiment, and I think I know why.

When probing for Z height, the V One doesn't seem to account for the notion that there may be copper pads there already, which are ever so slightly raised above the substrate. So with a Z height of 0.08, I'm pretty sure the nozzle got bent when it tried to lay down ink over one or more of the existing pads on the template. That in turn caused a slit at the bottom of the taper, which slowly oozed ink during the rest of the print.

At first I thought that perhaps next time I would deselect the template's pads, print, then recalibrate the Z height to be higher. But that only goes so far. Not only does the toolpath essentially print a pad on top of pad, but the traces extend into the pad as well. So I'd run the risk of a nozzle collision on one of the traces that terminates on a template's pad.

Any creative ideas on this? I'm not familiar enough with Gerber and Eagle's CAM processing to know how you might leave hints for the V One to watch for the speed bumps on a template. Or if such a thing is even possible.

[DavidYon47]

Wow, am I the only one with this problem?

[mewertow]

Hey @DavidYon47,

Your theory sounds reasonable - unless the V-One specifically probed the soldermask-covered traces and around them, it wouldn't know to adjust the height. This shouldn't be a problem for typical print heights at 0.1mm, but at lower print heights, the probe accuracy + copper thickness + soldermask thickness could start to get dangerous.

In general I would suggest the approach you described (selectively print only the low features first, then increase the height for the rest) - however it seems like this might not work for your case.

Without seeing your specific design, what you could try is to split the design into two separate Gerber files, with low features / alignment in one, and high features / alignment features on the other. Use Aligned Print and the same alignment features to align the two prints. If possible, leave the center clear for the calibration step. If this isn't possible, you'd have to turn the gear clockwise so no ink comes out during the 'Calibrate' step, then go to the Print step and use selective print to calibrate on-the-fly with low-risk features (i.e. features with a lot of ink, like large pads). Of course, that's not a standard procedure, so it's not guaranteed to work great unless you're pretty familiar with calibration.

• Matt

[DavidYon47]

Yeah, but I also think you're correct in that if the Z-probing hits the pads, then it probably takes those into account. So this might not be the issue I thought it was.

I've been playing with some new ink (haven't tried the solder yet), and I'm still going through nozzles like candy. This time around I don't think it was the pads, instead I broke two nozzles in a row due to some other sort of error in probing. The difference this time was that since I was going to actually bake, I had my mill pre-drill the holes and vias. (request for feature: a Pi Hat template that doesn't have the corner holes, since it just wastes time on my mill)

So what happened this time was that the ink cart started a trace that began at a through-hole, and the edge of that hole was higher than the probing accounted for. So the nozzle dove deep enough to clip the edge of the hole, bending it badly, and that was that.

When I backed things up enough to re-align/re-probe the board, the Z-probe also dived into the hole just enough to gouge across the board and then the software threw an error. That happened twice in a row (odd it didn't happen the first time), I then thought to pull the clamps further away so they were closer to the edges of the board. Since this was the 2nd side, I think I was getting some flex from clamping that the probing wasn't accounting for. At any rate, that allowed the probe to succeed, which allowed the ink to lay down without eating yet another nozzle, and now both sides are baked.

Then it was on to after-school kid duty.

I gotta say that the cost of the nozzles are adding up. They bend/break just by looking at them funny. I know that they are inherently fragile, but it would be nice if you could get the manufacturing costs down somehow so that it's not so painful on this end every time one breaks or clogs, which happens all too frequently.

[DavidYon47]

The good news is that the 0.5mm spacing FPC connector printed nicely with the new ink once I dialed in the Z height and flow rate.

I did have to use an xacto to separate some of the pads, but no biggie. Will be interesting to see how well it adheres once I get that far.

[cborn]

Interesting point on nozzle cost/fragility. Is there a reason the nozzle tip has to be so fine externally? My 3D printer nozzles are much thicker and more robust. Is it because the ink/paste will stick to a wider tube wall and smudge?