First Impressions ... Some good, some bad

[RoySalisbury]

Got my V-One yesterday and unpacked it and got it all setup. Packing was great. Very nice, well protected and secure.

I started with the "Hello World" sample just so I could make sure I had a baseline of what to expect and if anything went wrong there were plenty of examples to compare it too. I noticed that the "long runs" of ink were near perfect, whereas the shorter runs and short bends is where any defects would show (extra blobs of ink, not sharp turns but rather "curves"). All in all, it worked.

So now I started my own board. Nothing fancy. Some relays, NPN's, Resistors, and connectors. I made sure to use the DRC file for eagle that Voltera provides so all my routing and traces were within specs. But even the DRC let a few things slip through.

1. The traces were allowed to close to the edge of the board. While it could print fine, the baking would have had issues sitting on the rails. So I masked off that area with a restriction area and re-routed.

2. Vias just to close to traces and other components. Since the ink is not perfect (from what I saw in the Hello World) There might have been issues with the vias blending into some pads or traces. So I manually moved the vias to give better clearance.

Now for the printing. I was using one of the pre-printed Pi-Hat boards you have. Getting these aligned is a pain. At first it wanted to align on one of the holes, and then on one of the not yet printed pads, I re-did the alignment using two of the pads there WERE pre-printed.

1. I would recommend getting fudical marks on these to align to and allowing the software to choose from a list of YOUR boards to aid in alignment.

2. Probing. Allow certain areas of the board to be masked off from probing. More than once it was probing in the areas of the pre-drilled holes and the probe went into one of the holes and when it went to move it got caught on the edge of the hole (or solder trace) and when it shifted, the whole board moved with it (it was snug in the clams, but not overly solid).

Once I go to the printing stage, it did the initial test pattern first. That was fine, however you must be VERY CAREFUL in cleaning off the ink that you don't move the board when you do it (again, it was snug, but not overly solid).

After about halfway through printing I noted that it was actually printing on top of the re-printed traces at the top of the board and even outlining the holes. So, I stopped the print, cleaned off the board and started the whole process over. This time I excluded the pre-printed areas. This could be solved by being able to select the pre-printed board from a list.

One I got it all aligned, probed and print test done again, I let it print the entire circuit. There were a few issues.

1. Some pads did not print at all. It tried to print them, but only tiny bits of ink were noticeable. I selectively went back and tried to print these items, but until I adjusted the trace height (from .10 to .13) it would not print. I suspect that the probing was off in that area. Odd since a pad right next to it printed just fine.

2. Some vias were not "circles" of ink, but rather "blobs". I'm also guessing this was due to the probing and not printing at the correct height.

3. The pads for my relay are very narrow and the amount of ink printed was just not enough to actually use. For these I went back into Eagle and updated the footprint of the device so it is wider. Its a part that I had to create for Eagle, and use the exact specs from the data sheet, but increasing it will not be an issue.

Finally the solder. that was perfect. If anything I would be happy just using the printer for putting down the solder. Though one issue. The formulation you have dries on the pads pretty quick. I only did the solder for the Hello World demo, but by the time I got to putting the 555 chip in place, the solder had dried. Lucky for me that in the re-flow stage it all worked itself out and I got strong bonds(though one of the led's was raised up into a bead of solder and not flat on the board).

One last thing. And this is the BIG, BIG, BIG problem. It seems the the belts used to move the "bridge" forward and back seem to slip a lot when doing initial movements. For example when going form the "home" position to the PCB, it moves VERY fast and does not have any "ramp up". During this time it slips the belt drive (making a terrible noise). I would suggest changing the software to go slower and ramp up the speed if necessary. Its not far to travel, so it does not need to go super fast anyway. this happened 3 or 4 times to me.

Roy

[mewertow]

Thanks for posting @RoySalisbury, I've shared your feedback with the team, and hopefully we can incorporate some of your suggestions.

Pre-loading templates - Since the software can't distinguish the template file from your design, it won't always make the best choices for fiducials. As you found out, you can override the suggestions with your own choice (the pads on the template are what I usually use) - but it would be really nice if the software could somehow distinguish the template features from your design.

"Selective" Probing - This is one we've heard a lot of requests for - personally I'd find it really useful as well. Selective probing would allow dispensing paste on areas of a populated board, or touching up features after the fact.

Probe falling into holes - This one we're currently trying to address by updating our probing algorithm. It's a bit of a head-scratcher, but we're hoping we'll have a catch-all solution shortly.

Inconsistent ink deposits - As you suspected, this very likely has something to do probing / height mapping. Since you needed to increase the height, my guess would be the probe may have been off due to a hole in the area, or potentially it was because of a violent vibration of some sort - 30um is quite a big jump. We're hoping to detect and address this kind of noise with the algorithm update as well.

V-One printing on template features - This one is related to the fiducial issue - since the V-One can't distinguish between the features, it'll print everything unless otherwise instructed to do so. We'll see what we can do here.

Bridge bucking like a horse - Now that's a weird one... I'm checking with our production guys to see what they think. There is variation between rails in the units, but it definitely shouldn't be enough to cause what you're seeing. We'll see about pushing a software update with an acceleration adjustment - however, if you can get it to reproduce the issue and take a video, send that to support@voltera.io and I'll have a look. It may be that there is an obstruction on the bridge rail.

• Matt

[RoySalisbury]

I kept getting the probe falling into the pre-drilled holes on the Pi-Hat board every time, and think I came up with a quick solution. Since you say we can print on kapton tape and it will withstand the baking process, I just masked it off on the board. The probe no longer falls through and the tollerance is small enough that the probing will handle it.

I did both sides since this is a two-layer board. Its baking now, so will see how it works out.

Roy

[mewertow]

Good thinking - let me know how it turns out.

I just realized that you may be using the PI_HAT_RAW file, instead of the PI_HAT file - the PI_HAT_RAW file includes the holes for reference, so you can see the connections between pads and holes, while the PI_HAT file by itself doesn't include the holes at all, so you shouldn't have problems with falling into the holes. See the image below for clarification - PI_HAT_RAW is on the left, while PI_HAT is on the right.

Using PI_HAT instead might save you some headaches.

• Matt

[RoySalisbury]

Another issue I came across tonight. If the PCB is not perfectly flat, or if the bed is not perfectly flat, the board might not site exactly perfect on the bed. I clamped it down as tight as I could, but you could press down in one of the corners and see the PCB reflex a bit. Its really evident when it probes in that area as the probe will push down on the board. That throws off the height just enough that the printing is not even. This is what happened in my first print from my review above.

I would suggest a bit better clamping of the board so it can be pressed DOWN in all four corners. Perhaps some type of spring loaded pins attached to the clamping rails. Usually there is "some" room on the corners that are not going to get in the way of the print head.

BTW, I hope all these "future enhancements" that get suggested are going to be backwards compatable via an upgrade package.

It would also be nice to have a stand alone baking pad (separate product) so I can print another PCB while the first is baking.

Roy

[RoySalisbury]

I just realized that you may be using the PI_HAT_RAW file, instead of the PI_HAT file

Well ...: That would explain things.

I'll switch my board file over to the other one.

Thanks.

Roy

[mewertow]

Glad to help - hopefully when we have an updated probe algorithm to better account for holes, you'll be able to use either file without problems.

Its really evident when it probes in that area as the probe will push down on the board. That throws off the height just enough that the printing is not even. This is what happened in my first print from my review above.

That would definitely do it - probe deflecting the substrate is something we see a lot more when users try a flex substrate that isn't perfectly flush with the surface, though it can happen with rigid boards. Since the clamping bars are flat, screwing in too tight will cause them to bow, which may lead to uneven clamping. If possible, try clamping just enough so it's held in place, preferably with a few mm of board under the clamp (which is tough to do sometimes if the design goes close to the edge, though).

I hope all these "future enhancements" that get suggested are going to be backwards compatable via an upgrade package.

We try to release software updates on a semi-regular schedule as we address bugs and come out with enhancements. These updates are automatically downloaded through the V-One application on startup, and we'll also typically post in the community forums for major updates.

• Matt

[RoySalisbury]

@mewertow

Good thinking - let me know how it turns out.

Just to let you know, the kapton tape worked out great. survived the baking and pealed right off after it was done. However, using the correct eagle board works better.

Roy