Fine pitch footprints (0.5 mm and lower)

[a2retro]

@jazozaya I see what you mean about the merging of the two. I am hoping you can implement that sooner rather then later but in the mean time I will build some alternate footprints for current projects in the pipe.

Since it would still be nice to have a symbol/footprint combo, I have tried embedding the smallest pad allowable into a trace in a custom PQ208 footprint

and here i have joined new style footprint on the left to a stock PQ208 footprint on the right. Where they meet it is clean but where it connects to a stock footprint we get the bump. I am guessing that at .8mm and above that has not been an issue.

So (1) is with the pad on the right of the trace (best option) (2) is with the pad on the left of the trace, (you will notice a minor tick) , and (3) there's still the offset trace when finally connecting to a standard footprint

[jazozaya]

Hi @a2retro,

So we got a chance to work on some fine pitch goodness today.

I made some small tweaks to the algorithms so that single pass lines are actually truncated at pads. What this means is that they will not enter the outline of the pad, and if they do, they will enter by the tracepad_penetration_ value in the settings window. This prevents the V-One from double dispensing on fine pitch pads. Double dispensing is bad as it leads to shorts.

Below is a 0.5 mm pitch footprint:

If you look closely, the traces and the pads are now being drawn as two separate lines that are almost touching. The pads are still a tiny bit offset from the traces - this is actually a trivial bug and is next on my list.

We took a quick video of it dispensing and have posted it here: https://youtu.be/Z44uQ7EotqI

The print looks quite decent and at first inspection it doesn't look like there are any shorts. The board is currently baking so we'll have to update you on how it turned out.

The next steps are to merge the pad and traces together to form a single line and then to try dispensing solder paste and reflowing some components on.

Cheers,
Jesus Zozaya

[a2retro]

Thanks for the reply- looking good and waiting patiently for your next update

[emcniece]

Really enjoying seeing the V-Team keeping in touch and releasing updates. Keep up the great work!

[a2retro]

Did some selective print testing today of a PQ208 .5mm pitch foot print today.

This file has some 2.54mm holes for alignment and the PQ208 file is comprised of a footprint that has the smallest possible pad embedded in a short trace. In the actual eagle file where the footprint is used I have added longer segments to each pad.

For the most part I think this will be usable if there are a few issues corrected. I think Jesus is probably already aware but here goes any way.

1) Because the resulting line in Eagle is created as 3 discreet segments, The V-One also prints as 3 discrete segments. No surprise but because they are all the same width it would make sense to optimize this. In the meantime I guess having a pad embedded in a trace isn't the way to go.

2) I don't know why but the longest segment (the one added in the drawing) is printed in 3 passes where the smaller segment which is part of the footprint is printed as two passes.

3) At the end the actual pad was added as a dot to each line. This works on some but started smearing on others. It should be optimized out if it it's possible.

208 QFP_5mm1.GTL (20.3 KB)
208 QFP_5mm1.TXT (579 Bytes)

[jazozaya]

Hi @a2retro,

Thanks for posting - nice pictures! Love the contrast

For the viewers at home, I decided to print a portion of the file that @a2retro submitted. Here is the link

2) I don't know why but the longest segment (the one added in the drawing) is printed in 3 passes where the smaller segment which is part of the footprint is printed as two passes.

Actually - not quite. It looks like that in the video, but it's a bit more complicated.

Imagine you had a thin 10mm long trace along the X axis. To print it, the nozzle will come down on one side, (X=0) rotate the gear forward to apply pressure, and then move in X until it reaches the end of a trace (X=10). When it reaches the end, the printer will rotate the gear back to relieve pressure and raise a tiny amount, it will then 'backtrack' onto the trace for about 2mm (X=8) before fully raising.

This is what we call the 'anti-stringing' distance, and it is a configurable parameter in the settings panel. The default for conductive ink is 2mm and the default for solder paste is 0.1mm. This parameter exists to encourage the 'stringing' in the ink to fall back onto itself and avoid shorting.

In traces shorter than 2 mm, the nozzle will actually backtrack twice when it reaches the starting point. So it looks like a pinball bouncing back and forth...

---

The picture above shows the 3 segments of the pad. It looks like the tiny pad is collapsing into a single diagonal line. It should be pretty trivial to get the software to discard lines these small.

[jazozaya]

By the way,

I have fixed the small offset issue seen below. Pads will now line up with traces properly. The fix should be rolled out in the next release.

[DavidYon47]

@jazozaya: at the risk of hijacking the thread, any ideas as to if/when you might minimize.or eliminate the dimple issue with pads? Having the pads not be flat gets cumbersome as the component count goes up.

[jazozaya]

@DavidYon47,

There is one way to reduce it...

The default print height is 0.100 mm for conductive ink. You can try lowering the print height to 0.06 mm in the settings panel. We've seen that printing at a lower height reduces dimples and improves soldering significantly, however you do take a small hit in trace conductivity

You can try experimenting with lower heights, but please do not go lower than 0.04 mm... the nozzle will not be very forgiving if you do

[jazozaya]

@a2retro,

I showed your pictures to some of our team members and they loved them.

We've always struggled with getting good close up shots. Did you use a particular camera or lighting system? Or was it some photoshop magic?

[a2retro]

Actually, I used a cheap 1.3 MP USB digital microscope, the base would not tighten so i had to steady it with my hand. Also the USB driver has issues on windows 10 so it's a bit of a miracle that I was able to get them at all

[mharris]

You guys need a good macro lens on a decent camera.. that means interchangable lens camera or a DSLR.

We have a couple of these around the office: https://strobepro.com/collections/continuous-led/products/quadstar-1000w-continuous-light?variant=366019595 and I have them at home too - great portable suns with soft light, perfect for video or photos, or when you need some more light on your bandsaw

Uncropped with Nikon 105mm Macro, ambient light (hence very shallow DoF)

Uncropped with same, ambient light hand held (hence a bit blurry lol) Smallest track is 0.2mm.

Same deal, but further away. Magnifying desk lamp providing side light:

Keep in mind those are 24mpix scaled down, so when you are 1:1 on the image sensor, you can see the grains in the ink and solder paste.

Same lens on a 12mpix full frame DSLR, from a few years ago:

[mewertow]

Hey everyone,

There are a few software enhancements that we've had on the backburner that we expect will help with pushing the spec towards a lower standard pin-pitch, though we haven't had a chance to get to them just yet - this is hopefully something we'll have a chance to tackle in the coming months.

However, some of you may have noticed that we now stock nozzles with a smaller ID for printing high-resolution features - these can be found in the Experiment section of the store. If you'd like to experiment with fine-pitch printing, these nozzles are definitely worth trying out.

The key considerations for achieving finer tracewidth with the 150um nozzles are:

1. USE LOWER PRINT HEIGHT. The print height is directly tied to the minimum trace width that you can achieve. With a print height of about 40-60um (0.04 - 0.06 mm), we've been able to achieve tracewidths around 6mil (150um). Going any lower than 40um, however, will become a bit risky, and increase the chance of crashing the nozzle (there are some other safeguards you can take that can help prevent this, namely decreasing the Probe Pitch setting).

2. CAREFULLY ADJUST THE FLOW RATE. Use the smallest increment of adjustment, and take your time with the calibration step to get the right trace width.

3. REDUCE THE PASS SPACING PARAMETER. This is to ensure that pads are filled in correctly. You'll want to reduce this parameter to within 1 mil below your desired tracewidth. For example, if you're printing 6mil (150um) traces, adjust pass spacing to between 0.13mm-0.15mm to start.

If you're really trying to push the boundaries, there are other settings you will likely have to adjust - I'd encourage users to get in touch or post here, to address these requirements on a case-by-case basis. The smallest tracewidth I've been able to achieve is 4mil (100um), though it took a fair amount of trial and error - if anyone can beat that, I'd be excited to see your results!

[michael]

I got a couple extra nozzles in my order, but I'm not sure if they are the 150um extra fine types. Are the 150um nozzles always labeled as such, or is there some other way to determine this?

[michael]

I've figured out that the general minimum pad specs are 0.8mm/0.5mm but don't know how to interpret if say QFN-40, QFN-48 (like ESP32 chips), or SOIC-16 would be easy, difficult, or impossible to use. Do most people simply take the time to download and read the specifications for every single component while shopping? This is time consuming and error prone.

[mewertow]

Hey @michael ,

The 150um nozzles will have a blue plastic hub instead of black. Regarding the pad specs, typically I look at the specific part or package datasheet for the pin-to-pin pitch. However, there are also handy references that will give you the typical pin pitch range for particular package types - for example, see here, under the P column.

A quick side note regarding minimum pad specs - we've had pretty good results printing down to 0.65mm pin pitch with the conductive ink and standard black nozzles. You can also try going lower with the fine-pitch 150um nozzles, if you're up for some experimentation.

• Matt

[michael]

Thanks, I'll check the color of my spare nozzles to find if they are blue (15um) or black (standard.) The Wikipedia package reference you mention is quite useful, thanks for suggesting it. The pad pitch is given for about half the components I've ordered recently. I'll keep my fingers crossed for the other ones, but am optimistic since almost everything listed in the Wikipedia reference is .8mm or thicker. It seems the Voltera can print fine enough ink and solder for almost any component I might need.

[ka911]

How about we get some settings we can save as preferences? I would like to (need to) experiment with the 150um and just got them. But it would be nice to save my "settings" as "Print Profiles" that I can recall later without too much hassle without resorting to notes hidden somewhere in my old memory banks.

Any plans on "calculating" dispensing based on material flow/nozzle width? You have the variables Print Height, Flow Rate and Pass spacing - so can you calculate either "backwards" ?

Btw seems like Pass Spacing counts in 0.1mm steps and so default is 0.18 mm but pressing up or down select 0.1 or 0.2 - not 0.19 or 0.17 - is that on purpose? or a little bug

[jazozaya]

Hi @ka911,

How about we get some settings we can save as preferences?

Excellent request. Right now there's a 'Restore Defaults' button but there is no clean way to load or export a print profile. We'll add it to the feature queue.

The current software assumes that you are using the standard nozzles provided, we could certainly add a parameter that lets you indicate the nozzle in use and have it automatically calculate your pass spacing & height to get good prints.

Btw seems like Pass Spacing counts in 0.1mm steps and so default is 0.18 mm but pressing up or down select 0.1 or 0.2 - not 0.19 or 0.17 - is that on purpose? or a little bug

Not on purpose, definitely a little bug. I think all of the settings change in 0.1 increments, I understand this can be annoying.

[ka911]

It is actually rather random how each settings preference counts up / down... hmm maybe time for a settings preference overhaul. Some does .01 +/- - others do 0.1 +/- insisting on whole numbers - others do 0.1 +/- based on the number that is in the box to begin with - so a bit all over the place.

Reg settings - hmm maybe a "select nozzle" - then have sets of profiles for each? (Like I might name mine - rough/medium/fine/super fine - dependent on how critical each is. And it would be nice with a note field for each set of settings - so you can put in memo "Test with .5mm pin pitch ok - remember to do xyz BEFORE print etc"