Close

[R] Embedded electrical components

A project log for SecSavr Suspense [gd0105]

The best of resin with the best of filament; is such a 3D printer possible?

kelvinAkelvinA 07/29/2022 at 19:013 Comments

Hybrid additive manufacturing of 3D electronic systems: https://iopscience.iop.org/article/10.1088/0960-1317/26/10/105005

PDF: https://openresearch.lsbu.ac.uk/download/47301d36e4bef86d066c76d945893c3f7fc8702e786aef4dc988d9f775a6768e/1946572/ttn_microeng_3dp.pdf

I've got no idea why so much (additive manufacturing) research is behind an access wall that only university students or those willing to pay ££ per paper, but I was reading Material issues in additive manufacturing: A review and the last set of images looked very intriguing:

So I looked into this paper and the method they used, which is seen below:
I had a feeling that some kind of dispensing paste was more likely to happen than a decently conductive UV curable resin. If the cleaning roller works as intended, I'd likely be able to have a more streamlined process to what these researchers acheived in 2016. I doubt I'd have to use the oven until the part is finished since they won't be upside down or pushed around by resin in a vat. They use a clear resin with a Tg of 100C. 

Resistivity

For a reference, this is the milliohms/sq for copper.

I've highlighted 70um because that is the same thickness stated from Voltera, who sell a PCB that prints in gels.
I feel like I'm coming in, looking at and swiping every additive manufacturer's tech right now. First BCN3D, then trying to replicate Stratasys' Polyjet, continuous fibre reinforcement from Desktop Metal and CF3D, Nano Dimension and now my eyes are on Voltera.

As you can see, the gels are pretty distant from copper (and the other metals, seen below).

For a 1.6mm, 4 layer PCB, there does seem to be enough room to make any 1oz traces (35um) into thicker ones like 80um or 160um, which should help with resistance issues.4 Layer PCB Board Bulit On Rogers 12mil RO4003C and FR-4(id:10944752). Buy  China RO4003C, High Frequency PCB, Digital Radio Antenn - EC21
.16+.28+.16+.12+.16+.28+.16 = 1.32mm
.16+.32+.16+.12+.16+.32+.16 = 1.40mm

 That would leave enough for a 120um thick soldermask each side to acheive a 1.64mm PCB.

Anyway, it seems they're using "silver electrically conductive adhesive" and I was able to dig around and find something:

While technically expensive (£7,866/L), it sounds like the stuff in the paper by the description:
Is it me, or is that resistance really low? Perhaps there's an extra 0?
Nope.

Why is the resistivity 1/6 of copper? I don't recall silver being that much better than copper in the electrical resistivity table I found a day or two ago.

[26th Aug] It's because I didn't notice that the pastes are in ohms/sq not milliohms/sq. The first paste is 20mOhm/sq and the second is 10-50.

Discussions

Daniel Grace wrote 07/30/2022 at 03:08 point

Little known secret of research papers:

Most universities have an agreement with publications that encourage students to publish papers for credit. Those publications have contracts with anyone submitting a paper that restrict them from doing certain things with the papers. That includes openly publishing it for free or pay.

But there's been a loophole in as long as this has been a thing: they are allowed to share their paper one-on-one for free. The vast majority of people who have published papers will gladly email you a copy if you ask. They see ZERO money from anyone buying their paper through the publication and do not give two craps about the papers profits.

Yes, it's a hassle to have to email the authors, but I bet you could get any paper you wanted.

  Are you sure? yes | no

kelvinA wrote 07/30/2022 at 03:23 point

At least now I know.

  Are you sure? yes | no

kelvinA wrote 07/29/2022 at 22:01 point

As I started to suspect, it's been done with FDM: https://hackaday.com/2018/04/13/hybrid-3d-printer-creates-complete-circuits-case-and-all/

Is this like the 5 axis printer  made by a uni student where it never got picked up by anyone else in the community and was slowly forgotten? Or perhaps it was because the feature sizes were still too large -- a 0.2mm nozzle would only be able to make 0.4mm lines before Cura 5.0.

  Are you sure? yes | no