0
\$\begingroup\$

I am trying to design a rigid-flex PCB that currently requires some features that appear to be pushing the limits of manufacturability. These features include:

  • 22 Layer Rigid Section
  • 4 Layer Flex Section (1/2 ounce copper)
  • Blind Vias (1-14 and 14-22)
  • IPC 6013 Class 3 (high reliability)
  • Polyimide rigid section (needs high temperature stability)
  • 4 mil trace spacing in some places on the rigid section
  • Via-in-pad

I worked closely with three board houses to do Design for Manufacturability (DFM) on this design and they had some minor suggestions that I implemented, but none of them conveyed any concerns about being capable of building it. However, all 3 board houses failed to manufacture a single board that passes cross-section testing. They all came back and told me where it failed and that they no longer believe it's manufacturable. After 4 failed attempts at 3 different places, I believe them now. Unfortunately when I ask what we need to change to make it manufacturable they beat around the bush or ignore me (probably because they don't want to waste any more time and money on something they weren't able to build...?). Note that 2 of these board houses specifically specialize in rigid-flex boards.

The houses failed in the following places during manufacture:

  • House 1: Registration (they tried and failed here twice)
  • House 2: Completed build but failed cross-section testing (hole fill depression and wrap distance)
  • House 3: Over-etch

I am able to make some concessions in the design to remove some of these features if needed. I just need to know what combinations of them are making this board so difficult to build. I want to ask this community what combinations of the features listed above seem to be red flags that would make the board much more manageable without having to get rid of everything. I really need a 4 layer flex with 1/2 ounce copper, 4 mil trace spacing, and via-in-pads, but I may be able to reduce the rigid layer count, remove blind vias, change the rigid material, and/or move down to class 2 if needed.

Unfortunately I can't show design files, so I know any advice here will have to be general and the issue may actually be unrelated to the features I described above. I'd appreciate any general advice however on rigid-flex design for manufacturability, because the board houses I've worked with have not been as helpful as I would have liked.

\$\endgroup\$
4
  • \$\begingroup\$ Maybe you ought to get test boards made with test patterns and layouts that are easily testable so you can actually figure out which features are failing. \$\endgroup\$
    – DKNguyen
    Jul 20 at 19:55
  • \$\begingroup\$ You could try ripping apart layers and destructively analyzing the board to see if anything stands out. \$\endgroup\$
    – Voltage Spike
    Jul 20 at 19:59
  • \$\begingroup\$ I would look at reducing the high layer count and board thickness (assuming 22 layers is making ~0.125" thickness) \$\endgroup\$
    – Aaron
    Jul 20 at 21:54
  • \$\begingroup\$ I'm no PCB expert but aren't registration and over-etching per-layer issues? as in they are problems in the part of the process where each layer is manufactured separately? and therefore they would still occur even on boards with few layers? \$\endgroup\$
    – user253751
    Jul 20 at 23:47

0

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.