0
\$\begingroup\$

I'm looking at prototyping a PCB that needs to make many connections over a curved surface. To develop this, I'm looking at using thin/flexible PCBs. For a 2-layer design, I'm comparing a thin (0.2mm-thick) FR4 PCB to a standard Polyimide flex PCB. The 0.2mm FR4 seems like it would be a suitable, less expensive option ($104) versus a typical Polyimide flex circuit ($189). Can I just use really thin FR4 instead of polyimide flex to save a few bucks?

In an effort towards full disclosure, here's what I think matters about the design:

  • The PCB bend is a relatively large radius, say greater than 100mm / 4in.
  • The routing isn't particularly demanding in terms of space/trace
  • The PCB will contain SMT parts populated before bending
  • The PCB is soldered to panel-mount, through-hole components
  • Once the PCB is soldered, no further bend/flex of the PCB is intended
  • Once the PCB is bent in place and solder
\$\endgroup\$
6
  • \$\begingroup\$ Can you just buy a sample of 0.2-mm G10 and try it? \$\endgroup\$
    – The Photon
    Commented Jun 27, 2020 at 6:15
  • \$\begingroup\$ "The PCB will contain SMT parts populated before bending" - really bad idea. Why pinch pennies when it will seriously compromise reliability? \$\endgroup\$ Commented Jun 27, 2020 at 6:29
  • \$\begingroup\$ yep, if you're willing to spend $100, try one of the chinese (JLCPCB) or US-based (oshpark) suppliers of flex pcb; a lower-quality flex still sounds wiser than a high-quality rigid board if you need flex. \$\endgroup\$ Commented Jun 27, 2020 at 9:57
  • \$\begingroup\$ but: If you go the FR4 route, why not cut slots in it so that you know more or less where it bends, and can put the components where it doesn't? \$\endgroup\$ Commented Jun 27, 2020 at 10:00
  • \$\begingroup\$ @BruceAbbott Thanks for the feedback. I think I can limit the flexing where the components are, but with a FR4 board I won't have an explicit stiffener. \$\endgroup\$
    – W5VO
    Commented Jun 27, 2020 at 13:38

1 Answer 1

3
\$\begingroup\$

The base material can handle that amount of bending,

Just testing on hand with a 1.6mm thick 20x300 mm PCB I have on hand it can be bent to roughly a 0.4m radius before it starts strongly resisting, though I will note it is definitely into the plastic deformation levels of stress to do this. as it does not return to being perfectly flat.

0.7m radius appears to be the limit of the elastic range for my example PCB.

so for a 0.2mm thick pcb, I would certainly say you could even still be in the elastic deformation limits for a 100mm radius

For routing, avoid any sharp corners, your putting things under stress,

Components being fitted before bending could be an issue, ceramic capacitors really do not like being flexed, make sure you only use soft termination smd capacitors for this kind of project,

Make sure as much as possible any SMD components are perpendicular to the bend, the less width they are bent over, the lower the forces,

Use leaded SMD packages for IC's where possible, e.g. SOIC, or even SOT, QFN and BGA would risk lifting pads on that kind of bend.

So yes its possible, and may even have more desirable properties, will just cost you some time to make sure the components can survive it.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.