Asymmetric stackup for PCB motor

[charliehorse55]

I'm working on a PCB motor, the board needs to be thicker for mechanical strength (the board is 10"x12"), but I don't want the coil layers getting too far away from the working surface, as it decreases the strength of the magnetic fields.

I have an 8 layer board, with 5 layers of coils on inner layers (my last prototype put coils on the surface, but the plating process increases the thickness of the copper and causes that phase to have a lower resistance). All 8 layers are nearly entirely filled with copper. Before hearing about warping, I made a stackup that used minimal thickness on all dielectrics, as well as minimally thick cores for the first 2 of 3 cores. This gives minimal thickness between the coil layers and the top surface:

(using PCBWay's available materials):

Top 1oz copper
3.5mil dielectric
Inner 1oz copper
8 mil core
Inner 1oz copper
3.5mil dielectric
Inner 1oz copper
8 mil core
Inner 1oz copper
3.5mil dielectric
Inner 1oz copper
80mil core
Inner 1oz copper
3.5mil dielectric
Bottom 1oz copper

For a total thickness ~3.0mm

PCBWay has said of this stackup that "this will cause problems with warping", but I believe they're still offering to build it, so it can't be too bad?

* Am I likely to run into issues with this?
* Is there a better compromise to be had?
* Making the board thicker should help reduce warping, even if it's mainly in a core that's off-center, right?
* Would increasing the last dielectric thickness from 3.5mil to 7.5mil help much?
* What about changing the copper thickness on the bottom? The thicker copper could balance out the larger number of layers that are above the big core

[T3sl4co1l]

They'll make it, sure... who are they to say you're wrong, it's your money

How about a thin board with all thin layers (probably one or a couple will still have to be thicker cores), stacked on top of the main board?

Tim

[Berni]

Yep PCBs can indeed warp when one side has much more copper than the other (You basically create a bimetalic strip)

If this warping is a problem depends on a lot of factors. Like how big the copper imbalance is, how large the PCB is, how much warp you can tolerate, if you have large SMD components in the problematic areas..etc

But in general PCB motors are in most cases a bad idea since they need outside help to guide magnetic fields to where they are supposed to go and they don't tend to offer a lot of copper cross section for the coils as most of the space is taken up by FR4.

[Damperhead]

I don't think the situation is too bad, more of a compromise. I have seen many kinds of asymmetric structures, with and without hybrid material. They are mass-produced goods and yet they can be manufactured.
I would be more worried if sequential lamination was used asymmetrically in the structure. Generally, such structures are challenging for bow and twist.

However, I recommend using metric units, or alternatively imperial units, but don't mix them.

[charliehorse55]

They'll make it, sure... who are they to say you're wrong, it's your money

How about a thin board with all thin layers (probably one or a couple will still have to be thicker cores), stacked on top of the main board?

Tim

There are a ton of connections between all 8 layers, so it's all got to be on one board. Luckily, the enclosure is plenty rigid, so I could increase the number of mounting points, and use a PCB that will be quite floppy on it's own, but rigid once mounted.

I wonder how thin a 10"x12" board can be before it causes problems for fabrication/assembly?

[charliehorse55]

I don't think the situation is too bad, more of a compromise. I have seen many kinds of asymmetric structures, with and without hybrid material. They are mass-produced goods and yet they can be manufactured.
I would be more worried if sequential lamination was used asymmetrically in the structure. Generally, such structures are challenging for bow and twist.

Maybe it's fine as is? I'm ordering 10 boards, 5 bare, 5 assembled. I guess they can pick the flattest 5 boards to assemble. If something goes wrong and I only get 3-4 functional boards, that's fine. It'd be a nice learning experience. Just trying to avoid some disaster where the warping is so bad they have to start over on the PCBs with a different stackup.

I believe sequential lamination is an HDI thing? This board is pretty standard, 5/5mil and only through vias.

[Infraviolet]

"needs to be thicker for mechanical strength"
Does it? My suspicion is the torques generated by a PCB motor will be pretty low, if you want larger torques for use you'd probably want reduction gearing after it. With this in mind, it might well be the case that the PCB  motor suffers a lot less force than you expect.