Does anyone have any advice on routing some high speed signals on flex PCBs? I'm looking at jlcpcb because of the low cost but I'm having difficulty getting impedances to be decent.
My requirements are:
- 45 ohm single ended impedance (maybe? Def doable)
- 90 ohm differential impedance (usb spec)
- 5A current on power pins
But if we look at the capabilities: https://jlcpcb.com/capabilities/flex-pcb-capabilities
- 2 layer
- substrate thickness (PI) = 25 um
- 1 oz pour thickness = 35 um
- min trace width/spacing = 4/4 mil = 0.101/0.101 mm
- ε = 3.3
Is this feasible with this stackup? I'd like to do a 1 oz pour because of power traces, but there's also 0.5 oz (18um) and 0.33 oz (12um).
For the differential signals, when I'm doing impedance calculations, I can get to roughly 70 ohms using W=100um, S=200um. I don't think this is good enough. I think I can get away using a 0.33 oz pour but then I'm worried about the power pins.
And for the power traces, I'm needing 2.2mm, which is reasonable for the pins on a USB-C connector. But if I try using the 0.5 oz or 0.33 oz pour, it gets to be 4.2 mm and 6.3 mm, which seems impossible given the pins are tiny and very closely spaced together. Even with vias to the bottom layer, this seems problematic.
Anyone have any advice here? This is just for a hobby project, so I'm really not looking to change fabs because of costs.
do you have any control over the thickness of the substrate? see this question here:
https://electronics.stackexchange.com/questions/663304/flex-pcb-controlled-impedance
it looks like changing the substrate would be easiest from a design standpoint but there are ways around using a thin core. you could try ditching the ground plane on the other side and using a single plane coplanar wave guide.
i got 90Ω with width 55 and ground plane spacing 23 https://www.microwaves101.com/calculators/864-coplanar-waveguide-calculator
about the power traces: can you not use a slightly wider trace on the run and then match to the shape of the connector with a gradual taper near the end?
I can't change the thickness. I might try cross hatching the ground plane (suggested in the SE) and seeing how that affects the impedance as well.