I do not like riving knifes that are a lot, like .020, thinner than the blade because they will not stop the kerf from closing on the blade. When that happens, it is possible to put a wedge in the kerf but I prefer the riving knife give me a chance to complete the cut without that aggravation. They still will give reasonable protection against kick back, I think. But with a riving knife closer to the blade thickness you get both things - but you have to be really careful to get it centered on the blade.
I made a thicker riving knife for my old BT3100 and it worked great. I used 7075 T6 aluminum plate from McMaster Carr. This grade of aluminum is relatively hard and does not bend. It will fracture if you try to bend it without heating it. It is easier to cut than steel. But I find my Bosch jig saw works well to cut mild steel or aluminum. If I needed to put a bend in the material, as I believe my SS requires, I would use mild steel.
My PCS is at least my fourth 120V table saw and the first to give me a very hard time with full kerf blades. Part of the trouble was the fact I was using Infinity blades. First a 50 tooth with 40 ATB teeth and 10 rakers. Once the trouble started I switch to an Infinity 24 tooth ripping blade. Neither would rip the glued up 2x4 material I used for legs on my grandkids bunk bed without the saw hitting the thermal overload. Then I tried a 24 tooth Freud blade which was not new but was relatively clean. It worked OK but required a slower feed rate. Then I dug out a 24 tooth thin kerf blade (Freud) and could go back to normal feed rate. Since then I have ripped similar glueups with a thin kerf Fusion all around blade (40 teeth). It worked well. But the wood was not warping into the blade like the bunk bed legs seemed to. My conclusion is to retire the nearly new Infinity blades and to use Freud thin kerf blades in this saw. With them, I have had no issues. Including in 3 inch thick cherry glueups. I am using the 2.3mm riving knife the saw came with, however, which is only very slightly thinner than the 3/32 blades I am using. That helps to avoid one of the power robbing issues of the kerf closing on the blade.
I think the thermal overload on my PCS may be set somewhat conservatively. But I am not going to mess with it if I can solve the issue with a thin kerf blade. It is, for me, an expensive tool and I would rather it be well protected.