We all could give you rules of thumb but you need to add simulation to your toolbox. All these things you mention vary depending on stackup and the only way you will ever know is if you simulate your design with your stackup.
It would be great to have a SI tool to analyze a board layout, and the time to learn how to use it. But I do not at this time. I'm asking for some experienced eyes that could say, "Oh no! Don't do that!" or, "Sure, I've made accordions like that".
Perhaps you can tell me what effect changing the amplitude & gap (frequency) can have on a signal. (I.e. increasing the amplitude of an accordion does X to the signal, and decreasing the gap of an accordion does Y.). A hint like "too small a gap and you'll generate a lot noise" would suffice. I would guess that the effect of a small gap is more critical than that of a larger amplitude, but I really have nothing to base that on other than the "no sharp corners" rule mentioned above.
With my stackup and trace widths I can get the suggested Z0, I just need to adjust trace lengths to within suggested tolerances. If using "bad" accordions is worse that having mismatched trace lengths, that would be helpful information. In all the literature I've read, there is no mention of how (or how not to) match trace lengths.
1 of 1 people found this helpful
It depends on lots, there is no rule which be safe enough to garantee your board works well whatever your design is. In Hyperlynx 9.0, you can simulate various accordion types, do comaprison to decide which type is the best for your design if you spend two hours to learn this tool. If you like a thumb rule, just keep Gap >= 5h (h: distance to reference layer) .
Thank you. I used your 5h "rule" and we'll see how it goes when the prototypes come back!
I am happy to report that I received my prototype boards back and the layout was a success!
It passed a barrage of RAM tests and I have seen no problems with the flash.