How much time did you take to learn high speed design?

[colhany]

Hello everyone I'm an ECE student persuing a career in hardware engineering and PCB design. Sometimes I get frustrated when I see high-speed board designs and schematics, and I ask myself, will I ever be able to do this?
I've designed various PCBs up to 4 layer, but with no regard to signal integrity, except for maybe the spacing between signals. No high speed design techniques, no impedance matching, nothing more.
So a question for every high speed designer, how much did you take to reach this level? And if possible, how did you go from regular designs to high speed?
And is it possible for anyone to design high speed boards, or is it a must that you have further academic education (masters and PhD) to do it?
Thank you all.

[qdrives]

Not really an answer to your question, but I worked in electronics development for 20 years before doing a bit of high speed.
Knowing there is a thing called SI, PI and high speed is already half you need as it allows you to search for the answers.

[robertferanec]

- when designing a board for a proper chip manufacturer, they will provide design guidelines which will help you (e.g. search for iMX6 design guidelines as an example)
- I keep learning new stuff every day. Initially you work based on rules, later you start thinking about how it probably works and what is happening, then you get some confirmation about what you think is right / wrong ... and this goes and goes and goes .... I don't think there is a moment when you can tell yourself "I know everything about high speed".
- take a project which requires to use a CPU
- you can do some high speed designs without a special education based on rules, but if you really would like to understand what is happening, to be able to understand that you need to have some background (and to understand the background may be tricky without university education because of the physics and math there)

[colhany]

Thank you everyone

[miner_tom]

Like Robert says, there really is not a direct answer to your question. Your real teacher is experience. For example, if you have to have a PCB pass FCC or (other international standards) for radiated emissions, you will inevitably encounter a radiation lab. That is where the rubber meets the road. That is where you will see the results of poor signal integrity in terms of radiated emissions. That will teach you a lot.

And, BTW, there are ways to measure radiated emissions without a lab (using a spectrum analyzer or a scope probe) but those ways don't tell you if the emissions are a problem. In truth, everything radiates somewhat. What matters is what the FCC says matters. Does it pass class B? That is what is really important.

Tom