[m55au]

> the design of a modern motherboard, which is highly complex

The design of a modern motherboard is far from complex since all the complexity has pretty much been hidden inside the black boxes of highly integrated circuits. You as a designer do not in general need to concern yourself with the implementation details and just need to understand the interfaces, similar to, for example, using python's numpy as opposed to writing BLAS or LAPACK from scratch.

[foldr]

What knowledge and experience do you have of designing motherboards or other similarly complex PCBs?

Getting GHz frequency circuits to work requires much more than just understanding the digital interfaces.

If you think designing and laying out a modern motherboard is “easy”, then there’s a 1% chance that you’re a highly experienced and capable electronic engineer, and a 99% chance that you just have no idea about the existence of entire classes of problems that need to be solved to make a motherboard work.

[m55au]

Would not call myself an experienced electronics engineer (have only gotten into RF lately, but yes I know about signal integrity and EMI problems), but getting GHz circuits to work on a modern motherboard is not as big of a deal as you might think mainly because it is a "solved problem" with best practices and design tools, and there are just not really that many things to connect together.

I am not saying it is trivial or that Rossmann could do it with his knowledge (no, he could not) or that anyone could do it. I just claimed that in the grand scheme of things it is not "highly complex" (as opposed to pretty much anything analog, or proper RF, or getting the actual implementation inside the chips correct).

[foldr]

It seems that we agree then. Rossman is operating well outside his knowledge and expertise when critiquing the design of a motherboard. As far as I know the guy has never so much as blinked an LED with an Arduino, or designed a simple two-sided low frequency PCB. A motherboard is way more complex than anything a hobbyist electronics enthusiast will ever work on.

If you haven’t actually been involved in designing a motherboard yourself, I’m also inclined to believe that there might well be complexities that you’re unaware of. Power management is complex. Routing hundreds of length-matched high frequency traces is complex. Low power circuits are complex. Safely charging LiPo batteries is complex. If you can show me something similarly complex that you’ve done yourself, then I’ll be willing to believe your claim that all of this is not really very difficult.

[m55au]

I think it depends on what he is specifically criticizing.

I have not designed a motherboard, but plenty of smaller digital IC-based schematics and PCBs and in my experience the most problematic part has been either erroneous or lack of proper documentation or bugs in ICs which there are plenty of.

What I recommend you to do is to compare the schematics of, for example, Thinkpad T420 and Keithley 2001 both of which should be available with a search. What you will find is that the former is a collection of specialized chips (including one for battery charging) with mostly datasheet reference design based implementations + bunch of mosfets for enabling/disabling/routing signals, whereas the latter has almost none of that and plenty of analog "cleverness" and "raw" digital design, let alone the power input, which in my opinion was totally overdesigned. Or take any oscilloscope, where you need to design not only a computer but also an analog front-end + high frequency signal processing, not only routing. And I'm not even going into the actual RF board designs which you are probably well aware of is more physics than electronics.

Again I am not saying that it is easy for any guy on the street, but "relatively easy" to understand what is going on on the motherboard as opposed to the actual complex stuff. So if you see a common fault among many motherboards you can probably conclude that that specific area of the design was bad and criticize that. Whether the criticism is valid might be another question, but I would not automatically dismiss it.

Edit: I will put designing a motherboard on my todo list.