GARTHWILSON wrote:
The problem with reflections and the need for termination is associated with rise and fall times, not particularly frequency. The faster the t(r), the shorter the maximum transmission line length needs to be to avoid problems. Dr. Howard Johnson has a lot of articles on it linked
here. I was going to list the most relevant ones separately, but there are too many. He is quite an industry guru on this stuff.
Wow, what an amount of stuff to read.
I was also thinking about the rise and fall times as a particular reason for the noise, as sometimes using a ALS type made it worse compared to an LS type.
I know very well that my stuff is horribly outdated :-) But I actually started that whole thing in 1989, so back then, it wasn't really outdated ;-) Then I jumped on and off that project over the years...
But with the dying of 5V parts, especially the higher integrated ones, I feel I need to do a big cut very soon anyway. I have three projects still in the pipeline (Ethernet, I hope to get that done these days, board's done, working on the driver, a block transfer engine, and USB, but I may postpone the latter until after the switch). Until then I hope I have my 65k CPU ironed out into a working state, so then I'm going to switch to an all-new design, FPGA, high speed, 3.3V, maybe even lower, and so on. No details defined yet though.
For me this system is more about self-education and the love of tinkering with stuff. So now I'm going to educate myself about bus termination ;-)
Thanks for your comments
André