-
Categories
-
Platforms
-
Content
- Status
Bad programming. It's possible to have a line of logic that is scanned by a cpu later than the other portions of logic (example: subroutines). By the time one portion of logic executes, the other conditions have not been updated.
power supply problems. (I believe in testing the supply before anything else.)
hevans1944
Hop - AC8NS
Biggest source of glitches is inadequately by-passed power supply rails. Each time a high-speed logic device changes its output state, a current pulse occurs at the power supply input to the device. This pulse can then propagate as a "glitch" over the power supply rails to other devices. The solution is to always bypass the power pin on EACH device with a capacitor, typically 0.1 μF stacked ceramic type, connected as close to the device as physically possible between the power supply rail and the common rail. Back when logic ICs were sold in DIP packages, you could purchase sockets with this capacitor built into the socket. Good PCB layout is also important, especially when using the faster logic families with clocks approaching one gigahertz. Having power and signals on different layers of the PCB is virtually mandatory.
Glitches in a production circuit design are an indication of either poor engineering or defective components. I would suspect both, given the current explosion of products originating from the Pacific Rim nations, and the fact that a new generation of engineers needs to re-learn techniques from the previous century while discovering new techniques apropos to the technology of this century. As things get smaller and faster the problem of "glitches" and how to eliminate them becomes much more difficult to solve. And the test equipment you need to even "see" a pico-second "glitch" is very expensive. <sigh> Sorta glad I am retired and will avoid such trials and tribulations.
Glitches in a production circuit design are an indication of either poor engineering or defective components. I would suspect both, given the current explosion of products originating from the Pacific Rim nations, and the fact that a new generation of engineers needs to re-learn techniques from the previous century while discovering new techniques apropos to the technology of this century. As things get smaller and faster the problem of "glitches" and how to eliminate them becomes much more difficult to solve. And the test equipment you need to even "see" a pico-second "glitch" is very expensive. <sigh> Sorta glad I am retired and will avoid such trials and tribulations.
Why do they use ceramic instead of electrolytics or tantalum?
I have seen decoupling caps as electrolytics and sometimes i see them as tantalum and other times i see the molded ceramic caps. I'm not sure when to use electrolytic or tantalum decoupling caps.
But why Stacked ceramic? do you mean the multi-layer ceramic
davenn
Moderator
They tend to be more stable and they also have a smaller footprint on the circuit board for a given value
tend to be used for the larger values and the initial brute force smoothing. Then after that, smaller value ceramics will be used
yes
Dave
davenn
Moderator
do I really have to answer that ?
what do you really think a higher temp rating means ?
- Status
