Greetings,
I've been reading everything I can about decoupling, noise reduction etc. There were a few things I wasnt aware of like just how much a small distance from an IC can reduce the effectiveness of a bypass cap. When it got to combinations and different materials affecting the frequency response thats when the smoke started coming out of my ears.
So, I'd be grateful for any advice on what would at least be a solid strategy to use. I'm designing a Z80 based platform for experimental interfacing with a variety of different things so power requirements are expected to vary wildly. It will be clocked from a 3.6MHz crystal. So far, I've been working on a simple ROM programmer to give it its initial set of instructions (this will include interface code for an SD card which will become the primary software source). I've mentioned a bit about the ROM programmer in circuit help.
One thing to note is that the thing will be lit up like a christmas tree. It's for very low level experiments so everything including the address and data busses etc. will be displayed.
Question: If using capacitors a decade apart flattens the frequency response, how far should I go with this? If I put 100nF's across each IC, some 1uF's nearby, a few 10uF's and maybe a 100uF or 2 in the vicinity, would that be good? Similarly with the power, if I use a 4700uF at the supply, and a 470 plus a few decade steps down on the regulator, is that good?
I've been reading everything I can about decoupling, noise reduction etc. There were a few things I wasnt aware of like just how much a small distance from an IC can reduce the effectiveness of a bypass cap. When it got to combinations and different materials affecting the frequency response thats when the smoke started coming out of my ears.
So, I'd be grateful for any advice on what would at least be a solid strategy to use. I'm designing a Z80 based platform for experimental interfacing with a variety of different things so power requirements are expected to vary wildly. It will be clocked from a 3.6MHz crystal. So far, I've been working on a simple ROM programmer to give it its initial set of instructions (this will include interface code for an SD card which will become the primary software source). I've mentioned a bit about the ROM programmer in circuit help.
One thing to note is that the thing will be lit up like a christmas tree. It's for very low level experiments so everything including the address and data busses etc. will be displayed.
Question: If using capacitors a decade apart flattens the frequency response, how far should I go with this? If I put 100nF's across each IC, some 1uF's nearby, a few 10uF's and maybe a 100uF or 2 in the vicinity, would that be good? Similarly with the power, if I use a 4700uF at the supply, and a 470 plus a few decade steps down on the regulator, is that good?