| ??? 08/31/05 09:50 Modified: 08/31/05 10:04 Read: times |
#100328 - Huh Huh? Responding to: ???'s previous message |
Erik, in your former post you wrote:
a, in my opinion, much better way is to put small ferrites in series with the wires where they exit the board. I run ribboncables up to 5 ft driven by 74h connected that way and they cross 2 50 A switcing power supplies. Then you wrote: I have cables from 3" TO 6' and am running pulses at 2MHz. The beauty of using ferrites is that they "round" the pulses a bit and dramatically cut down on EMI and reflections. Erik, you didn't tell us in your former post, that you allow a drastical signal bandwidth limiting, when using ferrites. This changes everything! If you can allow a drastical signal bandwidth limiting then of course you can avoid series termination, to some degree. But take care, these ferrites form a low pass filter in combination with cable capacitance, otherwise you wouldn't observe a rounding of edges. Corner frequency of this filter is about f = 1 / (2 x pi x SQRT(L x C)) = 1 / (2 x pi x SQRT(1µH x 1.8 x 67pF)) = 14MHz if you use a 6' (1.8m) long ribbon cable showing a capacitance of about 67pF/m and if you use a standard SMD ferrite bead. But, if you now increase the length of your cable, then corner frequency decreases and transmission of your 2MHz signals becomes more and more problematic. Also, if you have some skew sensible signals to transmit, which are susceptible against relative timing errors, then unavoidable tolarances of ferrites and imperfections of ribbon cable can heavily erode signal integrity. (Think about, what happens if noise is superimposed on the smoothed edges. And noise is more probable in your system, because you have increased the source impedance of line by inserting the ferrites!!) So, when thinking about the use of ribbon cable for transmitting TTL signals, when there is NO short cable used and when NO relevant signal bandwidth limiting is allowed, then your solution with the ferrites will terribly fail!! Only series termination is practicable then! Kai |
| Topic | Author | Date |
| driving ribboncabel - ESD precautions | 01/01/70 00:00 | |
| on interconnect | 01/01/70 00:00 | |
| Alternative | 01/01/70 00:00 | |
| comments. | 01/01/70 00:00 | |
| Recommended scheme | 01/01/70 00:00 | |
| 74H? | 01/01/70 00:00 | |
| nope | 01/01/70 00:00 | |
| Don't forget series termination | 01/01/70 00:00 | |
| series termination | 01/01/70 00:00 | |
| re: series termination | 01/01/70 00:00 | |
| an, in my opinion, much better way is to | 01/01/70 00:00 | |
| Series termination is the correct method | 01/01/70 00:00 | |
| both? | 01/01/70 00:00 | |
| Series termination prevents overshot | 01/01/70 00:00 | |
| HUH? | 01/01/70 00:00 | |
| Huh Huh? | 01/01/70 00:00 | |
| read the original post | 01/01/70 00:00 | |
| Think about the newbies | 01/01/70 00:00 | |
| That is exacltly why I always push the u | 01/01/70 00:00 | |
| RS485 is a good alternative, indeed | 01/01/70 00:00 | |
| Use a relay driver and sleep well ... | 01/01/70 00:00 | |
ferite bead, series termination, or | 01/01/70 00:00 |