| ??? 02/20/09 01:48 Modified: 02/20/09 02:08 Read: times |
#162581 - Some hints... Responding to: ???'s previous message |
Mahmood said:
I thought I could get away with removing the terminating resistor for the transmitter, but for the receivers they all should have the 120 Ohms resistor. Only to avoid any misunderstanding: There should only be two such termination resistors. One at the one end (master), the other at the other end (farthermost receiver). The other receivers between must be connected to the bus without any termination resistors! Are you sure that it is actually a 120 Ohm resistor and not a 12 Ohm? Don't laugh... Mahmood said:
+12v power supply feeds the master board through LM7805 regulator. Then same power supply goes with Y and Z lines of transmitter to the receiver board where the +12V power supply goes to LM7805 regulator in the slave circuit. Should there be a seperate GND pin to the Y Z lines or would the power supply return line suffice? You should route the supply-lines (+12V/0V) and Y/Z-lines from the master to receivers via two twisted pairs, which are close to each other: Twist the +12V-line with the 0V-line and the Y-line with the Z-line. Take care that the +12V-supply is properly decoupled at both ends of cable (at each receiver!) to the local signal ground (0V), to which also the transmitted 0V is to be connected. As long as no high currents (no 120Hz ripple currents!) flow along the 0V line you can take the supply 0V as signal ground and can transmit it via the mentioned twisted pair. (From EMI reasons you should have only one ground, anyway.) Mahmood said:
I have another question, I placed the LM7805 on the metal Chassis of the clock case and made wires about 10cm from the regulator to the board. is this ok or would it introduce noise into the circuit and I should make the regulator leads as short as possible? Depends on how noisy the clock's chassis is. In any case you should decouple the regulator pins nearest the terminals. Properly decoupled at the regulator and at the board the 10cm to the board should not be a problem. If you insert a ferrite into the +5V-line to the board, then noise can be properly blocked, assuming that the ground connection is noisefree. (Not exactly knowing how your application actually looks like) best is to have pi-filters for the +5V-connection at both ends, properly terminated to the local ground plane. Kai |
| Topic | Author | Date |
| RS485 problem | 01/01/70 00:00 | |
| Just 1? | 01/01/70 00:00 | |
| The 120 ohm resistor | 01/01/70 00:00 | |
| VCC = 5V | 01/01/70 00:00 | |
| you need | 01/01/70 00:00 | |
| Scope? | 01/01/70 00:00 | |
| daisy chain | 01/01/70 00:00 | |
| Some hints... | 01/01/70 00:00 | |
| 100 meters | 01/01/70 00:00 | |
| 7 clocks | 01/01/70 00:00 | |
| Topology | 01/01/70 00:00 | |
| RE: topology | 01/01/70 00:00 | |
| don't | 01/01/70 00:00 | |
| Yes, don't! | 01/01/70 00:00 | |
| Getting there | 01/01/70 00:00 | |
| So, what was the cause of failure? | 01/01/70 00:00 | |
| not yet | 01/01/70 00:00 | |
| with additional components | 01/01/70 00:00 | |
| Where do you have this circuit from? | 01/01/70 00:00 | |
| from my Mentor | 01/01/70 00:00 | |
| Ben's site | 01/01/70 00:00 | |
| Mission Accomplished | 01/01/70 00:00 | |
| I'm not sure I agree | 01/01/70 00:00 | |
| exactly | 01/01/70 00:00 | |
| Maybe I misunderstood your earlier post | 01/01/70 00:00 | |
no | 01/01/70 00:00 | |
| yes | 01/01/70 00:00 | |
| then how could you ask .. | 01/01/70 00:00 | |
| With noise+etc reduction | 01/01/70 00:00 | |
| email updated | 01/01/70 00:00 | |
| here's info's........ | 01/01/70 00:00 | |
| That means you get | 01/01/70 00:00 | |
| Only when seeing nosie? | 01/01/70 00:00 | |
| Not exactly but as per requirement | 01/01/70 00:00 | |
| CAT5 ethernet cable | 01/01/70 00:00 | |
| 100 Ohm | 01/01/70 00:00 | |
| More details here | 01/01/70 00:00 | |
| did you check the FAQ? | 01/01/70 00:00 |