• Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

signal reflection

Status
Not open for further replies.

leedude

New Member
hi, i have some questions:

Is this correct:
signal reflection can be caused by physical imperfections in a wire inducing current in the opposite direction, thus causing interference and signal weakening.

i would also like to know how adding a resistor to ground can get rid of signal reflection.

thanks in advance for any replies.
 
Last edited:

crutschow

Well-Known Member
Most Helpful Member
Signal reflections are a high frequency effect caused by variations in the characteristic impedance of a line or a difference in the characteristic impedance of a line and the line terminating resistance. All lines have a characteristic impedance based upon the line capacitance and inductance (equal to the square root of the inductance divided by the capacitance). For high frequency lines this is carefully controlled (such as in RF coax) so the impedance is constant and doesn't cause internal reflections. These reflections are only important when the frequency of a sinewave has a wavelength that is comparable to the wire length, or the risetime of a digital signal is comparable to the propagation delay of the wire length.

If a line isn't terminated by a resistance equal to it's characteristic impedance then there will be reflection back from the line's end. Thus a resistor to ground (if it's the correct value) prevents the reflection.

For more info Google "characteristic impedance"
 

leedude

New Member
so, what about square-waves operating at say, 192khz.
the wavelength would be 1561 meters.
if the wire was 10m long, is there any chance of reflection then?
 
Last edited:

Mikebits

Well-Known Member
so, what about square-waves operating at say, 192khz.
the wavelength would be 1561 meters.
if the wire was 10m long, is there any chance of reflection then?
A squarewave is made of multiples of the fundamental frequency. In other words, there are many high frequency components in a square wave that are much higher than the fundamental frequency of 192KHz. 10m is probably too long to drive on a single wire. You will probably need some sort of line driver on twisted pair.
 
Last edited:

leedude

New Member
ok, so what is the difference between a line driver and the TX pin of my UART?
i would imagine that it has a higher current sourcing capability.

would this be correct:
the more current flowing through the signal/gnd pair, the higher the immunity to noise.
a termination resistor is added to increase current flow, so that a more powerful external influence(EM field) is required to induce a noticeable current and disturb the data stream.

what i need is to send data from my computer down a cable to about 5 receiving endpoints.
the data will be at 19.2kbaud.
so i would need a line driver on the output of my computer's UART, a twisted pair cable, and an inverting buffer on each endpoint receiver connected to it's MCU.
should i also put termination resistors at each endpoint?
if my idea of termination resistors is correct, this should reduce noise.
 

leedude

New Member
ok, ive read this:
http://www.allaboutcircuits.com/vol_2/chpt_14/1.html

and i understand this whole reflection thing now.
for the line driver, would a max232 IC do? bearing in mind my following plan to have only 1 unit driven my the computer.

it seems like the best thing to do would be a daisy-chain, with a max232 on each receiver, and a TX and RX socket, with the MCU retransmitting each received byte to the next node:


PC ---> UNIT1 ---> UNIT2 ---> UNIT3

I would still use a twisted pair cable.


now i just gotta think of a connector to use for the link.
 

blueroomelectronics

Well-Known Member
This would have been simple if you just explained what you're trying to do in the first place. RS232 is short range and realistically limited to two devices. Rs485 or RS422 are long range multidrop devices that simply need to be terminated at the endpoints.
 

leedude

New Member
hmm, yes it seems like i am shooting myself in the foot with rs232.
i think ill go with rs485.

from what ive read, termination resistors only need to be placed at the ends of the bus not every single node.

so, what is the best way of making connections to the bus from each node without causing reflections, what kind of connector would be good?
also, being as rs485 is differential, do the devices need to share a common ground?
 
Last edited:

blueroomelectronics

Well-Known Member
Use the same 3 wires you planned to use for RS232. A B GND. Some folks use 100ohm resistors between the devices and the cable ground.
 
Status
Not open for further replies.

Latest threads

EE World Online Articles

Loading
Top