Author Topic: RS485 unterminated  (Read 3446 times)

0 Members and 1 Guest are viewing this topic.

Offline rakeshm55Topic starter

  • Regular Contributor
  • *
  • Posts: 209
RS485 unterminated
« on: July 05, 2018, 02:02:26 pm »
Hi,
For a Board to Board communication UART communication , I plan to use a RS485 signal level Drivers. The trace length will not be more than 50cm.
RS485 is expected to operate inpoint to point mode.
UART speed is expected to be of the order of 12Mbps and Industrial grade temperature range is expected.

In this case inorder to reduce power consumption can I use RS485 in un-terminated mode.

 If the termination is avoid reflections then my bit period is of the order of 50ns, so a 5ns round trip delay can be accommodated. Will there be significant reflections for 50cm trace length??

 If terminated with a 120Ohms resistance the total power consumption is of the order of  200mW(3.3V RS485).


Any recommendations to reduce power for RS485.

 

Offline rstofer

  • Super Contributor
  • ***
  • Posts: 9963
  • Country: us
Re: RS485 unterminated
« Reply #1 on: July 05, 2018, 02:28:54 pm »
At the bit frequency you want, I doubt that the design will work reliably without terminating the lines.  Maybe, maybe not, but more likely not.

So, design the resistors in and leave them out for a trial.  This article does suggest that you would only need to terminate one end in the case of One Transmitter - One Receiver.  Check the signal integrity with a scope and see what you think.

https://www.maximintegrated.com/en/app-notes/index.mvp/id/763

 
The following users thanked this post: rakeshm55

Offline Giaime

  • Regular Contributor
  • *
  • Posts: 72
  • Country: it
Re: RS485 unterminated
« Reply #2 on: July 05, 2018, 02:37:37 pm »
In my experience, we were running 20-30cm of 3Mbps RS-485 in an highly noisy envinronment (6kW AC/AC IGBT converter) and it didn't absolutely work w/o proper termination  :-\

Maybe for DC power consumption you can try AC termination, it's documented in various RS-485 appnotes.
 
The following users thanked this post: rakeshm55

Online mikeselectricstuff

  • Super Contributor
  • ***
  • Posts: 14117
  • Country: gb
    • Mike's Electric Stuff
Re: RS485 unterminated
« Reply #3 on: July 05, 2018, 02:38:36 pm »
I'd be surprised if it didn't work unterminated with only 50cm You may want to bias the lines to avoid glitches at start/end of packet.
if it is terminated, you could reduce power by running the transceiver at reduced voltage, or higher value termination resistor
 


Youtube channel:Taking wierd stuff apart. Very apart.
Mike's Electric Stuff: High voltage, vintage electronics etc.
Day Job: Mostly LEDs
 
The following users thanked this post: rakeshm55

Online mikeselectricstuff

  • Super Contributor
  • ***
  • Posts: 14117
  • Country: gb
    • Mike's Electric Stuff
Re: RS485 unterminated
« Reply #4 on: July 05, 2018, 02:39:47 pm »
Might be worth looking at LVDS as an alternative.
Youtube channel:Taking wierd stuff apart. Very apart.
Mike's Electric Stuff: High voltage, vintage electronics etc.
Day Job: Mostly LEDs
 
The following users thanked this post: rakeshm55

Offline rstofer

  • Super Contributor
  • ***
  • Posts: 9963
  • Country: us
Re: RS485 unterminated
« Reply #5 on: July 05, 2018, 03:12:57 pm »
I haven't thought this thru but don't we shut the transmitter down when not in use?  There is usually an Enable pin.  If so, the resistor won't be dissipating anything when the transmitter is down.

https://www.st.com/content/ccc/resource/technical/document/application_note/52/57/b8/e4/08/38/43/cd/CD00249778.pdf/files/CD00249778.pdf/jcr:content/translations/en.CD00249778.pdf

 
The following users thanked this post: rakeshm55

Offline rakeshm55Topic starter

  • Regular Contributor
  • *
  • Posts: 209
Re: RS485 unterminated
« Reply #6 on: July 05, 2018, 03:36:43 pm »
For lvds drivers do they have fail safe mode?? When open or short at input do they o/p a known state??
Also do they support sleep/low power mode when unused??
 

Offline cowana

  • Frequent Contributor
  • **
  • Posts: 325
  • Country: gb
Re: RS485 unterminated
« Reply #7 on: July 05, 2018, 03:51:45 pm »
It's a valid technique to AC-couple your termination resistor with a reasonably high-value capacitor, so it only draws current during data transmission (and not when the bus is idle).
 
The following users thanked this post: rakeshm55

Online David Hess

  • Super Contributor
  • ***
  • Posts: 17426
  • Country: us
  • DavidH
Re: RS485 unterminated
« Reply #8 on: July 05, 2018, 06:09:36 pm »
Well, let's see ... 50 centimeters at roughly 50 picoseconds per centimeter is 2.5 nanoseconds one way and 5 nanoseconds two ways.  With a 50 nanosecond bit time I think you can do it without termination with appropriate filtering to limit transition time.
 

Offline Daixiwen

  • Frequent Contributor
  • **
  • Posts: 367
  • Country: no
Re: RS485 unterminated
« Reply #9 on: July 06, 2018, 07:46:38 am »
I'd recommend an AC termination too. It works very well.
I'd recommend also to use at least 0603 components so that you can easily adjust the values if necessary.
 

Offline rakeshm55Topic starter

  • Regular Contributor
  • *
  • Posts: 209
Re: RS485 unterminated
« Reply #10 on: July 06, 2018, 07:58:52 am »
I'd recommend an AC termination too.

What will be the value of capacitor to be used for AC termination??

Today I did try using RS485 receiver unterminated for 10Mbps UART and it worked, distance was hardly 4 inches - Room temperature 25C.
 

Offline Daixiwen

  • Frequent Contributor
  • **
  • Posts: 367
  • Country: no
Re: RS485 unterminated
« Reply #11 on: July 06, 2018, 10:50:11 am »
Use a scope to see how much margin you have. You could have more noise in the final application.
If you transceiver datasheet doesn't recommend any value, basically you just need a capacitor that is big enough so that the time constant RC is way over the transition time. But too high a capacitor and your consumption will increase, so experimenting may be necessary to find the correct value.
 

Offline Kjelt

  • Super Contributor
  • ***
  • Posts: 6586
  • Country: nl
Re: RS485 unterminated
« Reply #12 on: July 06, 2018, 11:08:29 am »
RS485 describes a 120 ohm resistor IIANM , what happens if you put in a larger value say 1k or 10k ?
The impedance changes but is it not better to have something than nothing ?
 
The following users thanked this post: rakeshm55

Online mikeselectricstuff

  • Super Contributor
  • ***
  • Posts: 14117
  • Country: gb
    • Mike's Electric Stuff
Re: RS485 unterminated
« Reply #13 on: July 06, 2018, 12:09:19 pm »
At high rates, I'm not sure an AC coupled terminator will save you much - as data rates approach reflection times, the cap is going to be conducting most of the time. Might be tuneable to optimise for a particular baudrate and cable length.
Youtube channel:Taking wierd stuff apart. Very apart.
Mike's Electric Stuff: High voltage, vintage electronics etc.
Day Job: Mostly LEDs
 
The following users thanked this post: rakeshm55

Offline JS

  • Frequent Contributor
  • **
  • Posts: 947
  • Country: ar
Re: RS485 unterminated
« Reply #14 on: July 06, 2018, 01:05:33 pm »
RS485 describes a 120 ohm resistor IIANM , what happens if you put in a larger value say 1k or 10k ?
The impedance changes but is it not better to have something than nothing ?
With a much higher resistance as you said the reflection would still be there, like 90% for the 1k and 99% for the 10k, so not helping much...

  Even if the round trip is much shorter than the bit period, you left the receiver end with 10k, you have a reflection of 99% when it finds the receiver, then, when it finds the transmitter you have a reflection of the same amplitude in the opposite direction and this goes back and starts again. The thing takes a while to extinguish and you might encounter problems. You could try out and find what's going on, probe the line and see what happens at both ends.

  If one is transmitting and one receiving you could increase the termination and the source resistance by adding a series resistor to the source output, if both are transmitting and receiving this won't work as now you have a series resistor to the receiver and signal amplitude would be low.
If I don't know how it works, I prefer not to turn it on.
 
The following users thanked this post: rakeshm55

Offline rakeshm55Topic starter

  • Regular Contributor
  • *
  • Posts: 209
Re: RS485 unterminated
« Reply #15 on: July 06, 2018, 05:15:34 pm »
the source resistance by adding a series resistor to the source output

So if I connect 120 Ohms series resistance at the source the round trip reflection would cease to exist. Will it any way impact my slew rate and data transmission rate say I would like 15Mbps
 

Offline JS

  • Frequent Contributor
  • **
  • Posts: 947
  • Country: ar
Re: RS485 unterminated
« Reply #16 on: July 06, 2018, 06:53:26 pm »
the source resistance by adding a series resistor to the source output

So if I connect 120 Ohms series resistance at the source the round trip reflection would cease to exist. Will it any way impact my slew rate and data transmission rate say I would like 15Mbps
No, let's say the transmitter has an internal 120Ω resistor, you add an external 120Ω in series with it and at the other end now you need a 240Ω resistor instead of the 120Ω. In effect you have 3 times the resistance so 1/3rd of the power.

If you go up and up the capacitance of the bus starts to limit the bandwitdh, how much up could you go depends on that capacitance.

JS
If I don't know how it works, I prefer not to turn it on.
 
The following users thanked this post: rakeshm55

Offline rakeshm55Topic starter

  • Regular Contributor
  • *
  • Posts: 209
Re: RS485 unterminated
« Reply #17 on: July 07, 2018, 07:02:11 am »

Why cant I have a series termination (Source termination) and do away with parallel termination??
Since in my case its a single point to point connection cant I have a series termination??

As shown in attached diagram. Will it not eliminate reflections??

Please suggest
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 17426
  • Country: us
  • DavidH
Re: RS485 unterminated
« Reply #18 on: July 07, 2018, 10:47:19 am »
Why cant I have a series termination (Source termination) and do away with parallel termination??
Since in my case its a single point to point connection cant I have a series termination??

As shown in attached diagram. Will it not eliminate reflections??

Series termination at the source without parallel termination at the load produces a reflection which causes the voltage to initially shift half way.
 
The following users thanked this post: rakeshm55

Offline rakeshm55Topic starter

  • Regular Contributor
  • *
  • Posts: 209
Re: RS485 unterminated
« Reply #19 on: July 07, 2018, 01:26:19 pm »
[quote author=David Hess link=topic=120854.msg1655066#msg1655066 date=1530960439

Series termination at the source without parallel termination at the load produces a reflection which causes the voltage to initially shift half way.
[/quote]
Thanks for pointing the issue....

This half way shift(1.5V) will remain for half of roundtrip time if i am not wrong....Also any voltage above 200mV will be considered as high....On similar lines any diff voltage below -200mV considered as low ... So wont this be ok..or will it be a show stopper....
 

Online David Hess

  • Super Contributor
  • ***
  • Posts: 17426
  • Country: us
  • DavidH
Re: RS485 unterminated
« Reply #20 on: July 07, 2018, 06:43:29 pm »
Series termination at the source without parallel termination at the load produces a reflection which causes the voltage to initially shift half way.

Thanks for pointing the issue....

This half way shift(1.5V) will remain for half of roundtrip time if i am not wrong....Also any voltage above 200mV will be considered as high....On similar lines any diff voltage below -200mV considered as low ... So wont this be ok..or will it be a show stopper....

As long as the input threshold requirements of the differential receiver are met it is not a problem.  But for instance if the source termination resistor values were such that the differential voltage was zero volts during the reflection, the differential receiver would have very low noise immunity.
 
The following users thanked this post: rakeshm55


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf