EmbeddedRelated.com
Forums
Memfault Beyond the Launch

RS485 chips and differential voltage

Started by Dennis Clark March 12, 2004
  I've been reading up on RS485 drivers like the LTC485, 75176B and others.
They look too simple in pin count to supply the + and - voltages needed
for differential drives.  Are there some chips that supply the +/- voltages
that other chips simple scavange and use?  Or do these chips I note actually
supply the +/- voltages for the differential outputs?  The chip specs don't
say.  I've a project that I'm going to propose using an RS485 network but
am not sure of all the pieces needed.  Can anyone supply me with a place
to look for the "ABC's" of RS485 networking to answer questions like:

Speed vs. termination resistors vs. line length.
Realistic upper limits of number of drivers on a bus (I've heard 8)
Comparisons of the driver chips available.

I'd appreciate any and all aid I get.

regards,
DLC
-- 
============================================================================
* Dennis Clark         dlc@frii.com                www.techtoystoday.com   * 
* "Programming and Customizing the OOPic Microcontroller" Mcgraw-Hill 2003 *    
============================================================================
Dennis Clark wrote:

> I've been reading up on RS485 drivers like the LTC485, 75176B and others. > They look too simple in pin count to supply the + and - voltages needed > for differential drives. Are there some chips that supply the +/- voltages > that other chips simple scavange and use? Or do these chips I note actually > supply the +/- voltages for the differential outputs? The chip specs don't > say.
I'm not sure that I'm answering your question, but I'll give it a try anyway. RS-485 does not use plus and minus supplies. It uses +5 volts only, and sends it differentially on the pair of signal wires. In other words, the equivalent of a "high" or "marking" RS-232 signal would have the + RS-485 signal at +5 volts and the - signal at ground. A "low" or "spacing" signal would have the + RS-485 signal at ground and the - signal at +5 volts. I've a project that I'm going to propose using an RS485 network but
> am not sure of all the pieces needed. Can anyone supply me with a place > to look for the "ABC's" of RS485 networking to answer questions like:
This will get you started: http://www.maxim-ic.com/appnotes.cfm/appnote_number/736 The authoritative organization for the standard is the EIA. You can find them at www.eia.org and do an incredible sillywalk to try to find out how to buy a copy.
> Speed vs. termination resistors vs. line length. > Realistic upper limits of number of drivers on a bus (I've heard 8) > Comparisons of the driver chips available. > > I'd appreciate any and all aid I get. > > regards, > DLC
"Dennis Clark" <dlc@io.frii.com> wrote in message
news:405152b2$0$198$75868355@news.frii.net...
> I've been reading up on RS485 drivers like the LTC485, 75176B and
others.
> They look too simple in pin count to supply the + and - voltages needed > for differential drives.
I think you're misunderstanding the term differential. These drivers are powered from 5V single supply. By setting the A line to 5V and the B line to 0V and vice-versa, you get a differential signal between A and B. A receiver only looks at the voltage DIFFERENCE between A and B, the absolute level against ground does not matter. So even when A and B swing between 12 and 13V, it would be a differential system, albeit with a huge DC offset. But this is of no importance for the receiver. There's no negative voltage required in this process.(I think you meant that, didn;t you?).
> Or do these chips I note actually > supply the +/- voltages for the differential outputs? The chip specs
don't
> say.
No they don't.
> I've a project that I'm going to propose using an RS485 network but > am not sure of all the pieces needed. Can anyone supply me with a place > to look for the "ABC's" of RS485 networking to answer questions like: > Speed vs. termination resistors vs. line length.
You can find this sometimes in the datasheets. But from experience I can tell you that 125kbit/s over a 400 meter line in an industrial environment works just fine. I vaguely remember the limit for 9600 baud is far over a kilometer. Termination is purely dependant on the wire impedance and should match it.
> Realistic upper limits of number of drivers on a bus (I've heard 8)
That is in the datasheets. The 'standard' driver (75176) follows the RS-485 spec, which says 32 drivers. But the more modern types from LT,AD and Maxim have a lower load and allow 128 driver on one bus. Meindert
"Dennis Clark" <dlc@io.frii.com> wrote in message
news:405152b2$0$198$75868355@news.frii.net...
> I've been reading up on RS485 drivers like the LTC485, 75176B and
others.
> They look too simple in pin count to supply the + and - voltages needed > for differential drives. Are there some chips that supply the +/-
voltages
> that other chips simple scavange and use? Or do these chips I note
actually
> supply the +/- voltages for the differential outputs? The chip specs
don't
> say. I've a project that I'm going to propose using an RS485 network but > am not sure of all the pieces needed. Can anyone supply me with a place > to look for the "ABC's" of RS485 networking to answer questions like: > > Speed vs. termination resistors vs. line length. > Realistic upper limits of number of drivers on a bus (I've heard 8) > Comparisons of the driver chips available. > > I'd appreciate any and all aid I get. > > regards, > DLC
Hi Dennis, RS485 is 0-5V differential signalling. (A=0-5/ B=5-0) RS485 standard calls for the ability to have 32 nodes in a network (some low power solutions sacrifice this ) IIRC;-) HTH -- ---- Morten Raider of the Lost Electron
Dennis Clark wrote:

> I've been reading up on RS485 drivers like the LTC485, 75176B and others. > They look too simple in pin count to supply the + and - voltages needed > for differential drives. Are there some chips that supply the +/- voltages > that other chips simple scavange and use? Or do these chips I note actually > supply the +/- voltages for the differential outputs? The chip specs don't > say. I've a project that I'm going to propose using an RS485 network but > am not sure of all the pieces needed. Can anyone supply me with a place > to look for the "ABC's" of RS485 networking to answer questions like: > > Speed vs. termination resistors vs. line length. > Realistic upper limits of number of drivers on a bus (I've heard 8) > Comparisons of the driver chips available. > > I'd appreciate any and all aid I get. > > regards, > DLC
Hi Denis, you don't need special voltages for RS485. Take a look at the documents at http://www.bb-elec.com/technical_library.asp?pathid=61 They have good explanations of what RS485 (and '422) means and where the pitfalls are when using it. HTH, Jens
On 2004-03-12, Meindert Sprang <mhsprang@NOcustomSPAMware.nl> wrote:

>> They look too simple in pin count to supply the + and - voltages needed >> for differential drives. > > I think you're misunderstanding the term differential. These > drivers are powered from 5V single supply. By setting the A > line to 5V and the B line to 0V and vice-versa, you get a > differential signal between A and B. A receiver only looks at > the voltage DIFFERENCE between A and B, the absolute level > against ground does not matter. So even when A and B swing > between 12 and 13V, it would be a differential system, albeit > with a huge DC offset. But this is of no importance for the > receiver.
True within limits. The recievers do have limits on the allowed common-mode voltage (the DC offset). A typical spec would be something like -7V to +12V. For such receivers, 13V would be outside the recievers specs, as would -8V. -- Grant Edwards grante Yow! Well, I'm on the at right planet---everyone visi.com looks like me!!!
"Grant Edwards" <grante@visi.com> wrote in message
news:4051cd7c$0$41291$a1866201@newsreader.visi.com...
> On 2004-03-12, Meindert Sprang <mhsprang@NOcustomSPAMware.nl> wrote: > > > I think you're misunderstanding the term differential. These > > drivers are powered from 5V single supply. By setting the A > > line to 5V and the B line to 0V and vice-versa, you get a > > differential signal between A and B. A receiver only looks at > > the voltage DIFFERENCE between A and B, the absolute level > > against ground does not matter. So even when A and B swing > > between 12 and 13V, it would be a differential system, albeit > > with a huge DC offset. But this is of no importance for the > > receiver. > > True within limits. The recievers do have limits on the > allowed common-mode voltage (the DC offset). A typical spec > would be something like -7V to +12V. For such receivers, 13V > would be outside the recievers specs, as would -8V.
I know. It was merely meant as an example to indicate that the absolute (=common mode) voltage on both lines does not matter for the differential signal. Meindert
On 2004-03-12, Meindert Sprang <mhsprang@NOcustomSPAMware.nl> wrote:

>>> [...] >>> A receiver only looks at the voltage DIFFERENCE between A and >>> B, the absolute level against ground does not matter. So even >>> when A and B swing between 12 and 13V, it would be a >>> differential system, albeit with a huge DC offset. But this is >>> of no importance for the receiver. >> >> True within limits. The recievers do have limits on the >> allowed common-mode voltage (the DC offset). A typical spec >> would be something like -7V to +12V. For such receivers, 13V >> would be outside the recievers specs, as would -8V. > > I know. It was merely meant as an example to indicate that the > absolute (=common mode) voltage on both lines does not matter > for the differential signal.
I was pretty sure you knew, but I wanted to make sure the OP didn't get bit by common-mode voltage problems (been there). -- Grant Edwards grante Yow! I selected E5... but at I didn't hear "Sam the Sham visi.com and the Pharoahs"!
Jim Stewart wrote:
> Dennis Clark wrote: > >> I've been reading up on RS485 drivers like the LTC485, 75176B and >> others. >> They look too simple in pin count to supply the + and - voltages needed >> for differential drives. Are there some chips that supply the +/- >> voltages >> that other chips simple scavange and use? Or do these chips I note >> actually >> supply the +/- voltages for the differential outputs? The chip specs >> don't >> say. > > > I'm not sure that I'm answering your question, but > I'll give it a try anyway. RS-485 does not use > plus and minus supplies. It uses +5 volts only, > and sends it differentially on the pair of signal > wires. >
3.3V parts interface nicely with 5V parts. Maybe out of spec.
Dennis Clark wrote:
> I've been reading up on RS485 drivers like the LTC485, 75176B and others. > They look too simple in pin count to supply the + and - voltages needed > for differential drives. Are there some chips that supply the +/- voltages > that other chips simple scavange and use? Or do these chips I note actually > supply the +/- voltages for the differential outputs? The chip specs don't > say. I've a project that I'm going to propose using an RS485 network but > am not sure of all the pieces needed. Can anyone supply me with a place > to look for the "ABC's" of RS485 networking to answer questions like: > > Speed vs. termination resistors vs. line length. > Realistic upper limits of number of drivers on a bus (I've heard 8) > Comparisons of the driver chips available. > > I'd appreciate any and all aid I get. > > regards, > DLC
Dennis, I'm in the process of typing up my notes on RS-485. I'm using it on a project and needed to get to know it better. I see you have been given a number of references, but I have a few more to add. Go to TI, National Semi., and Maxim, and B & B Electronics. I don't have the URL's, but it won't take you long to find them. In addition, Jan Axelson had a good article in Circuit Cellar ("Designing RS-485 Circuits") that is available on the internet. Now to answer some of your questions. Think of differential as being the digital signal and the inverse of it on the two lines. At least that is what you start with, depending on biasing. As you may suspect, the signal will attenuate over distance. As long as there is more than a 200 mV difference between A (the true or '+' line) and the B (or negative or '-' line), the receiver will be able to determine the state. If the difference is less than 200 mV, then the state of the bus is unknown. How does one insure that the bus state is known? You bias it by tying the A line to 5 volts through a resistor. Typical values range from 560 to 750 ohms. The B line is tied to ground with the same value resistor. You may have noticed that I specified a range of bias values. The reason has to do with termination. First, is termination needed? The answer is that depends on the signaling speed (bps) and the length of the transmission line. If the line length is much shorter than the electrical length of the bit window, then termination is not needed. When you get the documents mentioned you will see calculations to determine this. The idea is that the wave front, or Trise, needs time to dissipate before you are into the center of the bit. It is assumed that 4 or 5 round trips around the transmission path will attenuate Trise well enough. Remember the purpose of termination is to match the cable Zo to the load. 24 AWG wire has a typical Zo of about 120 ohms, so if termination is needed, there would be a 120 ohm, 1/2 watt resistor at each end. The problem with termination is that it makes the system less flexible if you plan to be adding and removing nodes at will. So you may want to do a trade-off where the speeds are slow enough that termination is not needed. The other problem with termination is that it is a load on the driver, and increases the current required. The number of drivers is not limited. Of course you can't have more than one driver on at a time because that will play havoc with the data integrity. The number of receivers is what is limited. A receiver can have 1, 1/2, 1/4, or 1/8 unit load (check with the receiver specifications). A unit load is about 12K ohms. So,if all of your receivers are one unit load, the spec. says that you can have up to 32 of them. More if you use receivers of a fractional unit load. Comparing chips is something I have to let up to you because they all have pros and cons. I'm out of time, but I hope this helps somewhat. Let me know if you have more questions. Sincerely, Dennis Zimmerman -- If sending a reply you will need to remove "7UP".

Memfault Beyond the Launch