Calculating Current Demands...

[RocketMan]

Dumb-Question time...   

I'm trying to calculate the power-supply requirements for a project I'm working on, but can't figure out what the RS-485 interface chip requires...

The datasheet for the part in question - a Maxim MAX13432E - lists a supply current (Icc) of 2ma, but I'm guessing that's when the chip isn't transmitting.  It has a listing for driver short-circuit current (Iosd) of 250mA... but I'm further guessing that this rating is a shut-down trigger for a fault.  The 'Typical Operating Characteristics' graphs suggest that it drives the transmitter to a maximum of approximately 60mA, with spikes to 120mA when the Output-High voltage reaches 3V or the output-low voltage reaches 1V.

Should I design my power supply to accommodate the short-circuit current, or would I be safe with designing for a 60mA current draw...?  It's the difference between building my board as-is, redesigning the board to fit a heat-sink, or placing the power supply off-board...   

[Jan Axelson]

From the data sheet:

The short-circuit output current is the peak current just prior to current limiting;

In other words, the chip will limit the current to that amount in case of a short. You don't need to provide that amount.

Most of the current will be through the termination resistors. For 120-ohm twisted pair, these would be  two 120-ohm resistors in parallel across the lines.

Jan

[RocketMan]

In other words... to figure out the current demand while transmitting, I just apply Ohms' Law to the driver voltage and resistance...?

With a 3.3V Vcc and (as you described...) two 120-ohm resistors on the line making a 60-ohm overall resistance, I get...  3.3 / 60 = 55mA

... and then add that to the current used when *not* transmitting, to get the total current used.

Now for a follow-up question... does it affect the network if I use drivers with different voltages?  That is - can I mix 3.3V nodes and 5V nodes on the same link?  I don't *see* a problem with such an arrangement, but I would rather double-check with someone who knows more about the subject than I do...   

[Jan Axelson]

The terminating resistors don't have the full supply voltage across them.

If you're using biasing resistors, those can greatly reduce the voltage across the termination.

What matters to the RS-485 interface is the differential voltage across the signal lines. As long as you maintain that you should be OK.

Jan