Switching Pull up / down resistors in/out of circuit digitally

[JeremyWilliams83]

Hello,
I am tying to come up with a circuit to allow me to switch some pull up and pull down resistors in or out of the circuit on some differential signal lines (RS485 fail safe bias resistors). I need to get the enable for the pull ups across a opto to my isolated driver circuitry. Is there a easy way to switch a pull up resistor in/out from a digital signal. Many micros have internally enabled pull ups so they must do it somehow. I tried using MOSFETs and BJTs but havent been able to make that work, I think I am getting base-collector/body diode current flowing when I don't want it to causing excessive current draw and bus loading.
regards,
-Jeremy

 

[itsmoked]

Use the optos themselves as a way to turn them on and off?

Keith Cress
kcress -

http://www.flaminsystems.com

 

[JeremyWilliams83]

I am actually using digital isolators not optos because the power consumption is much better and this needs to be as low power device as possible. The digital isolators have logic outputs...

 

[itsmoked]

Well... Just get a MAX13450E. It's totally bulletproof and has turn on-able pull ups along with wire reversal and a fault output.

MAX13450E

Keith Cress
kcress -

http://www.flaminsystems.com

 

[JeremyWilliams83]

I looked at the MAX13450 when i was initially designing the prototype, it has switchable termination but not bias. I am thinking of using a CMOS switch such as the
SN74LVC1G3157 but am a little concerned with possible over voltages exceeding the 5V rail of the analog switch (RS485 is spec'd to -7 to +12V) but since the port is isolated it shouldn't have much for common mode voltage (i don't know of any drivers that actually drive more than 5V). I have used an analog switch to switch termination resistors in and out of circuit and it seems to work OK.

 

[zappedagain]

How about using a fail-safe RS485 transceiver and not bothering with all this switching? TI SN65HVD12D comes to mind.

Z

 

[JeremyWilliams83]

I am using a fail-safe transceiver, however since this device is just a repeater/converter meant for existing industrial networks I have no control over the driver architecture of any of the connected devices and deal often with pretty old legacy PLC's that may require a bias but do not contain bias internally. I hacked in some analog switches (SN74LVC1G3157D) in place of the FET's I originally used and it seems to work well, it was a pain to solder leads to the SOT23-6 package, but I got it working enough to do some tests.