• State Resolved
  • Locked Locked
  • Replies 3 replies
  • Answers 1 answer
  • Subscribers 33 subscribers
  • Views 348 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

DS89C21: DS89C21

Pierre IZARD
Prodigy 20 points
Part Number: DS89C21

Tool/software:

Hello,

I would like to implement an AC termination for a reduced consumption with an inverted polarization to ensure a low output in case of failure of the input in high impedance or short circuit according to the following schematics.

Could you please confirm the polarization circuit will have the correct behaviour (low output in case of an input HI or in short circuit) with the receiver input impedance of the DS89C21 ? and that it won't interfere with the operational mode of communication knowing that it's a low speed control application (30KHz max) , in point to point with a distance of 50cm to 1m maximum.

Thanks in advance for your help, I have no mean to simulate it (only IBIS model is available).

Best regards

Pierre

  • 0
    TI__Guru 72901 points
    Pierre IZARD said:
    Could you please confirm the polarization circuit will have the correct behaviour (low output in case of an input HI or in short circuit) with the receiver input impedance of the DS89C21 ?

    I assume you mean if the circuit is input high Z (left open without a driver enabled). Yes, the output RO will be low because A-B should be negative during that case. 

    In the case where A is shorted to B (at the pins), the A-B voltage would be 0V so in that case it would be unknown (A-B needs to be above or below 200mV to guarantee a value). 

    In the case where A and B are shorted to the left of your diagram (left of the AC termination), A-B should be less than -200mV (I calculated -280mV) so the RO would be low. 

    Pierre IZARD said:
    and that it won't interfere with the operational mode of communication knowing that it's a low speed control application (30KHz max) , in point to point with a distance of 50cm to 1m maximum.

    Your operating speed is very slow and the distance you're communicating at is also short. You probably don't need to worry about reflections in your application.

    -Bobby

  • 0 in reply to BOBBY
    Prodigy 20 points

    Hi Bobby

    Thank you for reviewing and verifying my design. May I ask you for some additional opinions and confirmations?

    Do you confirm that the calculations of High Z and short circuit input levels, with the values ​​of the bias resistors, are not impacted by the receiver input resistances which are relatively low (9.5K typical and down to 5K minimum) and that there is no need to have lower bias resistors (1K or even 470R) to consider that the input resistors have no impact (which would significantly impact the consumption)?

    If we want to evaluate precisely the input levels in the different cases, we can do a simple simulation under SPICE by considering the receiver input resistors are connected to ground (the calculations are a bit complicated with the different resistors in parallel depending on the cases) but anyway there should be a significant margin.

    There should be no problem because of the divider bridge (with the 270R series resistors) given the driver output levels (which should be higher than 3V since the load should be 5K min = 2 * (Rbias // Rin) = 2 * (4.7k // 5Kmin)). The minimum input levels should be equal to 2.9V because of the divider bridge, which is well above the threshold.

    There should be no risk either that the bias circuit impact DS89C21 delays which are characterized in the datasheet?

    Thank you very much again,

    Best regards

    Pierre

  • +1 in reply to Pierre IZARD
    TI__Guru 72901 points

    Hi Pierre,

    I think in the actual application, your input voltage will most likely be between GND and 3.3V so the leakage coming from the device will in the range of hundreds of microamps. Using the leakage current value, if we say 3V on the input and 220uA leakage the input impedance would be 13.6kohms. 

    We can use the extreme case and say due to GND shifts you end up seeing 10V common mode and use the 5k minimum value. I believe this leakage will end up looking like a 5k resistor to GND on both inputs. I threw this into a simulation tool and saw that even during the short to the left of the AC termination, we still ended up seeing larger than 200mV difference between A and B pins. 

    In the case of an idle condition, you basically would have about Vcc/2 which meets the threshold requirement, so the only case we really needed to look at was the short scenario. 

    Pierre IZARD said:
    There should be no risk either that the bias circuit impact DS89C21 delays which are characterized in the datasheet?

    I can't think of anything that would cause it in your lower speed application. Technically the circuit you have set up changes the AC termination value of the system because from an AC stand point the Rc resistors and the External pullup/down resistors run in parallel with the termination. Since the distance and speed of your application is low/short, we don't really need to modify the AC termination resistor for it......I ran it anyways, the termination resistor would need to be about 122 ohms to make an equivalent resistance of 120 ohms. Using 121 ohms would work.