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.

TIC10024-Q1: comparator threshold questions

Part Number: TIC10024-Q1

Dear team,

Does our comparator threshold have hysteresis? For example, we choose THRES_COMP=2V, if the input signal varies from 1.9V to 2.1V, then the result will change, right? If we don't have hysteresis, our device will have fault report when input signal varies.

Could you please tell me the R1 and R2 value? VS is 9V~16V for my customer, we need to make sure the resistor divider is feasible.

Thanks & Best Regards,

Sherry

  • Sherry,

    Sorry for the delay here, I've assigned this thread to an engineer and they will reply by tomorrow.

    Regards,

    Eric Hackett 

  • Sherry,

    The 1.85 V - 2.25 V characteristic represents the range of values that the threshold for a particular device could fall in. This means that if one particular unit has a 1.85 V threshold, then input varying 1.9 V to 2.1 V would not reach the threshold. Of course, the same applies for if a unit's threshold is above 2.1 V.

    I'm unclear on the second half of your question. Are you looking to find the leakage currents for the device inputs?

    Best,

    Danny

  • Hi Danny,

    Thanks for your reply!

    For the first answer, I understand it. But in this way, our device's status is uncertain under some worst cases. The typical threshold is 2V, if the input signal is varying 1.9V~2.1V, then our device's output will be toggling, which is not expected for the customer. Because 0.1V input fluctuation is acceptable. In addition, for small portion devices, their threshold is 1.85V or 2.25V. For these devices, when input varies between 1.9V~2.1V, their result is opposite....

    For the second question, for example, the threshold is 2V, then the comparator's negative port is 2V, and the positive port is Vin*R2/(R1+R2). Vin is 9-16V. 16*R2/(R1+R2) should be larger than 2V, but we are not sure what is the result of 9*R2/(R1+R2)? So we want to know the R1, R2 value.

    Thanks & Best Regards,

    Sherry

  • Sherry,

    Thanks for your patience here, Danny is out of office at the moment but we will look into this and try to get back to you by early next week at the latest.

    Regards,

    Eric Hackett

  • Hi Sherry,

    First, to comment on your original question, the comparator does not have hysteresis. The exact threshold may be different from device to device due to manufacturing tolerance, but the threshold will be somewhere within the specified voltage range.  For example, the threshold may be 1.9V for one device, but 2.1V for another device.  A signal that is exactly 2V would cause one device to detect a Low voltage level, and the other device would detect a High voltage level. 

    The input pins are sampled in a sequence once every polling cycle and the comparator determines if the voltage is above or below the threshold.  Then the device's MUX routes a different Input pin to the Comparator for sampling.  The original input pin does not get sampled again until all other enabled inputs have been sampled and the cycle repeats itself.  Therefore hysteresis is not applicable in this type of architecture.

    The thresholds should be selected to ensure a High/Low voltage will be detected for the entire Min/Max range. To guarantee a Low voltage sample on an Input with a 2V threshold, the input voltage should always be below 1.85V.  To guarantee a High voltage sample, the input voltage should always be above 2.25V. If the input voltage falls between the Min and Max levels, this should be treated as an "unknown" from a system perspective because it may be sampled differently on different devices. 

    In the case of a Switch to GND, or Current Source (CSO), configuration, an open switch would allow no path for the current to flow and the voltage would be sampled as High.  If the switch was closed, it would in theory provide a low impedance path to GND and the input voltage would be sampled as Low.  The questions that need to be determined are:

    1. What is the impedance of the switch (and any other components on the net).
    2. How much wetting current is required to create the required voltage drop for this impedance.
    3. Which threshold fits best for this voltage drop.

    If it is simply a question of stability in the sampling, the device can be configured to require multiple consecutive samples before indicating a state change on the Input.  This is referred to as a Detection Filter and is set in the DET_FILTER bits of the CONFIG register.  This is useful in noisy environments or where the mechanical switch does not have a clean signal and avoids multiple High/Low/High... state changes while the voltage stabilizes.

    The resistor values on the input of the comparator vary with the threshold selected and the equivalent input resistance for each threshold is listed in the datasheet.  Once again, manufacturing tolerance will cause these to vary slightly with stronger/weaker resistor values in different physical devices.  This input impedance should be considered in parallel with the impedance of the external switch when determining the required wetting current and threshold voltage.

    Regards,

    Jonathan