3 Replies Latest reply on Oct 30, 2018 2:14 AM by gani

    MBR3 temperature response

    scarey_3621476

         I am developing a product using the MBR3. The product is looking to sense a user input for (possibly) hours at a time and we want to make sure moderate temperature changes while operating do not cause the output to remain on after the user releases the sensor. However, we have found that the unit has a fairly large change in detected capacitance based on temperature. We are taking steps to mitigate faster temperature changes, although I would be interested in information about mitigating that as well.

       

       

           Our main concern right now is that if the product experiences even a very slow temperature change while receiving a valid input it will cause the unit to continue to output even after the user is no longer activating the sensor for very reasonable temperature changes because the baseline has not changed. We have investigated the available options in settings including the auto-reset period but have not found anything to mitigate this problem since we want the output to remain on during the entire user activation and stop only when they release it.

       

       

      Here are my questions, then:

      1. What options do we have with the MBR3 chip to mitigate the change in detected capacitance due to temperature other than lowering the sensitivity?
      2. Are there options that becomes available if we upgrade to a different unit to help with the same problem?
      3. Is there a solution to the problem of temperature changes causing the detected capacitance to change so much during valid input that it causes false outputs after the user has released the sensor?
        • 1. Re: MBR3 temperature response
          gani

          Hi,

           

          1. On every start up/ reset, the MBR3 device calibrates the CapSense parameters (basically the co-ordinates of raw count and parasitic capacitance curve are fixed). From then (after calibration), the raw counts are generated according to the parasitic capacitance of the environment. With temperature, the parasitic capacitance of the sensor changes. So the raw counts changes according to the temperature. If the sensitivity is more, the change in raw counts will be more with temperature. So you may detect sudden rise in raw counts with temperature. To mitigate this, you can even increase the noise and negative noise thresholds or you can reduce the sensitivity. This will accommodate more swing of raw counts.

           

          2. We recommend you to go with PSoC 4000S/4100S/4100S plus devices (programmable) which contains latest CapSense IP as well as latest CapSense firmware version of component. Moving this to these components gives you much more flexibility in terms of hardware and software parameters. Also the latest CapSense IP supports the parasitic capacitance of 200pF.

           

          3. As mentioned in 1 you can increase the noise and negative noise thresholds and check once. Please ensure that auto reset is disabled.

           

          Thanks

          Ganesh

          • 2. Re: MBR3 temperature response
            scarey_3621476

            Hello Ganesh,

             

            Thank you for your reply. Can you help me understand your answers more clearly?

             

            1. I understand the automatic baseline setup, but are you referring to the 'finger threshold' when you talk about increasing the noise thresholds? I am using the EZ click software to program the MBR3 and do not see an independent setting for noise vs. finger thresholds like there is on PSoC devices. Is that setting not available with the EZ-click software and must be programmed some other way?
            2. Thank you for your recommendation of device. Can you explain what options become available on those devices to solve our problem and if they are specific to the devices you listed?
            3. I think perhaps you misunderstand - I am not talking about slow changes of temperature while the user is not activating the sensor. I agree increasing the noise thresholds will help with this. I am talking about changes in detected capacitance due to temperature change that happens WHILE the user is activating the device. The baseline does not update in this scenario, so if there is a big enough change the output will give a false positive when the user stops activating the device. Do you know of any solutions to this issue?

             

            Thank you for your help on this.

            • 3. Re: MBR3 temperature response
              gani

              Hi,

               

              1. Please refer MBR3 register TRM obtained from the ink below. Ez-Click do not have options for changing the noise thresholds. You have to change the noise thresholds through changing the configuration registers through I2C using any host/bridge control panel software. Please refer the steps mentioned in 5.2.4.1 Procedure to Configure CY8CMBR3xxx Using Bridge Control Panel in page number 68 of the MBR3 CapSense Design guide downloadable from the link below:

               

              You can download MBR3 register TRM from the link below:

               

              http://www.cypress.com/file/127221/download

               

              If you want to use bridge control panel you have to download PSoC programmer. The bridge control panel comes along with PSoC programmer.

               

              2. The programmable devices we mentioned contains latest CapSense IP along with latest CapSense component firmware.

              The programmable devices are more flexible in setting hardware and firmware parameters for each sensor. You can also alter the firmware flow in case of programmable devices. For more information please refer the PSoC 4 and 6 CapSense Design guide obtained from the link below:

               

              http://www.cypress.com/file/46081/download

               

              3. I understood your question clearly now. MBR3 device as such do not have any techniques internally to compensate this effect. You need a custom baseline update algorithm for this.

               

              Thanks

              Ganesh