4 Replies Latest reply on Jun 20, 2018 11:09 PM by PriyadeepK_01

    CSD or CSX (self or mutual) method?




      Does anyone know which method (CSD or CSX) is the better choice if you want use buttons with 10mm overlay-glass?


      thank you very much

        • 1. Re: CSD or CSX (self or mutual) method?

          Either CSD or CSX is OK.

          10mm overlay-glass should not be the main reason to select CSD or CSX.

          Are there any other requirements in your application?




          • 2. Re: CSD or CSX (self or mutual) method?

            Hi P. Kaesar,


            10mm overlay is quite thick. I would recommend to chose CSD. Please ensure the traces that connect the sensor pad to the device are short so that the Cp of the sensors is not very high.

            Also, I would recommend to allow an option to connect the surrounding hatched pattern either to shield or ground. Connecting the surrounding pattern to shield will provide high sensitivity. You can use this option if connecting the hatched pattern to ground is not sensitive enough.

            By the way, what is your end application?

            1 of 1 people found this helpful
            • 3. Re: CSD or CSX (self or mutual) method?

              do you think CSX measuring will not work with thick overlay glass + Airgap (0.2mm)?


              I already did some test's with CSD and 9 buttons on a PCB (only single Layer), outline max 50x55mm.
              Because the buttons needs to be as big as possible to get enough signal, I have no more space to add a hatched pattern/shield.

              Additional with CSD I need to wire 9 connections.


              So far it works good with thick glass, but I also have much interferences, means if I touch a button I also get some signals from other button, but the signal (diff-counts) are not such high than the touched button, so it is detectable.


              Now with CSX I hope to get follow advantages :

              - less interferences because capacity is measured between tx/rx and not to earth/human.
              - less interferences regarding environment changes.
              - only need 6 connections instead of 9 connections.
              - less power consumption?


              Finally the 9 buttons will be on a glass with transparent metal electrode (ITO).
              On the backside from that glass we will have a Display and on the top as mentioned thick overlay-glass incl. airgap.


              thank you very much

              • 4. Re: CSD or CSX (self or mutual) method?

                CSX definitely gives advantage being more pin efficient but may not necessarily give power consumption advantage. Having said that, CSD works better for very thick overlays. 10-15mm sounds a little too high for using  CSX.

                Regarding Airgap - it's generally recommended to not have any air-gap between buttons and overlay. Air-gap tends to vary from unit to unit leading to different sensitivities across units. If it's imperative to work with air-gap, you can tune the buttons for maximum possible airgap but that will mean that there may be slight hover effect for boards which have lesser air-gap due to manufacturing variations.


                What is the button size you are using, a 10mm diameter round button should be sufficient I believe. If the buttons are closely spaced, which it is in your case, it is expected to have interference from nearby buttons. Thick overlay only adds to this because the farther the field spreads in z direction (as intended), the farther it will spread in other directions.

                You can try using firmware techniques to minimize this interference (for example, set the thresholds higher than the signal from nearby buttons when finger is on center of one button). Ensure if user touches two buttons at a time, the button with the highest signal is the one that is detected ON. The later method can only be used if two buttons need not be touched simultaneously.

                You can also try changing the component settings to drive unused sensors to shield to get higher signal.


                Since your 9 buttons are on ITO, I would recommend you don't put the additional 560 ohm series resistance (or use a 0 ohm pad to allow testing) because ITO inherently has higher resistance. 

                I'm a little concerned the display might add noise to the buttons, best way is to test it out and try out some filtereing techniques (use those mentioned in AN92239). If possible, having a shield/ground hatched layer between display and buttons would help.

                1 of 1 people found this helpful