1 of 1 people found this helpful
This is primarily due to one limitation of SAR Sequencer(The part of SAR ADC which takes care of the sequencing and conversion formats without CPU intervention). Using the same signal for the positive half of one channel and the negative half of another channel under SAR Sequencer control is not feasible, and in general is not recommended due to the way that differential pairs are designated in the SarMux. In this case, it is possible to use the amxubus as a workaround and route both, but at the expensive of reduced performance for the negative input of channel 0 (which must route through an additional switch and traverse the high-capacitance amuxbus). The best option left is to use an additional pin for this purpose.
Thanks Vasanth. My design is remarkably low on pins, to the point way may not meet requirements unless we can save some. So even though this may not be ideal, it is a thermometer, so I'm not too worried about conversion speed.
Can you clarify: "not feasible, and in general is not recommended" ? It would seem options are: 1. maybe feasible but not recommended, 2. not possible.
Given I seem to have been able to route it, it seems it will operate, or am I mistaken?
If it will operate, can you suggest how to best deal with the less than ideal routing. Should I slowing the conversion or averaging in more values compensate for the longer path and switching?
This is what I have: Thanks.
I assumed that you were using PSoC Creator seeing the first schematic image. The Analog MuxBux component in PSoC Creator prevent the routing due to some limitation. As using Amuxbus was the only option left which not the best option, it was not recommended.
But as you are using ModusToolBox this issue is resolved. ModusToolBox 2.2 latest version allows this routing . You can use the dedicated pins for the purpose without any problem. As you are using the dedicated pins(SARBUS connections) for the purpose, there is no issue here. You have to re-use the one dedicated from different channels. The following image(not actual settings) should give more clarity. This should help you to save one GPIO pin.
This was done to increase the accuracy of the measurement. There is an excellent application note which talks about these issues.