CD74HC4067: Problem using SCT013 current transformer

Hi Akshay,

Quick question - after looking into the sensor used - is this application passing signals that go below 0V? If so this could cause issues in measurement because the device is not rated to pass below 0V and could sustain damage if voltage goes lower than -0.5V and the clamp diodes have broken their absolute current max. Or is the DC biased used to keep all voltages positive w.r.t ground? 

However if all voltages w.r.t. ground at the mux are 0V or above there are a few other possibilities that could be affecting the application:

On Resistance and On Capacitance - these are parameters of the mux that can attenuate and distort the signal - this part in particular has a high on capacitance at ~55pF and an on Resistance max of 270 Ohms. Since this device isn't spec'd at 3.3V the on resistance shown in the datasheet is actually lower than what the application you are pursuing might have. The on resistance also changes with input voltage so for a non-DC signal the resistance is variable which also can affect the measurement. If the application is low frequency, like <100KHz, the on capacitance shouldn't be causing much problems but it will change the output impedance under higher frequency conditions. The bias circuitry that is used in this schematic will lower the output impedance seen by the mux - therefore allowing the mux to have a larger impact on the signal compared to if it was just the ADC itself as a load. 

Another issue that can throw off ADC measurements when using a multiplexer is the leakage current - this parameter makes it so the current entering the switch does not equal the current leaving the switch. We didn't spec it directly on this device - however the off leakage can go up to +/-8uA and usually the on leakage is pretty comparable to that. This is very high and honestly I wouldn't be surprised if this is where a lot of that error is coming from. Leakage can vary with temperature and input voltage - which is why you could have possibly seen the increase with an increase in input voltage

I will be honest - I don't think this a good part for ADC applications - especially when we are working with smaller voltages that have smaller gaps between measurements. This part has a high leakage current, high on resistance, and a higher on capacitance. While I am not too concerned with the capacitance the leakage and on resistance could be causing errors. Also when adding more channels the resistances of each switch can swing up to 10 Ohms. So this means each channel will also have error associated with it. 

That being said - are there any reasons that you picked this part beyond it being a 16:1 mux?

Also would you possibly be open to a multi-chip IC. The issue is a lot of our 16:1 mux's aren't great for precision applicaitons but we do have a precision family that has very low leakage (pA range typical) which is teh TMUX11xx family where we would take 2 8:1's and 1 2:1 to create a 16:1 multiplexer. The draw back obviously is that there are more IC's in the solution but they are more prepared for precision applications. I think this is the best part family for this application. However the TMUX12xx family also offers low On resistance and a typical leakages in the nA range. However the smaller the leakage current the less error you will see.

Please let me know if you are open to other solutions because I am skeptical that this is the best part for this application and a multi-chip solution seems like it will produce better results.

Best,

Parker Dodson