DS 1882 question

[hummusdude]

the Maxim DS1882 digital potentiometer uses a split power supply for the analog signal side. The datasheet doesn't address my question specifically, which is...that seems more like an op amp kind of thing. Is there buffering?

I came to this because I was trying to design a basic audio pre-amp with digital gain control and I started with an X9C104 which is single supply and does not appear to do anything other than pass the original signal through some combination of resistors in series to get the desired resistance. But in posting to forums I was advised to go with the MCP4231 because it is supposedly better suited to audio applications. The MCP4231 is also single supply but my I made the noob mistake of assuming I could treat it just like an analog pot. As it turns out, it has a minimum input voltage level of - 300 mV. My signal goes to - 360 mV and my chip is toast. It seems likely that I pushed it to far.

So in posting to forums yet again I was advised that the DS1882 is yet an even better fit for audio, but it uses dual supply. So my question is, why does it need dual supply unless it is also buffering as an op amp would?

[Manul]

To put it simply, signals must be in the range of supply rails used to power resistor ladder switches. DS1882 has dual supply for the analog side to be able to accept bipolar signals (negative and positive). Also, separate analog and digital supplies generally reduce noise coupling from digital side to the resistor ladder. Some has internal voltage generation like X9C104, but it has more noise, because it is a charge pump. So DS1882 allows you to use separate, clean analog power supply for the ladder which expands the input signal range and has possible benefit of reduced noise. No buffers, as far as I understand.

[ledtester]

The Typical Operating Circuit shows buffering by external op amps, so I doubt the DS1882 itself does any buffering.

https://datasheets.maximintegrated.com/en/ds/DS1882.pdf

[hummusdude]

Maybe I'm missing something, but I don't get why it would be necessary to have a negative supply, or any supply really, for anything other than the switching. I know the X9C104 actually has 100 resistors, each attached to a transistor switch and that is how it manages to increase and decrease the resistance of the signal path. So I guess I'm confused about how the DS1882 utilizes the split supply unless it is doing something similar to an op amp. Put another way, what is happening with the signal path other than some combination of series resistances such that it needs positive and negative voltages to function?

[ledtester]

This video describes the basic operation of a CMOS analog switch:

https://youtu.be/h8ad1-Sy39A

A key point is that there needs to be a certain voltage difference between the gate voltage and signal voltage for the switch to conduct. The V- rail is likely used to allow the switch to conduct negative voltage signals.

From this stackexchange discussion some switches have internal charge pumps to allow them to switch voltages which are beyond their supply rails:

https://electronics.stackexchange.com/questions/100407/how-do-beyond-the-rails-analog-switches-work

[ledtester]

As it turns out, it has a minimum input voltage level of - 300 mV. My signal goes to - 360 mV and my chip is toast.

Indeed, the MCP4231 has the Absolute Maximum specification:

Voltage on all other pins (PxA, PxW, PxB, and
SDO) with respect to VSS ............................ -0.3V to VDD + 0.3V

I think the solution is to AC couple (level shift) your audio sign to a ground of VDD/2 so that your signal is fluctuating around VDD/2 instead of 0 V. Connect one end of the digital pot to this new audio ground. Then AC couple the wiper (your output) back to your normal audio ground.

[hummusdude]

That sounds right. I've read in several app notes from Maxim that a DC bias on the input is SOP. I was just curious about the presence of the negative supply voltage. I see you addressed that already...sorry for the oversight.