[Stephen Woodward] is familiar with digital potentiometers but is also familiar with their limitations. That spurred him to create the PWMPot which performs a similar function, but with better features than a traditional digital pot. Of course, he admits that this design has some limitations of its own, so — as usual — you have to make your design choices according to what’s important to you.
Perhaps the biggest limitation is that the PWMPot isn’t useful at even moderately high frequencies. The circuit works by driving two CMOS switches into an RC circuit. The switches’ inverted phase tends to cancel out any ripple in the signal.
The RC circuit is selected to trade response time with the precision of the final voltage output. The CMOS switches used are part of a 74HC4053B IC. While it might not solve all your digital potentiometer problems, there are cases where it will be just what you need.
We’ve looked at traditional digital pots before. If you prefer the hard way, grab a regular pot and a motor.
Interesting. I found a different approach which uses an opamp via stackexchange: https://arduino.stackexchange.com/questions/1844/how-to-emulate-an-analog-potentiometer-with-pwm/13850#13850
Both seem to need filtering to take care of the ripple.
That’s not the same thing. It’s just converting PWM to voltage with a low pass filter.
The idea of a digipot is to offer a digitally controllable variable resistance. This allows you to put the digital pot anywhere in a circuit, such as in the feedback path of an op-amp, to make a digitally controllable volume pot. You’re not just producing some voltage at the output – you’re altering some other signal voltage or current with your digitally variable resistance.
The PWMPot kinda doesn’t, because it appears to present a constant R, so it’s not a correct emulation of a digipot. What it is doing is essentially picking two input voltages A and B and switching between them rapidly, then filtering the resulting square wave down with a lowpass filter, giving you a weighted DC average between A and B chosen by the input PWM signal.
With the frequency limitations, it’s somewhat difficult to see where this would be useful.
“The idea of a digipot is to offer a digitally controllable variable resistance”
Variable voltage divider, not resistance. That would be a (digital) rheostat.
It does both. You don’t have to use both sides of the circuit.
You don’t have to use a digipot as a voltage divider. It’s technically two digitally variable resistances that have a common output. It can be a voltage divider, a rheostat (two rheostats) etc. like a real potentiometer.
It’s a potentiometer (variable voltage divider). How you connect it is up to you. You can use it as a rheostat (variable resistor), but that’s a subset of its features.
Some digital potentiometers have rheostat-only packages and those variants are labeled as such. My beef with your comment was you wrote “The idea of” instead of “one use case of”. That’s all. I’ve seen many beginners confuse potentiometers and rheostats, so that’s why I added my initial comment.
Your main point of post I agree with. That link that was shared is just a dutycycle to voltage converter. But if the supply voltage of that PWM signal is a signal, then it could function as a variable gain attenuator, not a 1to1 replacement of a potentiometer since it requires filtering and one of the terminals is connected to ground.
Known for centuries… Just Google for switched capacitor circuit or switched capacitor filter…
Btw, there are very capable analog switches in modern 74 series like 74lv.
Oh, and those usb switches/mux ICs are capable analog switches too. They are frequently used in switching IQ mixers in SDRs…
73
Analog pots are just a FET with variable gate voltage, right?
No. The dynamic resistance of fet is dependent not only on gate voltage, but also on io pins voltage.