The article referenced on Adafruit suggests that 30 Hz is the minimum practical. For its original use in welding helmets, it is important that it darkens extremely quickly, but it is not required to clear anywhere near as fast. This is a common feature of liquid crystal devices.
The Adafruit article also points out that this device is primarily capacitive, so the current draw when driven by AC is proportional to the frequency. However despite appearing as a capacitor, it is stated to be damaged by persistent DC, so you do need a non-polarised capacitor in series, probably around 1 µF.
I cannot see what a LPF has to do with this discussion.
Since it is specified to require 4 v (and not more than 5) for full darkness, you arguably would want to drive it in bridge mode from two Arduino pins, but I don't know how you can use PWM pins to do that - simply defining
The frequency had to be between 30 and 60 Hz as per the datasheet.
I chose 30.64 Hz, since its within that range. The only other frequency I could have changed it to with software alone was 61.04 Hz, that was above 60 so I was worried it would damage the display. I'm sure I could use external circuitry to make the frequency another value between 30 and 60.
Ignore the second PWM pin. Based on the earlier responses, I was going to connect one PWM pin to a capacitor, and then plug that into one pin of the shutter. Then I was going to ground the other pin of the shutter. The datasheet says these shutters take 5-10V peak, so I was going to step up the voltage range with some external circuitry, like an op amp, to 10V peak for max opacity.
I'm confused about your comment about 4V. I thought that was referring to the DC voltage and not AC. It says in the description "Starting at about 1.0VDC, the glass will start to darken." And then the next sentence says "at about 4V, the glass will be opaque", so I assumed that was referring to DC because of the previous sentence. Like I said, the datasheet says these shutters typically operate on 5-10 V peak.
They clearly do mean DC, but the point is that you have to apply that same voltage alternately in both directions, so you require an excursion of twice that to apply through the capacitor. And with the capacitor that you do need to block DC, it is suggested you have a resistor across the LCD to actually zero the voltage and clear it when you stop applying the alternating current.
