Like many others I don't believe that the dry Pac system has any value as it is left open to atmosphere even when the pump is stored and so the desiccant is compromised before you start.
It's also likely to be less efficient to try to remove the vaporised water at atmospheric pressure, rather than after it has become saturated, by the compression of the air.
I calculated that the volume of air required to raise the pressure of my 0.175 litre HW100S cylinder from 100 to 200 bar is approximately 20 litres. The volume of water likely to be stored in 20 litres of air is around 0.6 gms.
I therefore decided to install a filter in the filling whip that was capable of catching and retaining that volume of water.
I bought and use a small gold filter, that has a small tampon installed in the screw in cap. The cap is a relatively large block of metal and acts as a useful heat sink for the warm moist air entering the filter. I chill the filter before use to -18°C to encourage condensation of any water vapour. This should then be absorbed by the tampon before the colder, drier air continues on its way through the 10 - 15 grams of molecular sieve and self indicating silica beads packed into the filter body.
That volume of sieve in the filter, even at 10% absorption, should be more than enough without the tampon condensation trap.
Checking the condition of the silica beads after filling allows me to monitor the efficiency of the filtration.
The relatively slow rate of fill of a pump assists as the air moves more slowly through the filter, allowing additional exposure to the desiccant and keeping the temperature lower.
If you want to know all of the pros and cons of the options you should use the search facility and read some of
@sagalout posts on the subject.
