This is very, very mission dependent. Where does it land? How long do they stay? What is the mission hoping to accomplish, beyond producing enough fuel to return?
Add in all the other unknowns (How much water is locked in the Martian soil? How deep? Does that change over time? How?) and this question is very hard to answer.
But, Perhaps Less Equipment Than You'd Think
Hohmann transfer windows suggest a Mars mission will be ~18 months on the ground. Let's call that 500 Martian days. If they can mine 1.5 tons of water a day, they'll easily have >600 tons at the end of the mission.
The Korolev Crater has 530 cubic miles of surface ice - if they landed there, a team of 10 or so could probably mine 1.5+ tons of ice a day with shovels.
I... don't know that I'd sign up for that mission.
Pre-Mining Might be Better
If instead you landed one or more rovers in advance, you could feasibly have collected all or most of the water prior to the main mission touching down. This seems a much more reasonable answer to me, but implies we won't be seeing a manned Martian mission any time soon, since they'd have to launch the rover mission at least one Hohmann transfer window earlier than the manned flight, and Hohmann windows only come every two years.
EDIT: Let's Add Some Numbers
If our rovers arrive one Hohmann window ahead of the the manned mission, they'd have ~1,200 days to harvest material - they'd need to produce 0.5 tons water a day on average.
From there, we have to make many assumptions. Let's assert that Martian soil near the surface is 1% water by mass. Our system needs to extract the soil, separate out the water, and dispose of the slag. For 0.5 tons a day of water, we need to process 50 tons of soil.
If a rover can dig up and carry a "load" of 200 pounds of soil, and each rover is capable of 10 loads per day, one rover collects one ton per day. We need 50 rovers to process 50 tons of soil per day.
Recent Martian rovers have weighed around 1 ton. Since these new rovers will be performing more "industrialized" work, let's double the weight. This implies 100 tons of rover, plus whatever we need to separate the water from the soil, and whatever we use to store the water.
Obviously, these end numbers are very sensitive - if we assume that the soil is 10% water by mass, that changes the outcome by an order of magnitude - only 5 rovers are needed for a total of 10 tons. Similarly, if the rovers are more mobile, or carry more weight per trip, then the total weight scales linearly with those changes.
So the answer depends heavily on information we just don't have.