Closed loop air filtration based on EN148 filters

As I print a lot I really wanted a solution to keep things healthy.
I print a lot of ABS and this ABS emits fumes and fine dust, both things you don't want in your lungs.
I choose EN148 filters because they are meant to protect our lungs when used in their normal setup (you screw these to face masks, when spray painting, brazing etc.)

I did not have a lot of equipment to test this but:
When using dimafix spray for bed adhesion TVOC values (I have this air quality meter) go up (a lot) inside the printer enclosure.
With a double P3 filter for particles and an AX filter for gasses I was able to lower tvoc to almost normal levels.
As the filter is capable of this I assume the fine dust is also filtered out, I don't know this for sure as I do not have a fine dust meter capable of measuring those things reliably.
I actually have two of these, one is dual P3 + AX, the other one is dual P3 + A2B2.
This to filter as much as possible, I am fairly confident the thing works.
Do note, in order to get some airflow over the filters you should have a fan capable of high static pressure, and not those things for pc cases, I tried several fans and the only fan capable of getting any flow was the Sanyo Denki San Ace 9HV0812P1G001.
Keep in mind that the holes for the wires should be sealed, I used Super glue, this caused the white effect on the abs (see picture).
I had a PWM circuit for it but I wouldn't bother with that, they need to run at full speed to be effective.
The whole thing is assembled with M4 threaded rods,nuts and washers.

Some things I noticed:
-Dust always seemed to stick to my hot end fan, this is much less the case with these filters on.
-Having filters on both sides of the fan decreases the noise a lot.
-I had best performance with the P3 before the fan and then blowing the air out over the gas filter.
-The whole thing is loud, not extremely loud but still but for a desktop printer situation it may be too loud (it can be distracting).
-Sprays for bed adhesion are not that healthy tvoc wise, read the instuctions.

Any questions or remarks, please ask

Make sure your read this, http://pubs.acs.org/doi/abs/10.1021/acs.est.5b04983
I also want to thank Brent Stephens and many others in helping me come up with a working solution.