The most up-to-date version of this project is on hackaday.io/botsford. Material not posted there is highlighted.
A P100 mask is more secure than a cloth mask and it doesn't fog up your glasses. They are made of rubber and hard nylon and seal to the face without leaks if the straps are correctly adjusted and the height of the mask on the face is correctly adjusted. A suction test and a pressurization test are needed each time the mask is put on.
However, you can’t talk and be understood while wearing one. You have to remove the mask to talk–precisely the highest-risk situation, and talking seems unavoidable in social situations.
I therefore fitted a home-built intercom box to a P100 that I already had and put an electret mic inside. This project cost me $18 CDN ($14 USD) after taxes, not including the batteries. I field tested it between 2 July and 7 September, 2020. Below is the circuit that reached practicality and photographs showing the physical build. Caps shown polarized are aluminum-electrolytic. A mix of cap technologies was necessary due to issues of access impedance.
When I go out in this mask, I take a conventional cloth mask with me in case of a malfunction. You may think that a P100 mask is overkill against SARS-CoV-2, which is widely assumed to spread only by large respiratory droplets, which a simple cloth mask can filter out. However, aerosol transmission, such as the mumps virus utilizes, would require something better, such as a P100. There is some concern that aerosol transmission may actually be a minor route of transmission of SARS-CoV-2, and for all we know, we could be only one mutation away from its being the major route of transmission, which would be very bad news. Thus, it is good for society's pandemic responsiveness to have some margin for error built in, and this project provides some.
I bolted the project box to the exhalation valve with four 1.5-inch Robertson-head machine screws, 6-32 thread, 9/64” dia. These same screws fix the microphone board to the inside of the mask, fix the main circuit board inside the box, and provide conductors for the mic.
To avoid loss of continuity with the mic due to corrosion, use a nut, connector coating (MG Chemicals), and rubberized-fiber washer, in that order, under the mic board on each of two screws, and apply anti-oxidation paste to the mic lugs (on top of the board) before tightening down the top nuts. Then seal with more connector coating. Use two generous coatings, with drying between, to make sure no water can get in.
The centers of the bolt holes in the bottom of the intercom box form a 1.2" × 1.6" rectangle. I glued a cross-shaped spacer of ¼-inch acrylic to the box with epoxy to provide an escape route for the exhaled air. The bolt holes in the mask were carefully sealed with silicone caulking after assembly so that the mask continued to pass the suction test. After drilling, the mask was fully disassembled (a challenging task the first time, requiring the use my two largest jeweller's screwdrivers at once to remove each retaining ring) and cleaned to remove dirt and turnings that might otherwise cause air leaks. The speaker is protected from disinfection solutions by plastic film. I disinfect with 70% isopropanol spray. All exposed metal is protected from corrosion with connector coating.
The mic is protected from humidity by a 1.5" square of two layers of film cut from a sandwich bag and held in place with a 7/16" O-ring. Before the film was put on, I glued a 5/16" dia × 1/16" thick spacer ring on top of the microphone capsule with tire-repair rubber cement. Without this spacer, reproduction quality is unsatisfactory. After pushing the O-ring all the way down, I pulled it back up a little to give the film some slack; this also seems to help with sound fidelity
The battery is lithium-ion type, 9V. It connects to the intercom box with a 2.5"-long adapter (Kitronik 4132, Sparkfun 09518, or similar) that plugs into a 5.5 × 2.1 mm, center-positive barrel jack on the box. In use, the battery just hangs from the short adapter. When not in use, I unplug the battery plus adapter and put it in my pocket. The power seems to be good if I can hear a harsh noise burst on the "s" and "sh" sounds. (I understand that this is caused by the unavoidable nearness of my mouth to the mic.) If you do not hear this, try a new battery; this can markedly improve sound fidelity.
Another problem with the P100 is that there is no filtration on the exhalation valve, so others are not protected if the wearer is infected by SARS-CoV-2 in spite of taking unusual precautions. My build mitigates this by placing the intercom box over the exhalation valve to deflect air sideways into a cloth face-mask folded twice lengthwise and tied around the “waist” between the box and the mask. The cloth covers only the top and sides and is tied under the bottom. Unfiltered exhalation air is therefore directed downward, as with a face shield. This keeps the exhalation back-pressure low so as to avoid impairing understandability (otherwise, the nasals seem to drop out).
A problem I am living with is that there is now no way to perform a pressurization test upon donning the mask; I have to be content with suction tests only. (An obvious solution would be to replace the cross-shaped spacer with a ring-shaped spacer, and leave a gap in it at the bottom that can be stopped with a finger.) I have puffed finely ground stevia sweetener around the mask while inhaling through my mouth, with good results–almost no sweet taste. Huge taste without the mask.