Post-submission Field Notes

A project log for A P100 Mask Intercom

COVID-19 PPE taken to the next level. Anticipated to be good against aerosol transmission.

David Matthew MooneyDavid Matthew Mooney 10/05/2020 at 18:490 Comments


I have a tendency to lean in to my listener if I am not being understood while using the intercom, which breaks social distancing rules. Turning down the volume usually works better because that gives less distortion. I have to remind myself that the intercom is not a bullhorn, and it is not karaoke, it is just a voice replacement and it doesn't have to be any louder than my normal voice.


I am treating all condensate that forms inside the P100, which is plenty, as if it were full of SARS-CoV-2. Options seem to be: after use, drain mask with due care, blot the inside, dispose of the wet paper towel carefully, spray mask inside and out, and hands, with A) 70% isopropanol, or B) 99% isopropanol.


We have integration!

And the first field test thereof: A trip to the local cop shop by public transport to file a crime report ("No incident should be considered too small to report."), but that is a story for another day. My thoughts:

1) Intercoms are not magic. It seems that even the commercial units have limitations. When using this intercom, you need to speak up and you need to speak slowly and clearly. I understand from reddit that Walmart employees should do the same.

2) Now that the cold weather is here, condensate is rapidly forming on my cloth mask, making exhalation difficult. When that happens, I just slide the mask up or down to put a dry spot over the end of the exhalation tube. Five minutes later, it's time to shift it again. This is only an issue in cold but crowded places like bus shelters, where I can't simply remove the outer mask.  We'll see what a January in this town throws at me.


The tube for the exhalation valve kept falling out, so I fixed it as shown. The white blob is putty, which I added to prevent any air leakage. (The blob on the near side cannot be seen in this shot.) The wire is 20-gauge, tin-plated copper (header wire). Holes were drilled with a 3/64" twist drill. I am solving the understandability problem and the exhalation-filtration problem on separate masks, and will eventually integrate the results.

When my doctor saw me wearing the dual mask arrangement, he said, "I like your mask." Twice.

I got a COVID-19 exposure alert on my phone this month and had to go and get tested, then self isolate until the results were ready. Result: Negative. Thanks to Bill & Sal for bringing me groceries.

I have tried 3 different cloth masks with the exhalation tube. Two were OK but the third offered too much resistance to airflow. One needed a rubber band to cinch it around the tube, two didn't.


False alarm. People can only understand me while I wear a P100 mask with the intercom switched off if I am saying something they hear every day, such as "Bag with that," or "Medium, double-double." If I try to say something unusual, such as "What do you think of Tolkien?" I am likely to hear: "Couldn't say, I've never toked."


Lately, it seems that persons who deal with the public occupationally can understand me OK with the intercom switched off. How could I be so mistaken? I have clear memories of people telling me I sound like I'm speaking Mandarin if I tried to talk (English) while wearing a respirator. I now feel like a fool but I will be looking into this phenomenon. 


The photos show how I am filtering the exhalation air from a P100. There is obvious backpressure on exhaling, but that just pressure-tests the mask + new filter with every breath. It's not disabling; I can climb 2 flights of stairs without stopping or getting winded. Moderate coughs can be filtered if the straps are tight, but not violent coughs or sneezes, which leak up the sides of the nose. Fortunately, I don't cough or sneeze a lot.

10/19/2020: The gallery photos have been updated to show the new, tripartite intercom design.


A) I have decided to redesign the project to mechanically separate the intercom box from the mask and to connect them with a short mono audio cable. This will recover the ability to do a pressure test and to properly clean the mask once in a while. This will also give me more options for thermostatting the circuitry with my body heat in the winter. 

B) Without the intercom box present, a cloth mask can be worn over the exhalation valve to provide filtration (I have done it for the past two days with my spare P100), and it may seal better to the end of a simple, short tube than to the contours of a human face. Right now, this seal seems dicey because of interference from the filter cartridges, but it would be less dicey with a longer tube that pushes on the cloth mask more firmly. An exhalation-valve extender that friction-fits to the existing valve would make a great 3D printing project, but how long should it be? [20-10-2020: This extender would not be challenging enough to make a good 3DP project. All you need is a tube with an outside diameter of 1.5". I am assuming a model 9200, Workhorse brand.] Ideally, the tube should be made of the same plastic as the valve, to prevent the friction fit from loosening due to differential thermal expansion/contraction due to seasonal changes. I'd guess the valve is made of acrylonitrile/butadiene/styrene (ABS).


I am still undecided on what approach to use to clean up the exhaled air stream. We are going after one-half micron particles with only exhalation force as the power source, with as little back-pressure as possible, and a compact, rugged design. Options seem to be: filter, Andersen Impactor, cyclonic cleaner, chemicals, and bubble tube. Some impactor effect is probably present already in the two right-angle turns the air must make while escaping.


Another sub-project will be to thermally insulate the intercom box and the gap between it and the mask. This should reduce the variation in amplification due to changes in the weather that affect the temperature of the electronics. These changes can take you from understandable to not understandable from one day to the next. This can be fixed by readjusting the volume knob, but who wants to do that a lot? Electronic (partial) solutions such as nulling the LM741 would be something else to try.


I am currently designing a lucite box to go around the intercom-mask junction and direct all exhaled air out one circular opening facing downward. This will immediately allow pressure testing of the mask fit and someday soon, the connection of an exhaled-air filter to the mask. Following the KISS principle, I had hoped that a simple cotton ball stuffed into a tube would suffice for the filter, but no luck. Cotton balls in tubes are remarkably hard to exhale through.


Since I keep learning new, share-worthy things about this project due to the fact that I am using it every day, I decided that this project is "ongoing," not "finished," as originally advertised.

The gain in the first stage of the circuit may have to be adjusted. While indoors, I set the volume-control knob to mid-range, then adjust the first-stage feedback trimmer-pot to get best volume + excellent fidelity.


Spend some time getting the volume right; more is not always better if it leads to distortion.  The word "hour" is especially hard to get out clearly, but if you can, you're good to go. 

Reproduction fidelity continues inconveniently low under adverse but common conditions of background noise. Can someone point out a way to modify my circuit or physical build for higher fidelity? Am I flogging a dead horse in sticking with analog?

When you go out in this mask, take a conventional cloth mask with you in case of a malfunction or an encounter with a situation where only conventional masks are permitted or a situation of too much background noise. 

Op amps are temperature-sensitive; expect to have to turn up the volume on hot days to be understood and to have to turn it back down again on cold days to avoid feedback.


 I added a bypass capacitor to the LM386 chip to increase the amplification ten-fold, which is essential for dealing with noisy environments with even spotty success. The circuit diagram has been updated.