Where bad air comes from, and the engineering that fixes each cause

1. Intake siting — the cheapest control and the one most often skipped

Code requires NFPA 99 §14.2.4.2.2 requires compressor intakes to be located away from air contaminated by exhaust from vehicles, internal combustion engines, stationary engines, or building exhaust outlets. The paragraph is carried into accreditation criterion HBOV 3.0, which cites it explicitly (UHMS accreditation manual, 4th Ed.). We paraphrase rather than reproduce the code text, which is copyrighted.

Body recommends A Navy facilities maintenance manual turns the same requirement into two inspection questions you can ask on a walkthrough — "are compressor air intake lines piped to the weather and provided with an air intake filter?" and "are all air systems filtered before reaching the occupant/divers?" (NAVFAC MO-406, Appendix F items 11 and 2). Those two questions catch most of what goes wrong in a small installation.

Why it matters in practice, from an incident-analysis source rather than a sales page: contaminant sources found near compressor intakes include paint, cleaning fluids, cooking exhaust, rotting organic matter, and a petrol engine or generator powering the compressor itself. Any detectable carbon monoxide from a properly installed and maintained compressor indicates contamination by combustion gases or another source, and a 16 ppm cylinder reading "likely indicates that the compressor's catalytic converter is being overloaded by a significant amount of CO near the compressor intake" (DAN, Alert Diver, "Carbon Monoxide in Breathing Air"). That is a diving cylinder-fill case, not a chamber case — we say so on the FAQ — but the compressor physics is identical.

Vendor asserts At least one manufacturer markets a remote air intake option on exactly this basis, offering it "to meet OSHA 29 CFR 1910.430(b)" (Nuvair Q-5120). That is the manufacturer's characterisation of the regulation, not a quotation of it, and 1910.430 is the commercial diving standard rather than a hyperbaric provision. The hardware idea is sound regardless: move the intake to clean air rather than filter dirty air.

2. Oil carryover — how much oil a compressor actually delivers

This is quantified in manufacturer technical literature, which is what makes it usable instead of rhetorical (Parker Balston, Coalescing Compressed Air and Gas Filters):

• A well-maintained reciprocating compressor typically delivers 15–30 ppm liquid oil.
• An oil-sealed rotary screw compressor delivers 10 to more than 100 ppm, depending on the efficiency of its bulk oil separator.
• Condensation of oil vapour can itself produce approximately 0.2 ppm liquid oil.

The mechanism most often missed is oil vapour, because vapour passes straight through a coalescing filter. Its loading depends on discharge temperature, at 100 psig:

The stated consequence: where trace oil vapour is a problem, three-stage filtration is required — a Grade DX stage, then a Grade BX stage, then a Type CI activated-carbon cartridge, which achieves "0.003 ppm by weight maximum remaining oil content within a 0.05 ppm inlet challenge" (Parker Balston). Note the shape of that claim: an outlet figure at a stated inlet challenge. A filter rating with no inlet challenge attached is not a specification.

3. Filtration ratings that come with a test method

The better datasheets state the standard each number was measured against, which lets you compare two products honestly. This one does (Parker P3T series datasheet):

4. The 66 micron filter is a trash screen, not purification

Code requires NFPA 99 Chapter 14 requires a particulate filter of 66 micron or finer, per criterion HBOV 14.0, which also records at HBOV 14.1 that the filter is to be per ANSI/ASME PVHO-1 and "located as close to the source as practical" (UHMS accreditation manual, 4th Ed., citing §14.2.10.3).

Now put that number next to the table above. 66 microns is 6,600 times coarser than the 0.01 micron coalescing stage. A 66 µm screen stops rust flakes, thread sealant, swarf and insects. It does not touch oil aerosol, oil vapour, water vapour or carbon monoxide — the four things a breathing-air specification is actually about.

So a page that presents a 66 µm filter as the air-quality solution has misread the code. The 66 µm provision protects the chamber and its plumbing from debris. The air quality work is done by the coalescing, adsorption and drying stages, which the code addresses in a different way entirely: by naming an outcome, Grade E, and requiring a laboratory sample twice a year to prove you reached it. Both requirements are real. They are not the same requirement.

5. Moisture and dew point

Code requires For supplied-air respirators, the dew point must be 10 °F below ambient temperature at one atmosphere for compressor-supplied air, and −50 °F for cylinder air (29 CFR 1910.134(i)(5)(ii) and (i)(4)(iii)). Note that the first of those is relative to your own ambient temperature, so it is a different number in a Florida garage than in a conditioned equipment room.

Body recommends For hyperbaric service specifically, a UHMS safety article gives water limits by storage pressure — below 15 bar, 40 mg/m³ (500 ppmv, −27 °C); 40–200 bar, 50 mg/m³ (62 ppmv); above 200 bar, 35 mg/m³ (44 ppmv); to cylinders, 25 mg/m³ (31 ppmv) — states that chamber relative humidity is ideally 50–60 %, and gives an achievable practical limit of −48 °C or ≤35 mg/m³ (Burman, UHMS, §§4.1.3 and 5.1).

Two very different pieces of hardware answer this, at very different cost points. An integrated refrigerated dryer on an oil-free scroll package is published at ISO 8573-1 Class 4, pressure dew point 3 °C / 38 °F (Atlas Copco SF/SFe leaflet). A duplex desiccant dryer is published at −20 °F for laboratory service, with purge-saving control switching on at −15 °F and off at −10 °F (Powerex scroll medical air manual).

Which do you need? Work backwards from the two rules above rather than from a catalogue. A 38 °F pressure dew point comfortably satisfies "10 °F below ambient" in any normal conditioned space and keeps liquid water out of the chamber. A −20 °F desiccant answer buys margin you only need if you are filling cylinders, running in genuinely cold ambient, or working to the tighter society figures above. Remember that Grade E sets no moisture limit at all — see the grades page — so no one can tell you a dew point is required "by Grade E." It is required by your application.

6. Carbon monoxide and the catalyst stage

The chemistry, from two independent technical sources. Hopcalite is "a mixed oxide of copper and manganese, commonly identified as CuO/MnO₂"; the catalyst "oxidizes carbon monoxide" and "converts poisonous carbon monoxide (CO) into the much less hazardous carbon dioxide (CO₂)" (3M, "Gold Catalyst Technology" technical abstract). The same abstract describes hopcalite being placed upstream as a protective bed that sorbs sulphur dioxide, hydrogen sulphide, hydrogen cyanide and ammonia, and states that catalyst life "may be limited by accumulation of catalyst poisons," specifically H₂S and SO₂.

The operating limits matter more than the chemistry. "Properly sized filters usually have CO overload limits of 50 to 100 ppm at the compressor intake" — above that, CO passes through. CO-conversion capability decreases in humid environments; the conversion reaction generates heat in proportion to the CO load, and excess heat degrades the catalyst further. The expected outlet result from a working catalyst is 0 ppm (DAN).

Vendor asserts A combined cartridge containing activated carbon and a CO catalyst is commercially available, described as removing "carbon monoxide, hydrocarbons, oil, taste, odor, and particles," with a maximum useful lifespan of six months, a two-year shelf life, and a published price of $163.20 (Nuvair hopcalite and carbon filter). Read the fit before you budget from it: that is a high-pressure canister cartridge, for a CAN-35 canister and specific fill units, not a low-pressure panel element. We could not locate a low-pressure-panel CO catalyst datasheet that publishes a rated service life at a stated flow, so for LP panel service that figure is Not published.

7. What the code side of this looks like as a checklist

Each row is keyed by the accreditation manual to its NFPA 99 Chapter 14 paragraph (UHMS, 4th Ed.).

One deletion worth knowing about: the Fourth Edition deleted HBOV 12.1, which had required continuous monitoring "for volatized hydrocarbons and carbon monoxide at a location downstream from the oil filter when the compressor is running" (UHMS, 4th Ed.). Anyone telling you continuous downstream hydrocarbon monitoring is a current accreditation criterion is citing a withdrawn item. Continuous CO monitoring for oil-lubricated compressors remains an OSHA duty under 1910.134(i)(7) for respirator service.