Industrial Machinery

The 7-Step Checklist I Run Before Signing Any Compressed Air, Nitrogen, or Vacuum System Order

Lucian VargaOctober 29, 20265 min read8 views

At a glance

  • A field-tested seven-step pre-purchase checklist for industrial compressed air systems, PSA nitrogen generators, and vacuum equipment — built from six years of specification mistakes, plus the cost items that never show up on the unit-price line.

Who this is for, and what it covers

If you're the one who has to defend the purchase order for an industrial compressed air system, a PSA nitrogen generator, or an industrial vacuum system, this checklist is for you. Seven steps, in the order I actually run them. Same list whether you're buying one portable electric air compressor for a construction site or a 250 kW oil-free screw compressor feeding a pharmaceutical line — the depth changes, the steps don't.

What it doesn't cover: installation, commissioning, and long-term maintenance planning beyond the service-contract question. That's a different list.

Two things before you start. First, I'm not a neutral party on the price question — I think unit price is the least useful number in any quote, and by the end of this you'll see why. Second, I've made most of these mistakes myself, so the warnings are specific rather than theoretical.

Step 1 — Write the demand profile, not the equipment name

Most enquiries I get open with a product category: we need a rotary screw compressor, or we need a centrifugal compressor. That's an answer, not a requirement. Start from the demand instead.

Six fields, minimum:

  • Flow — in Nm³/min or cfm, at the reference conditions you'll fix in step 2.
  • Pressure — at the point of use, at the far end of the pipe run, not at the compressor outlet.
  • Duty cycle — hours per day, days per week, and the simultaneity factor. This is where projects go wrong quietly: the summed nameplate demand of the tools is often 2–3× the real simultaneous draw, and nobody checks.
  • Ambient conditions — altitude, summer ambient temperature, whether the room is ventilated.
  • Air quality requirement — see step 3.
  • Growth — what you expect to add in three years. Write it down now, because nobody remembers it at re-quote time.

I went back and forth between a reciprocating unit and a screw compressor for a low-duty application for about two weeks in 2019. The piston reciprocating compressor was cheaper to buy and cheaper to service at that duty. The screw was quieter and simpler to operate. I went with the reciprocating machine because the duty really was intermittent — roughly 40 minutes a day, five days a week — and buying a screw for that profile is buying convenience you never use.

Step 2 — Put every quote on the same reference conditions

This is the single step that kills bad comparisons. Two vendors quote 20 m³/min. One measured it at 7 bar with a 25°C inlet; the other at 7.5 bar with a 40°C inlet in a hot plant room. Those are not the same machine, and the difference can be 8–10% of delivered flow.

Ask for the data sheet format defined by ISO 1217 (displacement compressors, acceptance tests) and, where the supplier is a member, the CAGI data sheet format. Both force the vendor to state the conditions along with the number. If a quote gives you cfm with no conditions, that's not a specification, it's a marketing line.

Also ask, in writing: is that flow at full load only, and what does the unit deliver at 60% and 40% load? A quote that can't answer the part-load question is a quote for a machine nobody has characterised.

Step 3 — Translate air quality into a standard class

Oil-free is a word, not a spec. Neither is clean, dry, or instrument quality. Convert it.

ISO 8573-1 defines air purity by class for three contaminants: particles (count and size range), water (pressure dew point), and oil (mg/m³). Class 1 on oil means ≤ 0.01 mg/m³. Class 0 is not a purity level — it's a placeholder for whatever the equipment supplier and the customer agree on in writing, which means if there's no written agreement, the class is meaningless.

I learned that in my first year, 2017. I ordered two machines both described as oil-free, assumed the phrase meant the same thing on both datasheets, and approved the orders. It didn't. One unit's filtration package was scoped to a defined oil-content class at the point of use; the other stopped at the compression element, with dryer and filters excluded. The customer's audit caught it on measured air quality, not on paper. Filter retrofit plus re-testing came to about $2,400 and a lost week. Lesson: the number that matters is at the point of use, and it needs a standard reference behind it.

For the pharmaceutical industry specifically, traceability matters as much as the class. 21 CFR Part 211 requires that equipment surfaces contacting product not be reactive, additive, or absorptive, and that lubricants and coolants not contact product — that's a design and documentation question you want answered before the order, not during the audit.

Two extra rows for a PSA nitrogen generator

Nitrogen generators break the shared checklist in one specific way: purity and flow trade against each other on the same membrane or sieve bed. So quote both, always, together. Fields to add:

  • Purity as residual oxygen (ppm or %), not as a general purity percentage.
  • Flow at that purity, and the turndown range.
  • Dew point, and whether the quoted figure is at pressure or atmospheric.
  • Carbon molecular sieve or membrane, and the expected change-out interval.

I've watched a nitrogen gas generation system get signed off against a flow figure quoted at a lower purity than the process needed, because the two numbers sat in different columns of the same brochure.

Step 4 — Size the dry end and the storage, and mind the dew point basis

Dryers and receivers get treated as accessories. They aren't — they set the actual delivered quality.

Here's the step most people skip: pressure dew point and atmospheric dew point are not the same number. A dryer rated at −40°C pressure dew point is not delivering −40°C air at atmospheric pressure. If your process or your instrument air standard specifies a figure at atmospheric conditions and you've been comparing pressure-referenced ratings, you may be off by 15–20°C. Convert before comparing, and make the vendor state which basis they used.

Receiver sizing: ask what buffer the system needs to ride through a compressor unload or a demand spike without dropping below your minimum pressure. If nobody has calculated it, the answer is whatever tank was in the last project.

Step 5 — Price the energy before you price the machine

This is where total cost thinking stops being a slogan. For an industrial installation running multi-shift, the electricity to make the air will usually dominate the ten-year cost of ownership — often several times the purchase price. A machine that's 5% less efficient at your actual duty can cost more in eighteen months than the price gap between the two quotes.

The assessment method exists and is published: ISO 11011 covers energy efficiency assessment of compressed air systems, and the US DOE's sourcebook (Improving Compressed Air System Performance) walks through the supply-side and demand-side losses. Neither one is exotic. Both require the vendor to give you specific power (kW per m³/min) at your operating points, not just at best point.

I now build a small sheet before comparing any vendor quotes: unit price, freight, duties, installation, commissioning, first fill, dryer and filter power, and estimated annual kWh at my duty profile. The 15% cheaper quote has lost that comparison more than once.

Step 6 — Price the service and the spares, not the warranty headline

Ask three questions, and get written answers:

  1. What's the guaranteed response time for a breakdown, and is it in the contract or just an email?
  2. What's the lead time on the wear parts — filters, separator elements, valve kits, sieve charge? A two-week element lead time is a two-week outage.
  3. What does the five-year parts and labour total look like at my duty cycle?

Oil-free screw compressors and oil-flooded machines have very different service economics, and the gap widens on high-pressure hydrogen compressor work where seal and packing intervals drive everything. Same for vacuum: a vacuum pump supplier's service network matters more than the pump curve, because vacuum equipment tends to fail at the worst possible moment.

One extra row if you're buying vacuum

Ultimate pressure and pumping speed are not interchangeable across vacuum pump technologies. A figure in mbar without a stated pumping speed at that pressure tells you almost nothing. Ask for the curve, not the point.

Step 7 — Make the exclusions as explicit as the inclusions

Before signature, the scope sheet should list what is not included, in the same font size as what is. The recurring ones: freight and duties, offloading, foundation or pad, pipework and headers, condensate drainage, electrical supply and isolation, commissioning, operator training, and extended warranty.

The numbers said take the cheaper quote on one 2022 project — 12% lower, comparable spec on paper. Something about the response times bothered me. I asked both vendors the same seven questions from this list and only one answered all seven in writing. Went with that one at the higher price. Eight months later the other supplier's reference site was still waiting on a separator element, and I got to feel briefly clever. That's the whole test, really: whoever fills in the sheet gets the order.

Common errors, roughly in the order I made them

  • Comparing quotes at different reference conditions. Costs you 8–10% of flow without a single visible error on the paperwork.
  • Treating oil-free as a specification. It's a claim. Ask for the ISO 8573-1 class at the point of use.
  • Confusing pressure dew point with atmospheric dew point. The quiet one. Nobody catches it until an instrument air audit does.
  • Using summed nameplate demand as the flow requirement. Leads to oversized machines running at 30% load, which is the least efficient place a fixed-speed compressor can sit.
  • Buying on unit price. The cheapest quote is frequently the most expensive system, and the gap shows up in year two.
  • Assuming centrifugal and reciprocating behave alike at part load. They don't — different turndown, different control, different economics at your duty.
  • Leaving the exclusions verbal. Anything not written down is, in practice, not included.

Seven steps, one sheet, about ninety minutes of work per project. It has saved me considerably more than ninety minutes.

Reported by Lucian Varga, Casting and Forging Process Editor.