The Dryer Types and What Each One Can Actually Deliver
Why this matters
A compressed air dryer is selected on flow and price far more often than it is selected on the only specification that decides whether it works. Each family removes water by a different physical mechanism, and each mechanism has a floor on the pressure dew point it can reach that no amount of maintenance, oversizing or better ambient will move. Get the family wrong and you have bought a machine that cannot do the job on its best day, and every service call afterwards is spent tuning something that was never the problem.
Before you open any dryer
A dryer vessel holds stored pressure energy after the compressor stops. Close its upstream isolation, open the vent, confirm the vessel gauge reads zero with the vent still open, and lock and tag that isolation under 29 CFR 1910.147 before a fitting is broken.
On a twin-tower desiccant machine that is not enough. The offline tower is repressurised automatically by the switching valve on its own cycle timer, so a tower that reads zero can be brought back to full line pressure without a person touching anything. Isolate the tower from the switching valve as well as from the header, put the controller in a state that cannot switch, and verify zero at that tower's own gauge with its vent still open.
Changing a desiccant bed releases respirable media dust when the bed is opened, dumped or refilled. That is an inhalation route, so it needs a respirator selected and fit-tested under the employer's respiratory protection programme required by 29 CFR 1910.134, not a nuisance dust mask and not gloves.
A deliquescent bed produces a corrosive brine. Handle it with the chemical-resistant gloves and eye protection its safety data sheet specifies, and route disposal through the SDS and your local requirements rather than into a floor drain.
A refrigerated dryer contains a sealed refrigerant circuit. Opening it is refrigerant work performed by a technician certified for that refrigerant under 40 CFR Part 82, Subpart F, and the refrigerant is recovered, never vented.
The mechanism sets the floor
Read the table by the middle column, because that is the constraint. Everything else is a preference.
| Family | Mechanism, and the floor it sets | What it costs you besides input power |
|---|---|---|
| Refrigerated | Cools the air so water condenses on a surface and drains off. The surface cannot be driven below freezing without icing and blocking, so the floor sits above the freezing point of the condensate. | A pressure drop across the heat exchanger; nothing else once running. |
| Deliquescent | A consumable salt bed dissolves as it absorbs water. It suppresses dew point by a number of degrees below the inlet temperature, so it has no absolute floor at all. | Continuous consumable replacement plus corrosive brine handling and disposal. |
| Membrane | Water vapour permeates out through hollow fibre walls, swept away by a small stream of already-dried product air. Reaches low dew points, and the lower you set it the more sweep it takes. | Sweep air, which rises steeply as the target dew point falls. |
| Regenerative desiccant | Water adsorbs onto a solid bed; the offline tower is regenerated and the towers swap. Reaches the deepest dew points in normal industrial use. | Purge air, or heat, or both, depending on the regeneration method. |
Two consequences fall straight out of that column.
A refrigerated dryer cannot be made to reach a below-freezing pressure dew point. Not with a bigger compressor on it, not with a cleaner condenser, not in a cold plant room. Its mechanism is condensation on a surface and that surface would ice over. Take the actual rated floor from the manufacturer, because it varies with the machine, but the direction of the limit is physical rather than commercial.
A deliquescent dryer's output is relative, not absolute. It gives you a suppression below whatever the inlet temperature happens to be, so its delivered dew point rises and falls with the inlet all day. That makes it usable where the requirement is also relative, such as knocking down bulk moisture ahead of something else, and unusable anywhere a fixed pressure dew point has been specified.
Regeneration is the whole cost question on a desiccant machine
Regenerative desiccant dryers split by how the offline tower gets dried, and the split is worth more than the badge on the front.
Heatless regenerates by expanding dried product air through the offline bed. Simple, no heater, and the purge is a fixed fraction of throughput. That fraction is a fraction of the dryer's RATED flow, not of what is actually passing through it, which is where this bites: a heatless machine oversized by a factor of two purges 30 percent of the air it is actually delivering. The fraction is commonly in the region of 15 percent, but it moves with inlet pressure and cycle time, so take it from the manufacturer's data rather than from a rule of thumb. Purge fraction is the reason not to oversize a desiccant dryer.
Heated purge and blower purge replace some or nearly all of that purge air with electrical heat, which trades a compressed air demand for a direct electrical one.
Heat of compression uses the compressor's own discharge heat to regenerate, so purge approaches zero. It only exists as an option where the compressor is oil-free and its discharge is hot enough to do the regenerating, which is a condition on the compressor and not a feature you can add to a dryer.
Purge is not overhead sitting to one side of the system. It is a demand line the compressor has to carry, and treating it as anything else is what makes a plant look short of air. Its accounting belongs to its own article.
What moves a delivered dew point away from a family's floor
The floor is what the family can do. What it actually delivers on your site is worse than the floor whenever any of three conditions runs hot or heavy, and all three peak in the same hour on a summer afternoon at shift change: more flow through it than its rating, hotter air arriving at its inlet, hotter ambient around it. Manufacturers publish capacity at stated reference conditions with correction factors for each, and a dryer sized on rated flow without applying the corrections for your inlet and ambient is undersized in the season that matters. Measuring what it delivers, at line pressure and under load, is its own procedure.
Worked example: one plant, two requirements
A plant runs 200 scfm total. Fifteen of those scfm feed an instrument-air branch supplying pneumatic control valves on an outdoor rack; the rest is a general shop header inside a heated building.
Set the requirements from the pipe, not the machine. The coldest pressurised metal on the outdoor rack is 10 F in the design winter week, so with a common design margin of 18 F below the lowest metal temperature the outdoor branch needs a pressure dew point of 10 - 18 = -8 F. The coldest pressurised metal indoors is 62 F, so the shop header needs 62 - 18 = 44 F.
Family selection. The 44 F requirement is comfortably inside what a refrigerated machine reaches. The -8 F requirement is below freezing, so refrigerated is out for that branch by mechanism, and deliquescent is out because it has no absolute floor to specify against. That leaves regenerative desiccant or membrane for the outdoor branch.
Two architectures, and the arithmetic that separates them.
Option A is one heatless desiccant dryer on the whole 200 scfm. At a purge fraction of 15 percent of rated flow taken from the data sheet, purge is 0.15 x 200 = 30 scfm, which is 15 percent of the plant's total flow. That is twice the entire outdoor branch's 15 scfm demand, spent to give -40 class air to a shop header that needed 44 F.
Option B is a refrigerated dryer on the full 200 scfm, with a small heatless desiccant unit on the 15 scfm branch downstream of it. Purge is 0.15 x 15 = 2.25 scfm, which is 1.1 percent of the plant total.
30 divided by 2.25 is 13.3, so option A spends more than thirteen times the purge air of option B to satisfy the same two requirements. Against a 200 scfm machine, option A's purge alone is a permanent demand line larger than the branch that caused it.
What flips this. The split stops paying as the deep-dew-point fraction grows. Rerun it with 120 of the 200 scfm on the outdoor rack: option A still purges 30 scfm, option B purges 0.15 x 120 = 18 scfm, and the ratio falls from 13.3 to 1.67. At that point you are maintaining two dryers, two sets of filtration and two failure points to save 12 scfm, and the single machine is the better answer. The break is not at a fixed percentage; it is where the purge saving stops covering the second machine's complexity, and that is a judgement the plant makes with its own maintenance capacity in front of it.
It flips completely if the plant runs an oil-free compressor hot enough for heat-of-compression regeneration, because then option A's purge is close to zero and the reason to split disappears. That option exists only with that compressor, so it is a decision that had to be made when the compressor was bought.
The failure mode. A plant that specifies to its hardest branch and installs one desiccant dryer for everything does not see a fault. It sees a plant that is mysteriously short of air, because 15 percent of the compressor's output now leaves through the purge muffler, and the next capital request is for a bigger compressor. The dryer is working perfectly and the sizing is wrong one level up.
How to verify you picked correctly
Check the requirement first, because a wrong requirement makes a right selection wrong. Confirm the coldest pressurised metal came from a pipe walk in the coldest week you have data for, that the margin has its source named, and that it covers every branch, including the one that leaves the building and the one that was added after the dryer was bought.
Then confirm the selection was made against a corrected rating rather than a nameplate flow. Write down your actual inlet air temperature, ambient and flow, apply the manufacturer's correction factors, and check the corrected capacity still clears your requirement at the worst hour rather than at the reference conditions.
Then measure what arrives. A dryer selected correctly and installed correctly can still fail the gate because of what is upstream of it: a drain at the wet receiver that has stopped passing water sends the aftercooler's share of the load into a dryer that was never asked to take it, and that arrives as a dew point problem with no dryer fault behind it.
References
- 29 CFR 1910.147, control of hazardous energy, for isolating and verifying depressurisation of a dryer vessel, including isolation from an automatic switching valve on a twin-tower machine
- 29 CFR 1910.134, respiratory protection, for respirator selection and fit testing where desiccant media dust is disturbed
- 40 CFR Part 82, Subpart F, for technician certification and refrigerant recovery on a refrigerated dryer's sealed circuit
- Dryer manufacturer data for rated pressure dew point, reference conditions, correction factors and purge fraction; safety data sheet for deliquescent media and its brine
- See related: What a Desiccant Dryer Costs You in Purge Air; Why a Refrigerated Dryer Fails Quietly; What Pressure Dew Point Means and Why It Is the Specification