15.2 Crankcase Pressure Regulators and Gas-Pumped CPR Systems

15.2 Crankcase Pressure Regulators and Gas-Pumped CPR Systems

A crankcase pressure regulator (CPR) is the opposite sensing job from a BPR. Catalogs call it an outlet-pressure regulator or a hold-back valve. It sits in the suction line near the compressor and closes as downstream pressure rises, so the crankcase cannot exceed a maximum. Evaporators and suction headers upstream of the CPR are allowed to be higher during a hot pull-down. The motor is not.

CIRO also uses the same pressure-powered thinking on gas-pumped overfeed and vessel transfer: a regulator sets the gas pressure that pushes liquid, instead of a mechanical pump impeller. The formula sheet then asks you to think about valve capacity versus ΔP, not to design a piping isometric.

Why pull-down overloads a compressor

Compressor mass flow is roughly volumetric displacement × density at the suction flange (times volumetric efficiency). Warm rooms, post-defrost coils, and start-up after a weekend leave evaporators at high pressure. Density is high. The machine tries to pump a heavy vapor. Brake horsepower and amps climb. A motor sized for design suction (a freezer at 9 psig, a cooler at 20 psig) was not sized to sit at 50–70 psig suction indefinitely.

A CPR throttles the suction so crankcase pressure — the pressure the compressor actually inhales — stays at a set maximum. As the box pulls down, upstream pressure falls, the CPR opens, and eventually it is wide open and out of the way at design conditions.

Reciprocating compressors are the textbook CPR application: displacement is nearly fixed, and high density becomes high mass and high BHP immediately. Screw compressors often unload with a slide valve or VFD, so they have another way to shed load — but high suction still raises mass flow at a given slide position. Some packages still use suction throttling. Read the stem: if it says motor overload on pull-down and suction regulator, it wants a CPR, not a BPR.

Setpoint and failure modes

Set the CPR from the compressor manufacturer’s maximum suction for that motor and application, then confirm amp draw during the worst pull-down you actually run. Too low a setpoint starves the compressor after it should already be at design suction: long pull-downs, low mass flow, possible oil-return problems. Too high a setpoint (or a valve stuck open) is no protection: the motor runs into overload, overloads trip, or windings cook.

A CPR is not a capacity-control strategy you leave throttling all day at design. Continuous throttling is a suction penalty: extra ΔP, higher specific volume at the compressor than at the evaporator, wasted BHP. If the plant always needs the CPR to be in the yellow, the compressor is undersized for the suction you are asking, or someone is using a CPR to hide a slide valve or VFD that is not working.

Gas-pumped overfeed and vessel transfer

Mechanical liquid pumps (Chapter 18) move overfeed with an impeller and a motor. Gas-pumped systems move liquid with pressure on a liquid surface — a transfer drum, pumper drum, or pair of vessels — and a set of fill, gas-supply, vent, and liquid-outlet valves.

A typical cycle:

1. Fill. The transfer vessel is vented to the LPR or separator so pressures equalize. Liquid gravity-fills (or equalizes) into the drum. No pump NPSH, no rotating seal.
2. Isolate. Fill and vent valves close.
3. Push. High-pressure gas (high-side vapor, hot gas, or an intermediate pressure) is admitted onto the liquid surface through a pressure regulator. That regulator is the CPR-type control in this application: it sets the pumping gas pressure, which sets how hard you push, which sets the ΔP available to move liquid to the evaporators or the receiving vessel.
4. Empty. Liquid leaves the bottom through a check and the liquid header.
5. Vent down. Gas supply stops; the drum is vented back to the LPR or suction so it can fill again.

Two drums staggered in phase give more nearly continuous overfeed. Oil pots and dump traps use the same gas-pressure transfer idea on a smaller scale.

Why a regulator on the gas, not a wide-open hot-gas valve. The destination is a low-side evaporator or vessel. Unregulated discharge pressure on a transfer drum can overpressure the low side, slam liquid into evaporators, and lift internal or ASME relief. The CPR-type gas regulator holds a maximum pumping pressure: high enough to overcome elevation, friction, and destination pressure, low enough not to treat the liquid header like a high-side pipe.

Mechanical pump versus gas pump is an exam contrast, not a brand war:

• Mechanical pump: continuous flow, motor kW, NPSH, seals, cavitation, overfeed ratio set by pump curve and valve position.
• Gas pump: no impeller, cycling vessels, uses high-side gas energy, extra valves, pressure cycling of drums, flow set by gas pressure minus destination pressure (the ΔP in the Cv formula).

Modern overfeed plants are mostly mechanical pumps. Gas-pumped drums still appear on oil transfer, older recirculators, and CIRO stems that want you to name the pumping force.

Cv = GPM √(SG/ΔP) — capacity versus ΔP, not a hydraulic project

The CIRO formula sheet includes the liquid sizing relationship

Cv = GPM × √(SG / ΔP)

which rearranges to

GPM = Cv × √(ΔP / SG)

Teach three operator facts, not a full line-sizing study:

1. For a fixed valve (fixed Cv) and a fixed SG, GPM follows the square root of ΔP. If available ΔP falls to one-fourth, flow falls to one-half, not to one-fourth. If you double ΔP, flow rises by √2 ≈ 1.41, not by 2.
2. If you need the same GPM at a smaller ΔP, you need a larger Cv. Gas-pumped and gravity systems often have little extra pressure. A valve that was “fine” on a high-head mechanical pump can be too small on a drum that only has a few psi of push.
3. Pipe size is not Cv. A 2-inch valve and a 2-inch pipe are not interchangeable numbers. The formula uses Cv and ΔP.

Worked example — same gallons, two available drops. A transfer valve must pass 40 GPM of liquid ammonia. Take SG = 0.65 as a round exam value.

If the gas regulator holds a pumping pressure that leaves ΔP = 4 psi across the valve:

Cv = 40 × √(0.65 / 4) = 40 × √0.1625 = 40 × 0.403 = 16.1

If the destination pressure rises, or the gas regulator is set timid, and only 1 psi remains across the same duty:

Cv = 40 × √(0.65 / 1) = 40 × 0.806 = 32.2

Same 40 GPM, four times less ΔP, needs about twice the Cv (because √4 = 2). That is the whole exam idea: do not invent a pump curve; do see that starving ΔP kills capacity unless the valve is larger.

Worked example — fixed valve, ΔP changes. Cv = 12, SG = 0.65.

• At 4 psi: GPM = 12 × √(4 / 0.65) = 12 × 2.48 = 29.8 GPM
• At 1 psi: GPM = 12 × √(1 / 0.65) = 12 × 1.24 = 14.9 GPM

ΔP fell by 4×; flow halved. Operators who say “we only lost a little head, flow should be almost the same” fail this item.

Elevation is part of available ΔP, which is why this is not a full hydraulic design. Liquid ammonia is lighter than water, so a given feet of lift costs fewer psi than the same feet of water (head in feet ≈ 2.31 × psi / SG). You are not asked to run a pipe-loop calculation. You are asked not to ignore that a drum in the basement pushing liquid to a penthouse spent some of its ΔP on lift before the feed valve even sees it.

Exam traps

• Mixing BPR (minimum evaporator P) with CPR (maximum crankcase P). Pull-down motor protection is CPR.
• Thinking a CPR raises suction at the evaporator. It limits suction at the machine; the room can still be hot.
• Calling a gas-pumped drum a mechanical pump because liquid still moves. The force is gas pressure on the liquid surface.
• Treating Cv = GPM or Cv = pipe size.
• Assuming flow is proportional to ΔP instead of √ΔP.
• Using SG = 1 for ammonia liquid without the stem saying water.