What a Pressure Compensated Pump Changes About the Circuit

Why this matters

Swapping a fixed-displacement pump for a pressure-compensated one is usually sold as an energy fix, and it is one. What matters more in the field is that it moves the working-pressure control off the relief valve and onto the pump, which changes what the relief valve is for, changes what a hot relief line means, and adds one setting nobody looks at until the machine misbehaves. A tech who reads a compensated circuit as if it were a fixed-pump circuit will adjust the wrong device, and the machine will keep working while running at its worst possible operating point.

It is also not automatically the lowest-heat answer. On a single-function machine it can be beaten by a spool choice that costs a small fraction of a pump, and the worked ledger below shows by how much.

What the compensator actually does

A variable-displacement pump can change how much oil it sweeps per revolution, usually by tilting a swashplate that sets piston stroke. A pressure compensator is a small spool valve that senses discharge pressure and, above its setting, uses system pressure to push the swashplate toward zero displacement.

So the pump delivers full flow at any pressure below the setting, and at the setting it delivers only as much flow as the circuit will accept at that pressure. Dead-head it and displacement falls to nearly zero: the pump keeps turning, holds the setting, and delivers only its own internal leakage plus the small control flow the compensator itself consumes.

That is the whole mechanism, and every consequence below follows from it. The compensator does nothing at all below its setting, which is the part people forget.

What it changes about the relief valve

On a fixed-pump circuit, the relief valve is the working pressure limit, and on any machine that dead-heads it passes the entire pump output every time. On a compensated circuit the compensator is the working limit, and the relief becomes a genuine backstop: it should never pass flow in normal operation.

Two field consequences follow immediately:

• A warm relief return line on a compensated circuit is a fault, not a condition. On a fixed-pump machine it may be normal. Here it means either the compensator has failed to destroke or the relief is set below the compensator, and the machine is running on the relief with full flow going to heat.
• The adjustment you want is usually the compensator, not the relief. Turning the relief up on a compensated circuit changes nothing about working pressure, because the pump already stopped delivering flow at the compensator setting.

The margin between the two settings

The compensator setting must sit below the relief setting, and the gap between them is a real design number rather than an afterthought.

Too small a margin and normal transients push flow over the relief. A compensator takes a finite time to destroke, tens of milliseconds on a typical unit, so a sudden dead-head produces a pressure overshoot above the setting before the swashplate gets there. If the relief sits inside that overshoot it opens on every cycle.

Too large a margin and the relief no longer protects anything useful, because the components see the higher number on any fault that holds the pump at full stroke.

A common shop default is to set the relief a few hundred psi above the compensator, on the order of 10 to 15 percent of the compensator setting, and then tune to whatever the machine builder's own documentation specifies, because the builder owns the overshoot behaviour of that specific pump. Set both, record both, and label them at the machine, because the two adjusters look identical and are frequently mixed up.

What it does not change

• It does not make the machine faster. Below the compensator setting the pump delivers exactly the same flow a fixed pump of the same displacement would, so stroke times are unchanged. Speed is a flow question and it is covered in the card on what pressure and flow each decide.
• It does not hold a load. Holding is a valve function and the card on directional valve center conditions covers why no spool does it.
• It does not reduce contamination sensitivity, it raises it. The compensator spool and the swashplate control piston run some of the tightest clearances in the machine, and they are the parts that stick. A compensated pump that will not destroke, or that hunts, is a contamination symptom more often than a spring or a setting problem, and the card on contamination as the dominant failure cause carries the mechanism.
• It does not remove the need for a relief. The overshoot above, plus any failure that leaves the swashplate at full stroke, are exactly the events the backstop exists for.

Worked ledger: the same machine, two pumps

One clamp machine, taken from the same family used throughout these cards. Fixed pump rated 8 gpm at rated shaft speed. A 3 inch bore cylinder, 7.07 square inches of piston area. Cycle: 5.5 seconds extend at 900 psi, 40 seconds of clamp hold with the directional valve on a closed center, 4.1 seconds retract. Call it 50 seconds, so the hold is 80 percent of running time, at 72 cycles per hour. A 20 gallon reservoir with about 22 square feet of surface, which at the midpoint of the bare-steel-in-still-air rule of thumb sheds roughly 27.5 BTU per hour for each degree F the oil sits above the room.

Hydraulic power is flow times pressure divided by 1,714, and one horsepower is 2,545 BTU per hour.

Option A, fixed pump, relief set at 2,000 psi. During the hold the pump has nowhere to go, so all 8 gpm crosses the relief. 8 times 2,000 divided by 1,714 is 9.33 hp, every bit of it heat because nothing moves.

There is a correction to apply on this side and only this side. The relief's full-flow pressure sits above its cracking setting by the valve's override, and for the direct-acting valve on this machine that is 2,600 psi at 8 gpm, as worked in the card on relief valve override. So the honest figure is 8 times 2,600 divided by 1,714, or 12.14 hp. That correction does not apply to Option B, because on the compensated side no flow crosses the relief at all, so the relief's curve never enters the arithmetic. This is not a comparator being left uncorrected; it is a term that genuinely exists on one side only.

Option B, pressure-compensated pump, compensator at 2,000 psi, relief backstop at 2,300 psi. During the hold the pump destrokes and delivers only internal leakage and control flow. Say the pump datasheet gives 0.3 gpm at 2,000 psi for this unit, which is illustrative here and comes off that pump's own leakage curve in practice. 0.3 times 2,000 divided by 1,714 is 0.35 hp.

The comparison. 12.14 hp against 0.35 hp is a factor of 35 on standby loss. Against the uncorrected 9.33 hp figure it is a factor of 27. Either way the reduction is more than an order of magnitude, and it comes from the pump refusing to make flow it cannot use.

What that does to oil temperature. Extend and retract contribute the same small heat in both options, about 132 BTU per hour from line and valve losses. Adding the hold:

• Option A: 12.14 hp for 80 percent of the time is 9.71 hp average, which is 24,700 BTU per hour, plus 132, call it 24,800. To shed that at 27.5 BTU per hour per degree F, the oil would have to sit about 900 F above the room. It cannot, so the machine climbs until it trips or fails, which is why this machine already carries a cooler.
• Option B: 0.35 hp for 80 percent of the time is 0.28 hp, which is 713 BTU per hour, plus 132, call it 845. Divided by 27.5, the oil settles about 31 F above the room. No cooler needed.

The finding that stops this being a sales pitch. The card on what a hydraulic system is trading away runs the same machine with a fixed pump and a tandem-center directional valve, which routes pump flow to tank at low pressure during the hold, and lands at about 4.8 F above the room. On this machine, a spool choice beats a pump change on heat, at a small fraction of the cost and with one fewer adjustment to get wrong.

So the compensated pump is not the answer to "my single-function machine runs hot." It is the answer to "several functions share one supply and pressure has to be available at all times," which is precisely the case where a tandem center cannot be used, because the first valve in the chain would steal the supply from the rest.

The failure mode. The one that shows up most is a relief set below the compensator, usually because a tech chased a pressure complaint by adjusting whichever valve was easier to reach. The machine still works. Flow crosses the relief continuously at full pump output, the standby saving disappears entirely, and the only external symptom is heat and a warm relief return line. Every number in Option A comes back on a machine that was bought as Option B. Check the two settings against each other before you conclude a compensated pump is not saving anything.

How to verify a compensator is doing its job

• Dead-head the circuit deliberately, where the machine's own documentation permits it, and read the gauge. It should hold steady at the compensator setting. Steady above the setting means the pump is not destroking. Climbing to the relief means it is not destroking at all.
• Listen and feel at the pump during that hold. A destroked pump gets noticeably quieter and its case drain runs cooler than a pump at full stroke against a relief. A pump that stays loud during a hold is still making flow.
• Touch-check the relief return line, briefly and with the back of a hand, only where it is not hot enough to burn. On a healthy compensated circuit it should be at tank temperature.
• Read the two settings against each other, not individually. Compensator first, at flow; then the relief, with the compensator temporarily backed off so the relief is genuinely the limiting device, then restore both and record them. Back the locknut off rather than removing any cap or spring cover from a pressurized valve, keep out of line of the adjuster, and wear eye protection.

References

• Pump manufacturer documentation for compensator adjustment range, destroke response time, case drain and internal leakage at pressure, which own every pump-specific number this card only shows how to use
• Machine builder documentation for the specified compensator and relief settings and whether a deliberate dead-head is permitted
• See related: What a Hydraulic System Is Actually Trading Away; What a Relief Valve Does in a Hydraulic Circuit and What It Costs; What a Directional Control Valve Is Doing in Each Position; Why Contamination Is the Dominant Failure Cause in Hydraulics