Hydraulics
Pumps, relief valves, contamination, and reading pressure like a symptom.
- Pressure is a result, not a cause
- Pump wear and internal leakage
- Reliefs set the ceiling
- Contamination and filtration
- Heat as information
What it is
Hydraulics does the jobs pneumatics cannot: very high forces, precise control of position under load, and the ability to hold a load still. It buys that with pressures an order of magnitude higher, a fluid that has to go somewhere, and a set of hazards that are genuinely different in kind.
Pressure is a result, not a cause
This is the single most important idea in the chapter and the one most often got backwards.
A pump makes flow. It does not make pressure. Pressure appears only where that flow meets resistance. Put a healthy pump on a circuit with nothing to push against and the gauge reads almost nothing — and that is a correct reading, not a fault.
So when you look at a gauge, you are not reading the pump's health. You are reading how hard the system is currently having to work, which is a fact about the load. From that one idea:
- No load, no pressure. Normal.
- Pressure at the relief setting means the flow has nowhere useful to go.
- The pressure a machine works at is set by what it is pushing, not by the relief valve.
How it works
The circuit
Force = pressure × area, exactly as in Chapter 10 — which is why hydraulic cylinders develop such large forces at modest sizes. Speed = flow ÷ area, so a slow actuator is a flow problem and a weak one is a pressure problem. The Chapter 10 distinction transfers unchanged.
What does not transfer is the return path. A pneumatic circuit exhausts to atmosphere and forgets about it. A hydraulic circuit has to bring every drop back to tank, through a filter, at a temperature the oil can tolerate — and most of the failure modes below live in that return journey.
Pumps and what wear does to them
Gear pumps are cheap and robust; vane pumps are quieter; piston pumps handle the highest pressures and can be variable-displacement, delivering only the flow the system is asking for.
The failure that matters diagnostically is internal leakage. A worn pump still moves oil, but an increasing fraction slips back through the clearances instead of going out of the port — and that fraction rises with pressure. So a worn pump behaves like a healthy one with no load and falls short exactly when the machine works hard.
That produces the signature worth memorizing: normal pressure, but everything gets slower under load.
Relief valves set the ceiling, not the working pressure
The relief valve exists to cap the system, and it does that by dumping flow back to tank once pressure reaches its setting. Two consequences:
- A relief stuck open, or set far too low, dumps flow continuously. The machine is slow or dead and the gauge never rises. It also heats the oil quickly, because all that energy has to go somewhere.
- A relief doing its job — the machine at full pressure and not moving — means the flow has nowhere else to go. That is a symptom of something downstream, not a fault in the relief.
Turning a relief valve up to make a machine work is the hydraulic equivalent of raising an overload setting: it removes a protection and leaves the cause.
Reading the gauge and the actuator together
Contamination is the dominant cause
Most hydraulic component failures trace back to dirt. Particles the size of the clearances inside a valve or a pump erode those clearances, and the wear produces more particles, which is a loop that only goes one way.
Filters sit in three places and do different jobs: a suction strainer protecting the pump, a pressure filter protecting downstream valves, and a return filter catching what the system generated. Most have a bypass so that a blocked filter does not starve the pump — which means a blocked filter is not obvious from the machine's behavior, and its indicator is the only thing that will tell you.
Water is the other contaminant. It degrades the oil, corrodes surfaces and destroys lubricity. Cloudy or milky oil is water, and it needs a cause found rather than a top-up.
Heat is information
Hydraulic oil temperature is one of the most under-read diagnostics available. Every psi of pressure drop across something that is not doing useful work becomes heat, so a system running hot is a system throttling flow somewhere — a relief lifting when it should not, a valve partly shifted, an undersized line, or a pump wearing.
Temperature also changes viscosity, which changes internal leakage, which is why some hydraulic faults appear only after an hour of running. If a fault has a warm-up time attached to it, oil temperature is the first thing to log.
What normally fails
- Symptom
- Everything on the machine has got slower, worse under heavy load, pressure looks normal
- Likely cause
- Pump wear — internal leakage rising with pressure
- How common
- Very common
The signature is that it looks healthy unloaded and falls short exactly when it works. Because pressure still reaches its setting, the gauge is reassuring, and the fault gets attributed to the machine for months. Comparing cycle times against the machine's own history is what catches it.
- Symptom
- Valves sticking, pumps failing early, repeated component failures across the machine
- Likely cause
- Contamination — particles at the size of the internal clearances
- How common
- Very common
The dominant cause of hydraulic failure, and it presents as many unrelated component faults rather than as one. Check the filter indicators and take an oil sample before replacing the third valve. Fitting new parts into dirty oil is buying the same failure again.
- Symptom
- Oil hot, machine slow, no obvious mechanical fault
- Likely cause
- Flow being throttled somewhere and turned into heat — often a relief lifting continuously
- How common
- Common
Heat is wasted energy and the energy came from somewhere. A relief passing when it should be closed, a directional valve only part-shifted, or a pump losing flow internally all convert pressure into temperature. Log the oil temperature rather than judging by hand.
- Symptom
- A cylinder drifts down under load, or will not hold position
- Likely cause
- Internal leakage past the piston seal, or a leaking holding valve
- How common
- Common
Distinct from an external leak and much harder to see, because nothing appears on the floor. Isolating the cylinder's ports and watching for drift separates the cylinder from the valve, and doing that before ordering either is worth half an hour.
- Symptom
- Pump noisy — whining, rattling, or a sound like gravel
- Likely cause
- Aeration or cavitation, usually a restriction or air leak on the suction side
- How common
- Common
The suction side is the part nobody looks at because it is not pressurized. A blocked strainer, a collapsed suction hose, a low oil level or an air leak at a fitting will all starve the pump and destroy it fairly quickly. The noise is the warning, and it is not subtle.
- Symptom
- A machine that was fine loses power gradually over weeks, oil looks milky
- Likely cause
- Water ingress — a cooler leak, washdown, or condensation in the reservoir
- How common
- Occasional
Water destroys lubricity and corrodes the surfaces that depend on it. Milky oil is diagnostic on sight. Find the route in — heat exchangers that leak the wrong way and reservoir breathers are the usual candidates — rather than just changing the charge.
- Symptom
- A machine loses its smooth stop, or an accumulator-assisted function becomes weak
- Likely cause
- Accumulator precharge lost through a failed bladder or a leaking gas valve
- How common
- Occasional
Accumulators lose nitrogen slowly and nothing announces it. The function they supported becomes gradually less effective, which reads as general deterioration rather than as a specific fault. Precharge should be checked as a routine task, not diagnosed after failure.
How to troubleshoot it
Isolate, discharge accumulators, and block raised loads
SafetyChapter 2, with the hydraulic specifics: the pump being off does not empty an accumulator, and bleeding pressure is what lets a raised platen fall. Confirm zero on the gauge and put something physical under anything held up.
Look at the oil before you look at the machine
Level, color, smell, and temperature. Milky means water, dark and burned means heat, low means a leak you have not found yet. Then read the filter indicators. Five minutes here regularly explains the whole callout.
Take the pressure reading and the movement observation together
The Fig 11.2 move. Gauge reading alone is ambiguous and so is watching the actuator; paired, they place the fault in one of four families. Take the pressure at the pump and, where you can, at the actuator too.
Listen to the pump
SafetyWhining or rattling is aeration or cavitation, and it points at the suction side — strainer, hose, oil level, or an air leak at a fitting. This is one of the few faults where the diagnosis arrives before any instrument does.
Establish whether the relief is passing
A relief dumping flow continuously heats the oil and starves the machine. An infrared reading on the relief's return line against the rest of the tank line is a quick indicator; a noticeably hot return with the machine idle says the relief is doing work it should not be.
Separate internal leakage from external
External leaks are visible and mostly obvious. Internal leakage — past a piston seal, across a valve spool — shows as lost speed and lost holding with nothing on the floor. Isolating a cylinder's ports and watching for drift settles which component it is before anything is ordered.
Compare against the machine's own history, not against a specification
Cycle times, oil temperature at a known point in the cycle, pressure at a known load. Hydraulic degradation is gradual and a single reading rarely proves anything; the same reading against last quarter's proves a great deal.
Find out where the dirt, water or heat came from
A cleaned valve in dirty oil fails again, and a new pump in hot oil wears quickly. The component you replaced is usually the consequence, and the filter indicator, the breather or the cooler is the cause.
Common technician mistakes
Turning the relief valve up
WhyThe machine is not doing what it should, the adjustment is right there, and more pressure sounds like more capability. It is Chapter 5's overload setting in a different fluid: the relief is a protection, raising it removes headroom that somebody calculated, and it does nothing at all about a flow problem — which is what most "weak machine" complaints turn out to be.
Searching for a leak by feel
WhyA pinhole jet is nearly invisible and running a hand along a hose is the fast way to find where the wet patch starts. It is also how injection injuries happen, and they are catastrophic out of all proportion to how they feel at the time. Cardboard, always, and stand out of line.
Topping the oil up without asking where it went
WhyThe level is low, oil is in the store, and the machine runs properly again — a complete, satisfying fix that takes four minutes. But the oil went somewhere, and the two candidates are a leak you have not found and a cooler leaking into water you are discharging. Both matter more than the level.
Reading low pressure as a fault
WhyA gauge near zero looks wrong on a machine that is not working. But an unloaded circuit should read near zero, because pressure is a response to resistance — so the reading is only meaningful taken against what the actuator is doing at that moment. Half of hydraulic misdiagnosis starts here.
Ignoring the filter bypass indicator
WhyMost filters bypass when blocked so the pump is not starved, which is sensible engineering and it means a fully blocked filter produces no symptom at all. The indicator is the only warning, it is easy to miss, and by the time behavior changes the unfiltered oil has been circulating for months.
Replacing the pump for a relief valve fault
WhyThe pump is the obvious suspect when a machine loses power, it is the expensive part, and replacing it feels decisive. But a relief passing continuously produces exactly the same symptoms, costs a fraction to check, and will do the same thing to the new pump's oil temperature.
Hands-on challenge
Scenario
Press P-4 — 'it's lost power, we need a new pump'
Press P-4 has lost roughly a third of its cycle speed over two months. Production have asked for a new pump.
What you find:
- System pressure reaches its relief setting of 2,600 psi normally and holds there while the press is under load.
- Cycle time is fine from cold and noticeably slower after about an hour of running.
- Oil temperature at the reservoir reads 155 °F; the machine's normal operating figure is recorded as 115 °F.
- The relief valve's return line to tank is hot to the touch even while the press is idle between cycles.
- The return filter indicator is in the red, and nobody can say when it was last changed.
- Oil is dark but not milky, and the level is correct.
Write down: which of the four pressure-and-movement families this is, what the worse-when-hot behavior tells you, what the hot relief return line means, why a new pump would probably not fix it, and what you would do in what order.
Show how to approach it
Take the pressure reading and the movement observation together, then work out which of the four families this falls into.
- Full pressure with slow motion is the bottom-right cell. Pressure reaching its setting means the pump can still build against resistance; losing speed means flow is going somewhere other than the cylinder.
- Worse when hot is the decisive detail. Internal leakage rises as oil thins, so a fault that appears after an hour and clears overnight is leakage, not a mechanical restriction.
- The hot return line points at the relief. A relief passing continuously converts pressure into heat and dumps flow to tank — which is both the lost speed and the temperature rise, from one cause.
- The pump is not yet exonerated, but it is not first. A worn pump gives the same slow-under-load signature; the difference is that it would not make the relief's own return line hot with the machine idle.
- The blocked return filter is the reason, not the fault. It has been bypassing for an unknown time, so the oil has been running unfiltered, which is what wears reliefs and pumps in the first place.
- Order the work: prove the relief, sample the oil, change the filter, then re-measure cycle time hot before deciding anything about the pump.
Knowledge check
Five questions. Each is answerable by reasoning about where the flow is going, rather than by recalling a component's function.
Question 1 of 5