Why Heat in a Hydraulic System Is a Symptom Rather Than a Fault

Why this matters

"It runs hot" gets treated as the fault, and the response is usually a bigger cooler. Sometimes that is right and often it is expensive theatre, because oil temperature is not a quantity a machine has, it is the balance point between the heat the circuit makes and the heat it manages to get rid of. A reading of 150 degrees F can mean the circuit started wasting nearly twice as much power as it used to, or it can mean the cooler stopped working, or it can mean the shop got hot. Those three have three different fixes and one identical number on the gauge. This card is about which side of the balance moved, and it hands the question of which component is making the heat to the sibling HowTo on finding it.

Temperature is a balance point, not a quantity

Every watt of hydraulic power that does not leave the machine as mechanical work at an actuator leaves it as heat in the oil. The sibling card on what a hydraulic system is actually trading away owns that accounting. What matters here is what happens next.

Heat leaves through the reservoir walls, the lines, the components and a cooler if one is fitted, and every one of those paths rejects heat in proportion to how much hotter the oil is than the air around it. So as the oil warms, rejection rises, until rejection equals generation and the temperature stops climbing. That equilibrium is the number on your gauge.

Two consequences fall straight out of that and they govern everything below:

• Work in rise above ambient, never in absolute temperature. The balance is driven by the difference between oil and air. An absolute reading compared across two seasons, or across two bays, is comparing two different problems.
• The rise is roughly proportional to the generation. Double the wasted power, and the equilibrium rise above ambient roughly doubles. It does not double the temperature.

What sets the generation side

Generation is wasted hydraulic power, and it has a short list of homes: flow going over a relief valve at pressure, a pressure-compensated pump idling at full pressure on standby, flow forced through a throttling control, internal leakage across worn clearances, and pressure drop across undersized lines and loaded elements. Every one of those is a place where pressure falls without an actuator moving.

There is a feedback here worth naming, because it makes a mild problem worse than it looks. Hotter oil is thinner, thinner oil leaks more across the same clearance, and more internal leakage is more generation. The loop is bounded, because rejection rises with temperature too, but it means the machine finds its new equilibrium higher than the initial fault alone would suggest.

What sets the rejection side

Rejection is surface area, air movement and the temperature difference. On a machine with a cooler, most of the rejection is the cooler, which means most of the rejection is a component that can fail: a fouled core on the air side, a dead or reversed fan, a thermostatic bypass stuck open, a water-side circuit that has scaled, or a cooler that has simply been boxed into a corner with no air path.

The cue that separates a working cooler from a failed one is the temperature difference between its inlet and its outlet. A cooler rejecting more heat shows a larger drop across it; a cooler that has failed shows a smaller one. That direction catches people out, because a machine whose generation has gone up will show its cooler working harder, not less, right up until the cooler runs out of capacity.

Why the same reading means opposite things

Put the two sides together and the gauge is ambiguous by construction. A high rise above ambient says generation over rejection went up. It does not say which term did it, and the two fixes point in opposite directions: one is a repair inside the circuit, the other is a repair to the cooling path, and doing the wrong one leaves the machine exactly as hot.

That is the whole reason a bigger cooler so often disappoints. Adding rejection to a machine whose generation doubled brings the temperature back down while the circuit keeps burning the same wasted power, which shows up later as component life rather than as a gauge reading.

One measurement band, stated before it is used: treat a rise-above-ambient ratio within about 1.2 of the machine's commissioning baseline as inside the ordinary spread of ambient measurement and duty-cycle variation. Beyond that, something moved.

The error on the numbers, before they enter any arithmetic

Two different kinds of number get used below and they behave differently.

The rise above ambient is a difference between two readings usually taken with two different instruments in two different places, so their offsets do not cancel. Treat it as carrying both instruments' errors.

The drop across a cooler is a difference between two readings from the same instrument on the same kind of surface, minutes apart. A fixed systematic offset in that instrument is common to both readings and cancels in the difference, which is exactly why this comparison is worth more than either absolute reading. What survives is the instrument's random spread on each reading, and independent spreads combine in quadrature, so two equal terms multiply by the square root of two. On a small drop that residual matters; on a large one it does not.

One gate, two machines, the same gauge reading

The gate: correct to rise above ambient, compare the ratio against the commissioning baseline, then read the drop across the cooler.

Both machines below are the same model on the same duty, both reading a reservoir temperature of 150 degrees F, both in a 70 degree F shop, and both with a commissioning record showing a 45 degree F rise above ambient at that duty. Both were read at operating temperature with the machine cycling under its own controls, guards in place, nobody in the motion path, and the surface readings taken from arm's length rather than by hand, because line and manifold surfaces at these temperatures burn skin on contact.

The ambient correction, printed for both. Raw reading 150 degrees F. Ambient 70 degrees F. Rise = 150 - 70 = 80 degrees F. Baseline rise 45 degrees F. Ratio = 80 / 45 = 1.78. That is above the 1.2 band from the section above, so both machines have a real change. Two machines, two identical corrected figures, and so far no information about which side moved.

Machine A, the cooler check. Cooler inlet 152 degrees F, cooler outlet 138 degrees F, drop = 14 degrees F. The commissioning record for this cooler shows a 12 degree F drop. The drop grew. Both readings came from one instrument on the same steel surface finish, so its fixed offset cancels and what is left is the residual spread on a 14 degree difference, which is small against a 2 degree change but not nothing, so this was read three times and the drop held.

Rejection did not fail on Machine A, it increased, which is what the section above says a working cooler does when hotter oil arrives at it. Generation rose. The circuit walk found a cylinder that holds a load against a closed centre and a piston seal passing enough flow that the pump was sitting on its relief for most of the hold. The fix is the seal, and the temperature follows.

Machine B, the cooler check. Cooler inlet 151 degrees F, cooler outlet 148 degrees F, drop = 3 degrees F, against the same 12 degree F baseline. The drop collapsed to a quarter of baseline, and on a 3 degree difference the residual spread from the section above is a real fraction of the number, so this one was read at three points along the core rather than once. It held.

Rejection failed on Machine B. The core was packed on the air side, in a bay that had been fitted with a new dust collector six months earlier. The fix is cleaning the core and moving the air intake, and the circuit is untouched.

What each machine would have got from the other's fix. Machine A given a bigger cooler comes back to a normal gauge reading with a leaking piston seal still dumping pump flow over the relief, so the cylinder still drifts under load, the pump still runs at pressure it does not need, and the fault surfaces later as a pump. Machine B given a seal job comes back at 150 degrees F, because nothing in that circuit was ever generating more than it was designed to.

The failure mode in one line. Both jobs would have been closed as "runs hot, cooled it" by a shop that read only the reservoir gauge, and only one of them would have been fixed.

What the temperature is actually costing while you decide

Two costs, both real, and both routed rather than asserted.

Fluid life. Oxidation is a chemical rate, and a common rule of thumb for mineral oil is that the rate roughly doubles for every 10 degrees C (18 degrees F) of rise above about 60 degrees C. That is a chemistry approximation stated for bulk oil, not a specification for your charge, and the fluid manufacturer's own data governs the number for the product actually in the machine. The shape of it is the useful part: temperature costs fluid life geometrically, not proportionally.

Elastomer life. Hose covers, rod seals and O-rings carry a continuous service temperature rating from their manufacturer, and running near the top of it shortens everything at once. A machine that has climbed 35 degrees F above its commissioning rise has moved every elastomer in the circuit closer to its limit, which is why a hot machine tends to start leaking from several places in the same season rather than one.

How to verify which side moved

• Record the ambient in the same breath as the oil temperature, every time. Without it the reading is not comparable to anything, including itself last month.
• Compare rise to rise, and say when your baseline is not corrected. If the commissioning figure was taken at a different duty cycle or a different ambient and you did not adjust it, say so in the same sentence as the comparison rather than presenting the ratio as clean.
• Read the cooler's drop with one instrument, back to back, so the offset cancels, and read it more than once when the drop is small.
• Check the air or water path before condemning a core. A fan running backwards, a shroud missing, or a wall built two feet from the cooler produces the same collapsed drop as a fouled core and costs nothing to find.
• Then, and only then, find the component making the heat. The sibling HowTo on finding where a hydraulic system makes its heat is the next step once this gate has said generation.

References

• Fluid manufacturer documentation for the continuous and maximum bulk oil temperature of the specific charge, and for its oxidation behaviour with temperature
• Machine builder documentation for the commissioning oil temperature rise, the cooler's rated capacity, and the continuous temperature ratings of the hoses and seals fitted
• See related: What a Hydraulic System Is Actually Trading Away; How to Find Where a Hydraulic System Is Making Its Heat; What a Relief Valve Does in a Hydraulic Circuit and What It Costs