Coolant Pressure Test Carts for Data Center Liquid Cooling
Every direct-to-chip loop gets pressure tested before the rack it serves is energized. The reasoning is simple: an assembled loop contains dozens of mechanical joints — manifold ports, blind-mate QD faces, hose crimps, cold plate fittings — and the only practical way to verify all of them at once is to put the loop under pressure and watch what happens. A joint that weeps during commissioning costs a wrench turn. The same joint found after power-on can cost a rack, and possibly the racks below it.
The trouble with shopping for a pressure test cart is that most of what turns up is built for hydraulics or HVAC. A 10,000 psi test bench cannot resolve the half-psi drift that matters on a 75 psi hold, and a rig plumbed in carbon steel with nitrile seals will contaminate a loop that has to stay clean enough for coolant service. Data center loop testing is its own job: low pressure class, fine resolution, coolant-compatible wetted path.
We build three carts for that job in Van Alstyne, Texas, around 316L manifolds machined on our own machining and turning centers. All three fill, flush, and pressure test across the full 0-100 psi class, run PG25, PG30, or DI water, and ship in 10 to 15 business days.
PARTS ON THIS PAGE: RH-FC-150 · $18,700 RH-FC-300 · $26,600 RH-FC-600 · $43,700 RH-FC-HK-UQD · $2,600 RH-FC-ADP-U · $2,000 RH-FC-FE25 · $265
Why loops get hydro and decay tested before energization
Component-level proof testing should be finished long before hardware reaches your dock — our manifolds, for example, ship hydro-tested to 225 psi with a serialized test cert. What the factory cannot test is the field assembly: the UQDB blind-mates made as racks slid home, the hoses routed and torqued on site, the adapter that got swapped at the last minute. The field pressure test verifies the loop as built, not the parts as shipped.
Timing is the other half of it. On an ORV3 rack the blind-mate connections sit where you cannot see them, so a visual walk-down catches almost nothing. A hydrostatic hold followed by a timed decay test checks every joint simultaneously, hidden ones included, while the loop is still cold and the rack is still dark. Most commissioning specs also want the decay numbers in the turnover package, so the test doubles as documentation.
What the test rig actually needs
- A positive-displacement test circuit. A centrifugal pump is a flow machine — deadhead it against a closed loop and the pressure wanders with pump temperature, and it cannot hold setpoint at all once you shut it off. You want a gear pump feeding through a regulator: bring the loop to test pressure, isolate it, and read the gauge with the pump out of the circuit.
- A 0-100 psi class range, regulator-adjustable. Direct-to-chip loops are built from 150 psi class hardware, but commissioning tests run at or modestly above working pressure. The gauge should match: 0.1 psi resolution on a 100 psi full scale will show a decay that a 5,000 psi hydraulic gauge never registers.
- A decay readout, not just a gauge. Stopwatch-and-clipboard works, but a transducer that computes psi per minute removes the arithmetic and the arguments. At facility scale, logged data with pass/fail criteria closes out the paperwork the same day as the test.
- A coolant-clean wetted path. The test fluid usually becomes the first operating fill, so everything it touches — pump, manifold, hoses — should be 316L stainless with EPDM or PTFE seals, filtered through 25 micron absolute elements.
- Containment and mobility. Locking casters, a drip pan under the manifold, and a footprint that clears the freight elevator and a 30-inch door.
Three carts, sized by team
The RH-FC-150 runs on 120V wall power in a 26-inch-wide frame. Its fill pump is positive displacement, so one pump handles both fill and test duty — that is what keeps the price under nine thousand dollars. The RH-FC-300 is the cart most teams should look at first: separate flush and test circuits, so the centrifugal handles flow and the gear pump handles pressure, with a 55-gallon reservoir that gets through a full row without stopping. The RH-FC-600 adds a VFD, dual pumps, and a touchscreen that runs the decay routine for you and writes pressure, temperature, flow, and decay rate to CSV — worth the step up on jobs where every loop needs a documented pass/fail record.
For hookup, a stainless hose kit with UQD08-compatible ends (RH-FC-HK-UQD, $2,600) and a 14-piece UQD-to-NPT/JIC adapter set (RH-FC-ADP-U, $2,000) connect any of the three carts to whatever the loop presents.
| Model | Price | Flush flow | Test circuit | Decay readout | Sized for |
| RH-FC-150 | $18,700 | 15 GPM | 316L gear pump (fill and test) | Manual timed decay, 0.1 psi digital gauge | One tech, rack by rack |
| RH-FC-300 | $26,600 | 30 GPM | Dedicated 316L gear-pump circuit | Timed decay with psi/min transducer readout | Two to three techs, row scale |
| RH-FC-600 | $43,700 | 60 GPM | Dedicated 316L gear-pump circuit | Automated routines, pass/fail limits, CSV logging | Commissioning crew, facility scale |
RH-FC pressure test cart lineup
The test procedure, in outline
- Fill. Push coolant into the loop through the cart's 25-micron filtration at moderate flow. First fills of new piping shed debris, so plan on changing elements more often than you will in service (spares: RH-FC-FE25, $265 for three).
- Purge. Circulate until the return runs clear of bubbles, working high-point vents and cycling the QDs. Trapped air is the enemy of the decay test — it makes the loop spongy and masks small leaks.
- Pressurize. Switch to the gear-pump test circuit and bring the loop to the test pressure in your commissioning spec through the regulator, then isolate the loop from the cart.
- Stabilize. Give it 10 to 15 minutes before starting the clock. Seal seating, hose stretch, and temperature equalization all move the gauge early, and none of them are leaks.
- Hold and record. Run the timed hold your spec calls for and record the decay rate. Depressurize fully before breaking any connection.
Reading the decay numbers
- A drop in the first few minutes that flattens out is normal. That is hose expansion, seals seating, and fluid finding room temperature — the reason the stabilization step exists. Start the clock immediately and you will fail loops that are tight.
- A slow, linear decline that never flattens is a real leak. The usual suspects, roughly in order: QD faces that did not fully seat, adapter O-rings, drain or vent valves left a quarter turn open, hose crimps.
- A fast fall toward zero is an open path — a half-seated blind-mate or a cracked vent. Walk the loop before repressurizing.
- Decay that tracks temperature is physics, not a leak. A liquid-filled loop is hydraulically stiff, so a couple of degrees of fluid temperature change reads as several psi on the gauge. Flush hard and test immediately, and warm coolant cooling to room temperature will mimic a leak. Let it equalize — or log temperature alongside pressure on the RH-FC-600 and correct for it.
- On pass criteria: your site spec governs, and there is no universal number. A common form is a maximum allowable drop over a fixed hold after stabilization — 2 psi over 30 minutes, say. The trend matters more than the absolute figure; a decay that is still accelerating at the end of the hold is telling you something.
Common questions
Can I pressure test a loop with the fill pump?
What decay rate counts as a pass?
Should I test with water or with the operating coolant?
What is the lead time, and do you take purchase orders?
Put a test cart on the commissioning schedule
All three carts ship from Van Alstyne, Texas in 10 to 15 business days. POs accepted on net-30 terms. If your loop presents something unusual, send the details — custom manifolds are quoted in 24 to 48 hours.
See Cart Specs & Pricing