Field Notes
Liquid-cooled load bank purge and coolant sample custody record
Before a commissioning ramp test, the record should show the liquid-cooled load bank, loop boundary, hose IDs, purge witness, leak status, coolant sample ID, chain of custody, baseline pressure, flow, temperature, alarms, exceptions, and ramp release decision.
Direct answer
Before a liquid-cooled load bank commissioning ramp test, the record should identify the load bank, CDU or loop served, rack row or test header, hose IDs, supply and return connections, caps removed, purge point, purge method witness, air or debris observed, leak-check status, coolant source, sample point, sample ID, container count, collection date and time, collector, custody transfers, lab or onsite test receipt, baseline pressure, flow, temperature, conductivity or chemistry where monitored, alarms, exceptions, and ramp-test release decision.
The record should prove that the loop was not ramped only because hoses were connected and the load bank powered up. It should show purge evidence, coolant identity, sample traceability, leak status, sensor baseline, and who accepted the ramp boundary.
Use this as documentation guidance only. The commissioning plan, CDU manual, load-bank manual, coolant supplier instructions, laboratory instructions, owner standard, electrical safety plan, liquid-cooling safety plan, and qualified team control actual purge, sampling, testing, valve operation, energization, and ramping.
Why purge and sample custody matter
Liquid-cooled load banks are used to prove heat rejection and hydronic readiness before real IT load is exposed. That makes the ramp test valuable, but it also raises the consequence of hidden air, debris, wrong coolant, undocumented water, or a leak that appears only after flow and temperature rise.
The weak record says load bank connected, purged, sample taken, ramp approved. The strong record shows hose IDs, connection photos, purge witness, sample ID, custody handoffs, baseline trend, leak check, and exact ramp limit.
Avtron, ASHRAE, Dell, Dober, Trane, Vertiv, Chilldyne, Lenovo, and TI sources all support the need to treat liquid cooling as a controlled fluid, pressure, flow, temperature, and monitoring system. EPA and lab sources support sample identity and custody documentation.
Start with loop and ramp boundary
The first page should name the commissioning step, load-bank tag, CDU tag, manifold, loop, test header, rack row, planned ramp size, current load, target load, coolant source, sample point, witness team, and stop authority.
Do not record only the load bank name. A liquid-cooled load bank may be connected through hoses, adapters, quick disconnects, a CDU, temporary skid, facility water loop, or technology cooling loop. The record should say exactly what fluid path is being tested.
If the ramp covers only one loop, one row, one CDU, or one load step, state that boundary so later teams do not treat the record as a full-system release.
Record purge evidence and air or debris status
Record who witnessed purge, what loop was purged, when purge started and ended, whether air, debris, discoloration, foam, odor, or pressure fluctuation was observed, and what condition was accepted before ramping.
Do not turn this into a field procedure. The purge method belongs to the commissioning plan and equipment manuals. The record only preserves what the qualified team did and what evidence supported the release.
If purge evidence is missing, the ramp record should be held even if the load bank can still pull load.
Tie coolant samples to source and loop
The sample record should identify the coolant source, loop, sample point, pre-flush or post-flush status, container count, sample ID, collection time, collector, requested analysis, field readings where taken, and whether the sample represents the exact loop being ramped.
Dell, Dober, ASHRAE, Lenovo, and Trane sources all show why water, glycol, conductivity, treatment, and fluid maintenance context matter in liquid cooling. The article does not set chemistry acceptance. It requires traceability before the ramp decision relies on sample evidence.
If the sample label says CDU sample but the ramp uses a temporary load-bank hose loop, fix the identity before ramping.
Preserve chain of custody
Chain of custody should show who collected the sample, who held it, who transferred it, when it was transferred, how many containers moved, what analysis was requested, and who received it at the lab or onsite test station.
EPA and lab chain-of-custody sources show common custody fields: sample ID, matrix, collection date and time, container count, preservation or handling notes, and relinquished or received signatures. Use the lab's required form when a lab is involved.
A sample with no custody trail may still be useful for informal troubleshooting, but it should not be treated as the controlled chemistry baseline for a commissioning ramp.
Trend pressure, flow, temperature, leaks, and alarms
Trend or record baseline supply and return temperature, pressure, differential pressure where available, flow, pump command, pump status, coolant level, leak sensors, conductivity or chemistry where monitored, and alarms before the ramp begins.
TI monitoring material, CDU documentation, and liquid-cooling deployment sources support monitoring around leak, flow, pressure, and temperature. The ramp release should preserve normal values before load changes make the event harder to interpret.
If baseline flow, pressure, temperature, or leak status is missing, hold the ramp or record the authority that accepted the missing evidence.
Separate sample hold from ramp hold
Not every pending lab result automatically stops a ramp, and not every clean quick test automatically releases it. The commissioning plan should define what field checks are required before the ramp and what lab results can follow.
The record should separate: sample collected, sample custody complete, field chemistry accepted, lab result pending, lab result required before ramp, purge accepted, purge held, ramp allowed, ramp allowed with monitoring, or ramp held.
That separation prevents a missing lab report from being hidden inside a general ramp approved note.
Record table
Use a compact table so commissioning, facilities, liquid-cooling vendors, load-bank technicians, and the owner review the same evidence.
| Record field | What to capture | Why it matters |
|---|---|---|
| Ramp boundary | Load bank, CDU, loop, manifold, row, ramp size, stop authority | Defines what is being released |
| Hose evidence | Hose IDs, supply/return, caps, QDs, drip control, restraints, purge point | Shows the physical path used for the test |
| Purge record | Witness, time, loop, air, debris, discoloration, accepted condition | Proves the line was prepared before ramp |
| Sample identity | Sample point, sample ID, coolant source, container count, collector | Ties chemistry evidence to the loop |
| Custody | Relinquished/received names, times, lab receipt, handling notes | Protects sample traceability |
| Baseline trend | Pressure, flow, temperature, pump status, level, leak sensors, alarms | Shows the loop condition before load |
| Exceptions | Missing sample, broken custody, air, leak, alarm, wrong hose, pending lab | Makes hold points visible |
| Ramp decision | Release, release with monitoring, hold, retest, lab result required | Defines the commissioning action |
Before-ramp checklist
Run this checklist before the commissioning ramp test.
- Load-bank tag, CDU or loop, manifold, and ramp boundary are identified.
- Supply and return hoses are tagged and photographed.
- Caps, quick disconnects, drip trays, and temporary adapters are documented.
- Purge point, purge witness, purge time, and accepted purge condition are recorded.
- Air, debris, discoloration, foam, leaks, and pressure fluctuations are exception-listed.
- Coolant sample point, sample ID, container count, and collection time are recorded.
- Chain of custody shows collector, transfers, and lab or onsite receipt where applicable.
- Baseline pressure, flow, temperature, pump status, level, leak sensors, and alarms are recorded.
- Pending lab results or chemistry holds are separated from ramp approval.
- Ramp release decision, stop authority, monitoring window, and rollback owner are written down.
Weak versus strong record
Weak record: Hoses purged, coolant sample taken, load bank ramp OK.
Strong record: Load bank LCLB-3 was connected to CDU-2 secondary loop at Test Header C using supply hose S-3 and return hose R-3. Photos showed clean caps, seated quick disconnects, drip trays, and the purge port. The commissioning witness recorded purge from 8:12 a.m. to 8:21 a.m. with air observed for the first two minutes and clear coolant before close. Sample CDU2-C-0610-01 was collected at the post-purge sample point at 8:26 a.m., labeled with two containers, relinquished to the lab courier at 9:04 a.m., and logged as received at 9:42 a.m. Baseline flow, pressure, temperature, leak sensors, level, and alarms were normal for 30 minutes. The ramp was released to 25 percent only, with lab results required before the next step.
The strong record ties hose path, purge evidence, sample custody, baseline trend, and ramp limit together.
Common mistakes
The most common mistake is recording that the loop was purged without showing which hose pair and which purge point were used. That makes the purge note hard to connect to the actual ramp path.
Another mistake is collecting a coolant sample but losing identity. A bottle with no loop, time, collector, container count, or custody trail cannot carry the same weight as a controlled sample.
Other mistakes include unlabeled hoses, supply and return confusion, no cap photos, no drip-tray status, no leak sensor baseline, no pending-result hold, no lab receipt, and no stop authority for the ramp.
When to hold the ramp test
Hold the ramp if hoses are unlabeled, supply and return cannot be verified, purge was not witnessed, air or debris remains unresolved, leak sensors are active, pressure or flow is unstable, coolant source is unknown, sample labels are wrong, custody is broken, or required field chemistry is missing.
Also hold if the CDU or load-bank manual requires support, if a fitting was disturbed without retest, if baseline trends are missing, or if the commissioning plan requires lab results before increasing load.
A hold should name the load bank, loop, sample ID, missing evidence, required correction, owner, retest evidence, and whether a lower ramp step can proceed.
Owner handoff
The owner handoff should include hose photos, purge record, sample labels, chain-of-custody form, field readings, lab receipt, baseline trend, leak-check notes, alarms, exceptions, ramp decision, monitoring window, and next-step requirements.
Keep the record with commissioning ramp files, coolant chemistry files, load-bank reports, CDU trend exports, and leak-response records.
Questions before ramp
Which loop is being ramped? Which hoses connect it? Was that exact hose path purged? What sample represents the loop? Who collected it, who handled it, and where was it received?
What baseline pressure, flow, temperature, level, leak status, and alarms were normal before the ramp? What lab or field chemistry evidence is required before this step? Who can stop the test?
Answer those questions before increasing load.
Compliance and safety limits
This article does not approve coolant chemistry, prescribe purge steps, set ramp rates, authorize valve operation, define sample preservation, or replace commissioning. It is a record structure for preserving hose purge, coolant sample custody, baseline trend, and ramp-release evidence.
The commissioning plan, CDU manual, load-bank manual, coolant supplier instructions, lab instructions, owner standard, electrical safety plan, liquid-cooling safety plan, and qualified team control the work. If those documents conflict with this checklist, use the controlling document and record the decision.
Do not connect hoses, purge lines, collect pressurized samples, handle coolant, energize load banks, or increase commissioning load outside the qualified team's authority.
Sources checked
- Avtron, Liquid-Cooled Load Banks ReadinessUsed for liquid-cooled load-bank commissioning and readiness context.
- ASHRAE, Emergence and Expansion of Liquid CoolingUsed for data center liquid-cooling, CDU, loop, and integration context.
- Dell Technologies, Liquid Coolants GuidanceUsed for water and water-glycol coolant, TCS/FWS, CDU, and maintenance context.
- Dober, Glycol for Direct-to-Chip CoolingUsed for data center glycol and coolant chemistry context.
- Trane Technologies, Coolant Distribution Unit IOMUsed for CDU safety, fluid, alarm, operation, and maintenance context.
- Vertiv, Deploying Liquid Cooling in the Data CenterUsed for liquid-cooling deployment and operational readiness context.
- Chilldyne, FAQ Guide to Data Center Liquid CoolingUsed for liquid-cooling loop and operational risk context.
- ASHRAE TC 9.9, Water-Cooled ServersUsed for water-cooled server loop and water-quality context.
- Lenovo, Water specifications for secondary cooling loopUsed for secondary-loop water specification context.
- Texas Instruments, Leak Detection, Flow, and Pressure MonitoringUsed for leak, flow, pressure, and sensor monitoring context.
- EPA, Sample Chain-of-Custody FormUsed for sample identity, container, handling, and custody fields.
- EPA, Sample Handling and CustodyUsed for physical sample handling and custody documentation context.
- Pace Analytical, Chain of Custody FormsUsed for laboratory chain-of-custody form and instruction context.