Instrument and path control
Record calibration status, wavelength setting, connector cleanliness and reference path loss.
Quality and evidence control brief
Optical Power & DOM Testing for Transceivers
Separate calibrated optical measurement from module-reported DOM, then record the margin, conditions and lane-level exceptions.
Optical power testing should preserve units, wavelength, path loss, warm-up state, instrument identity and calibration status. A number without these conditions cannot be compared reliably.
DOM is valuable operational telemetry, but its readings and thresholds should be compared with an external measurement and the applicable specification, especially on multi-lane modules.
MingApex coordinates China-based sourcing, OEM execution and independent quality-control evidence. It is not represented as the product manufacturer or an accredited laboratory; the commercial record names the supplier, method and applicable evidence boundary.
01 / Release record
Define the exact evidence record before inspection starts
A usable Optical power and DOM record identifies the subject, baseline, method, limit, sample plan, traceability chain and release owner.
| Required control | What must be recorded | Why it matters |
|---|---|---|
| Product and order identity | Serialized optic, lane/channel map, wavelength, reach and temperature class. | Evidence cannot be transferred safely when the SKU, revision, order line or lot is ambiguous. |
| Approved reference | Approved datasheet/order limits and, where available, approved sample baseline. | The comparison baseline must be frozen before inspection results are known. |
| Method and equipment | Calibrated optical power meter/test set, controlled patch path, DOM readback and stated temperature. | A result without method, equipment or conditions cannot be reproduced or audited. |
| Acceptance limits | Tx output, Rx input/margin, lane balance, DOM correlation, thresholds and alarm behaviour. | Pass/fail criteria must come from the approved requirement, not from the observed result. |
| Sampling plan | Serialized units across requested speed/reach variants and all optical lanes. | Sample quantity and selection determine what the decision can legitimately cover. |
| Traceability record | Instrument ID/calibration date, setup photo, raw readings, DOM capture, operator, date and unit serial. | Raw files, photos, readings and serials must resolve to the same tested units. |
| Decision and owner | Pass/hold by channel and unit, with margin and retest condition recorded. | Open deviations need an owner, closure rule and release status. |
02 / Control points
Checks required for an auditable decision
Each check below produces evidence or a named deviation; a tick-box without the underlying record is not sufficient.
Tx output power
Measure every lane/channel after stated warm-up and compare with approved limits.
Rx power and margin
Record applied/input power and margin to the project threshold; avoid unsupported sensitivity claims.
DOM correlation
Compare module-reported values with external measurement and document the tolerance used.
Thresholds and alarms
Check readable thresholds, lane mapping and alarm transitions without assuming factory defaults are correct.
03 / Product anchors
Live product examples for defining the inspection family
These live SKUs help define the physical product family and inspection variables. They do not inherit evidence from another sample or lot.
Live product page10G SFP+ 1550nm ER 40km LC TransceiverLong-reach power-budget and traceability referenceLive product page25G SFP28 1310nm LR 10km LC TransceiverSFP28 coding, FEC and DOM evidence referenceLive product page100G QSFP28 SR4 100m MPO TransceiverFour-lane optical and MPO inspection referenceLive product page100G QSFP28 LR4 10km LC TransceiverFour-wavelength DOM and production-lot referenceLive product page400G QSFP-DD FR4 2km LC TransceiverQSFP-DD firmware, DOM and change-control reference
Product pages define commercial and physical starting points. Inspection evidence applies only to the recorded sample, lot and order line.
04 / Release workflow
From specification lock to documented release
A defensible record starts with the instrument and optical path, not with a copied DOM screenshot.
- 1
Freeze the specification and evidence scope
Identify the exact product, order line, approved reference and buyer requirement for Optical power and DOM.
- 2
Define method and acceptance limits
Record the instrument or inspection method, sample plan, units, tolerances and pass/fail rules before results are known.
- 3
Capture traceable observations
Tie raw readings, photos, files and exceptions to serial numbers, lot identity, date and responsible operator.
- 4
Issue a controlled decision
Record pass, conditional pass, hold, fail or retest together with deviations, owner and closure condition.
- 5
Bridge the decision to shipment and repeat orders
Release only the recorded quantity and preserve approved identities, revisions and change-control triggers for the next order.
05 / Decision states
How pass, conditional release, hold and retest are handled
The same observation can lead to different decisions depending on the approved scope, open deviations and shipment risk.
Pre-order definition
Freeze wavelength, lane map, units, instrument, path loss, warm-up condition and numerical limits before measurement.
Conditional release
A small DOM correlation deviation may be conditionally accepted only with stated tolerance, owner and impact; an out-of-spec external reading is not closed by DOM.
Hold or retest
Expired calibration, dirty/changed path, wrong wavelength, unstable temperature, missing lanes or out-of-limit reading requires a controlled retest.
Repeat-order bridge
Repeat the agreed sample plan on the production lot and compare lane statistics and DOM correlation with the approved baseline.
06 / Buyer Evidence Center
Downloadable evidence references
Use these redacted sample records to specify required fields. They are process references, not proof for an unrelated product or order.
07 / Risk control
Failure modes that invalidate a release decision
DOM used as the only meter
Telemetry is not a substitute for controlled external measurement.
Unknown calibration status
Instrument drift can move a result across the acceptance limit.
Wrong wavelength or path
Measurement conditions may not represent the quoted optic.
Lane averages hide failures
One weak lane can disappear inside a module-level average.
08 / Buyer questions
Quality and evidence questions
What does this record prove?
The result applies to the serialized units, wavelength channels, test path, instrument status, temperature and acceptance limits recorded. DOM alone is not a calibrated power certificate.
What must the buyer provide?
Provide optical standard, operating wavelengths, acceptance ranges, lane requirements, temperature and evidence format.
Can an open deviation be accepted?
A small DOM correlation deviation may be conditionally accepted only with stated tolerance, owner and impact; an out-of-spec external reading is not closed by DOM.
Does the result automatically cover a repeat order?
Repeat the agreed sample plan on the production lot and compare lane statistics and DOM correlation with the approved baseline.
Controlled next step
Define the optical power and DOM evidence record
Send the optic family, lane map, numerical limits, temperature and required external/DOM comparison.