---
title: "Average Optical Power and OMA: Two dBm Values Can Answer Different Questions"
description: "Leviathan Systems — Technical content Published 2026-10-20; updated 2026-10-20."
canonical: "https://www.leviathansystems.co/articles/optical-average-power-versus-modulation-amplitude"
last-updated: "2026-10-20"
---

# Average Optical Power and OMA: Two dBm Values Can Answer Different Questions

Leviathan Systems — Technical content
Published 2026-10-20; updated 2026-10-20.

Average optical power describes the time average of optical power; optical modulation amplitude (OMA) describes a defined separation between optical levels. Both can be expressed in dBm because both can be referenced to 1 mW, but they retain different meanings. A satisfactory average-power entry cannot satisfy an absent OMA requirement. Compare the specified quantity and method for the identified optic before treating the evidence as complete.

## Read the waveform quantity before the unit

For a two-level nonreturn-to-zero (NRZ) optical signal, call the high optical power P1 and the low optical power P0. A binary zero need not mean zero emitted power. In the simplified model, OMA = P1 - P0, while the dimensionless linear extinction ratio is r = P1 / P0. These are a difference and a ratio, respectively.

For ideal fixed levels occupied for equal amounts of time, average power is Pavg = (P1 + P0) / 2. Maxim's HFAN-02.2.2 application note gives these relationships; the legacy MAX3910 data sheet's optical-power table explicitly retains the 50% mark-density assumption and its surrounding text specifies 50% duty cycle. Equal occupancy is a condition of this simplified mean.

Keysight's FlexDCA Average Power documentation describes the average optical component and permits watts or dBm as display units. Its instrument measurement is separate from determining a high-to-low amplitude. The field name therefore matters even when both report columns say dBm.

References: [A1](https://www.analog.com/media/en/technical-documentation/tech-articles/hfan0222-optical-modulation-amplitude-and-extinction-ratio.pdf), [A2](https://www.analog.com/media/en/technical-documentation/data-sheets/max3910.pdf), [K1](https://helpfiles.keysight.com/scopes/FlexDCA-UG/Content/Topics/Eye-Mask-Mode/NRZ-Measurements/nrz_average_power.htm)

## Two invented NRZ signals have the same average

Consider two ideal NRZ signals with fixed levels, equal time in each level and instantaneous transitions. Every value in this example is invented. It represents no module, measurement, signal-quality ranking, bit error rate or acceptance result.

Signal A has P1 = 0.8 mW and P0 = 0.2 mW. Its average is (0.8 + 0.2) / 2 = 0.5 mW. Its OMA is 0.8 - 0.2 = 0.6 mW, and its linear extinction ratio is 0.8 / 0.2 = 4.

Signal B has P1 = 0.6 mW and P0 = 0.4 mW. Its average is also (0.6 + 0.4) / 2 = 0.5 mW. Its OMA is 0.6 - 0.4 = 0.2 mW, and its linear extinction ratio is 0.6 / 0.4 = 1.5.

Both signals spend equal time above and below the same mean. Signal A's levels are each 0.3 mW from that mean; Signal B's levels are each 0.1 mW from it. The average discards that difference in separation. Reporting only 0.5 mW leaves the two signals indistinguishable on that one metric, even though A's OMA is three times B's. This is why an average-power value alone cannot determine OMA.

References: [A1](https://www.analog.com/media/en/technical-documentation/tech-articles/hfan0222-optical-modulation-amplitude-and-extinction-ratio.pdf), [A2](https://www.analog.com/media/en/technical-documentation/data-sheets/max3910.pdf)

## Apply dBm to the quantity you actually calculated

For average power, the conversion is 10 log10(Pavg / 1 mW). For OMA expressed in dBm, it is 10 log10(OMA / 1 mW), using the linear power difference for OMA. The reference is shared; the quantity inside the logarithm changes.

Both invented signals therefore have average power of approximately -3.01 dBm. Signal A's 0.6 mW OMA is approximately -2.22 dBm. Signal B's 0.2 mW OMA is approximately -6.99 dBm. Retain the labels average power and OMA beside those numbers. A's OMA exceeds its mean in linear power because separation between levels and time-averaged power are different quantities.

Do the level arithmetic in linear units first. For Signal A, the high and low levels convert to approximately -0.97 dBm and -6.99 dBm. Averaging those logarithmic values gives approximately -3.98 dBm, equivalent to 0.4 mW, rather than the stipulated 0.5 mW time average.

Subtracting the two level values in dBm instead gives approximately 6.02 dB, the logarithmic extinction ratio of 4. Signal B's ratio of 1.5 corresponds to approximately 1.76 dB. Neither ratio is OMA. An OMA result in dBm must also retain its identity: it is not fiber loss in dB. It contains no comparison across a fiber path. The related power-versus-loss guide develops that separate input/output question.

References: [A1](https://www.analog.com/media/en/technical-documentation/tech-articles/hfan0222-optical-modulation-amplitude-and-extinction-ratio.pdf), [A2](https://www.analog.com/media/en/technical-documentation/data-sheets/max3910.pdf), [F1](https://www.flukenetworks.com/knowledge-base/dsp-fta-series/db-vs-dbm), [K1](https://helpfiles.keysight.com/scopes/FlexDCA-UG/Content/Topics/Eye-Mask-Mode/NRZ-Measurements/nrz_average_power.htm), [K2](https://helpfiles.keysight.com/scopes/FlexDCA-UG/Content/Topics/Eye-Mask-Mode/PAM-Measurements/pam_outer_oma.htm)

## PAM4 outer OMA has its own measurement definition

PAM4 has four symbol levels, so the two-level example cannot simply be carried across. In Keysight's named FlexDCA Outer OMA measurement, PAM4 outer OMA uses levels 3 and 0. Their average levels are determined within defined windows in qualifying symbol runs, and the lower value is subtracted from the higher one. Optical results can be displayed in watts or dBm. The documentation imposes pattern conditions; this is more specific than taking the largest and smallest samples visible in a waveform.

A PAM4 time average depends on the whole waveform, including its middle levels and their occupancy. Its highest and lowest levels alone do not establish that average for an arbitrary signal. Do not transplant (P1 + P0) / 2 into an arbitrary PAM4 calculation, or turn a DOM (digital optical monitoring) entry into an inferred waveform result. A DOM field needs the identified equipment's own reporting definition and limitations.

The practical distinction is between a measurement name and a unit label. Confirm the exact OMA definition required by the optic's specification and the method represented by the returned result. The cited instrument documentation explains a named measurement; it supplies no universal requirement for every PAM4 optic.

References: [K1](https://helpfiles.keysight.com/scopes/FlexDCA-UG/Content/Topics/Eye-Mask-Mode/NRZ-Measurements/nrz_average_power.htm), [K2](https://helpfiles.keysight.com/scopes/FlexDCA-UG/Content/Topics/Eye-Mask-Mode/PAM-Measurements/pam_outer_oma.htm)

## A completed average-power entry leaves an absent OMA criterion open

Suppose a buyer receives a summary containing the invented -3.01 dBm average-power entry above, while the requested evidence separately calls for OMA. The supplier has supplied no OMA result. Copying -3.01 into the OMA column would erase the distinction demonstrated by Signals A and B. The proper conclusion is that the OMA question remains unassessed; the available average alone establishes neither satisfaction nor failure of it.

Ask for the exact optic identity, applicable specification revision, operating mode and optical lane or channel. Keep transmitter and receiver quantities distinct. Have the appointed equipment and test owners identify the required field, its conditions and measurement location, then establish whether the returned evidence uses that same definition and method. Numerical comparison comes after that match.

If average power satisfies its own applicable criterion, record that conclusion at its actual scope. Keep the absent OMA result and any other required waveform qualification visible as separate unresolved evidence. Even an available OMA result answers only its defined criterion; neither this arithmetic nor the shared dBm unit establishes complete optical suitability.

References: [A1](https://www.analog.com/media/en/technical-documentation/tech-articles/hfan0222-optical-modulation-amplitude-and-extinction-ratio.pdf), [K1](https://helpfiles.keysight.com/scopes/FlexDCA-UG/Content/Topics/Eye-Mask-Mode/NRZ-Measurements/nrz_average_power.htm), [K2](https://helpfiles.keysight.com/scopes/FlexDCA-UG/Content/Topics/Eye-Mask-Mode/PAM-Measurements/pam_outer_oma.htm)

## Bring a defined physical question to the verification inquiry

If the review also leaves an accessible physical question, Leviathan's physical-layer testing and installation-verification service is the relevant scope discussion. It can address agreed cable identification, labels, physical-connectivity checks, defect records and handoff documentation. Identify the links, permitted access and requested physical evidence before agreeing work.

Instrument measurements require confirmation of suitable equipment, qualified personnel, method, coverage, acceptance criteria and reporting limits for the engagement. Cable identification and connectivity records cannot supply missing OMA or waveform qualification. Use the physical-verification project inquiry with a nonconfidential outline of the accessible path and the physical question to resolve. Optical specifications, waveform evaluation and test decisions remain with the appointed equipment and test owners.

## Primary sources

- [A1 — Maxim HFAN-02.2.2: Optical Modulation Amplitude and Extinction Ratio, Rev 1 (April 2008)](https://www.analog.com/media/en/technical-documentation/tech-articles/hfan0222-optical-modulation-amplitude-and-extinction-ratio.pdf). Checked 2026-10-20. Page 2, Definitions and Relationships, supplies the bilevel power difference, level mean and extinction-ratio relationships. Useful for the bounded ideal NRZ model; its performance discussion and example specifications are not acceptance criteria here.
- [A2 — Maxim MAX3910 data sheet, Rev 1 (January 2003): Optical Power Relations](https://www.analog.com/media/en/technical-documentation/data-sheets/max3910.pdf). Checked 2026-10-20. Page 8, Table 1 and its surrounding text, corroborate the optical-power relationships with stated 50% mark density and duty cycle. Only this mathematical model is used; legacy component limits are not applied to modern optics.
- [K1 — Keysight FlexDCA User Guide: Average Power](https://helpfiles.keysight.com/scopes/FlexDCA-UG/Content/Topics/Eye-Mask-Mode/NRZ-Measurements/nrz_average_power.htm). Checked 2026-10-20. Defines the average optical component and watts/dBm display options for the named instrument measurement. Useful for distinguishing reporting quantity from display units; it does not define an arbitrary module's DOM implementation.
- [K2 — Keysight FlexDCA User Guide: Outer OMA](https://helpfiles.keysight.com/scopes/FlexDCA-UG/Content/Topics/Eye-Mask-Mode/PAM-Measurements/pam_outer_oma.htm). Checked 2026-10-20. Defines PAM4 outer levels 3 and 0, pattern-dependent measurement windows and optical watts/dBm reporting. Useful for preserving the named measurement's scope; its settings are not a universal test recipe or optic acceptance limit.
- [F1 — Fluke Networks: dB vs dBm (February 28, 2014)](https://www.flukenetworks.com/knowledge-base/dsp-fta-series/db-vs-dbm). Checked 2026-10-20. Defines the 1 mW reference and distinguishes a power level from a power comparison. Used only for units; typical source outputs on the page are not optic specifications or project limits.

## Related Leviathan resources

- [Physical-layer testing and installation verification](/services/network-testing)
- [Discuss an accessible physical-verification scope](/contact?service=network-testing&from=/articles/optical-average-power-versus-modulation-amplitude)
- [Optical power in dBm versus fiber loss in dB](/articles/optical-power-dbm-vs-fiber-loss-db)
- [Long-reach optics on short links: receiver power windows](/articles/long-reach-optics-short-link-receiver-power-window)
- [400G, 800G and 1.6T optics: read the complete interface](/articles/400g-800g-1600g-optics-selection)

## Define the physical evidence your optical review needs

Share a nonconfidential outline of the accessible fiber path, link identities and requested physical checks. Our team will follow up to discuss scope, requirements and availability.

[Discuss the scope](/contact?service=network-testing&from=/articles/optical-average-power-versus-modulation-amplitude)