Two payloads that look identical in a render can differ several-fold in price. Six components explain almost all of it: thermal core resolution, radiometric calibration, zoom optics, stabilisation class, laser ranging and environmental sealing.
- 6 driversExplain almost all variance
- 256→640The largest single step
- 30xWhere optics can outcost the sensor
- 0Ways to add radiometry later
Key takeaways
- The thermal core dominates. Stepping from 256×192 to 640×512 multiplies core cost, and pixel pitch and NETD grade multiply it again.
- Radiometry is factory calibration and per-unit validation — real cost and real value for inspection, wasted budget for pure detection.
- Optics scale non-linearly: a 30x optical zoom assembly can outcost the sensor behind it.
- Do not economise on stabilisation or integration support. A gimbal that cannot hold its pointing spec wastes every dollar spent above it.
On this page
The Thermal Core Dominates
Stepping from 256×192 to 640×512 multiplies core cost — that is roughly six and a half times the pixel count, on a detector whose yield falls as array size grows. Pixel pitch and NETD grade multiply it again: a 12 µm array is a harder manufacturing problem than 17 µm, and cores are graded and priced on the sensitivity they actually achieve.
This is the first line item to interrogate on any quote, because it is the largest and the most often obscured. Two quotes both saying “640 thermal” can describe materially different cores. Ask for pixel pitch, published NETD with its measurement conditions, and spectral band.
Whether the step is worth paying for depends on altitude and target size rather than on general goodness. Resolution buys pixels on target, which buys working altitude — 256×192 vs 640×512 works through the arithmetic, and thermal sensitivity explained covers when NETD rather than resolution is your limit.
The honest guidance: large-signal work such as finding a hot bearing or wet insulation does not need the most sensitive core available. Small-signal work such as detecting a person in cool grass at distance does.
Radiometry Is Calibration, Not Software
Thermometry versions carry factory calibration and per-unit validation. Somebody characterised that specific detector against reference sources across its temperature range and stored the result on the unit. That is real labour on real equipment, per unit, and it is why the thermometry configuration costs more.
It is also why it cannot be added later. There is no firmware key withholding a capability — there is no calibration data to unlock. Radiometric vs non-radiometric covers the decision, and the decision is genuinely one-way.
Real value for inspection programmes; wasted budget for pure detection missions. A SAR or patrol programme that will never issue a temperature is paying for a capability it cannot use. An inspection programme without it will buy the payload twice.
Optics Scale Non-Linearly
Germanium thermal lenses and long optical zoom trains grow expensive fast with focal length. Germanium is a costly material that must be precisely figured and anti-reflection coated, and a longer focal length means more of it — germanium lenses, f/1.0 apertures and athermalisation covers why thermal optics are the way they are.
On the visible side, a 30x optical zoom assembly can outcost the sensor behind it. A continuous zoom block is a mechanism: multiple moving groups, precision actuators, and a design that must hold focus and alignment through the whole travel while being shaken by an aircraft. What 10x to 40x costs in size, weight and light covers the engineering.
This is where hybrid zoom figures do the most damage to procurement. Digital multiplication is nearly free, optical magnification is not, and a specification written around a hybrid number invites bids that economise on precisely the expensive part — see hybrid zoom explained.
| Cost driver | What increases it | Worth paying for when |
|---|---|---|
| Thermal core | Resolution, finer pixel pitch, better NETD grade | Small targets, long standoff, low thermal contrast |
| Radiometric calibration | Per-unit factory characterisation | Reports quote temperatures — cannot be added later |
| Thermal optics | Focal length, aperture speed, germanium volume | Detection range matters more than field of view |
| Zoom optics | Optical factor, mechanism precision | Identification at standoff is the mission |
| Stabilisation class | Tighter pointing accuracy, more axes | Always, above modest zoom |
| Laser ranging | Emitter, receiver, calibration | Coordinates must be handed to someone |
| Environmental sealing | Ingress protection, corrosion resistance | Marine, dusty or all-weather operations |
Where Not to Economise
Stabilisation and integration support. These are the two line items where saving money reliably destroys the value of everything else on the quote.
A cheaper gimbal that cannot hold ±0.01° class pointing wastes every dollar spent on resolution and zoom above it. The logic is unforgiving: at 1,000 m, 0.01° of error moves the aim point 17 cm while 0.1° moves it 1.7 m, so a payload with excellent optics on a mediocre mount produces unusable imagery at exactly the range you bought it for. What ±0.01° means at 1,000 metres has the arithmetic.
Integration support is the less visible one. A payload with incomplete documentation and no responsive engineering contact costs weeks of your own staff time, and those weeks do not appear on any quote. An integrator’s checklist for evaluating suppliers covers how to test for it before ordering — the single best predictor is how fast a complete integration manual arrives when you ask.
The cheapest quote is often the one missing a line item. Maintaining link-specific variants, factory thermometry calibration, complete documentation and stocked spares all cost money that shows up in the price. A materially cheaper bid usually means one of them is absent.
The Purchase Price Is the Smaller Half
Six drivers explain the quote. They do not explain what the payload costs you over its life, and for most programmes the operating side is larger.
Spares and repair turnaround, firmware support duration, whether the variant still exists in three years, training time per operator, and the downtime when a unit goes back to the manufacturer — these accumulate quietly. The five-year cost of a thermal drone payload puts numbers to it, and warranty, spares and lifecycle covers what belongs in the contract.
There is also a fleet-level cost that per-unit comparison hides. Standardising on one payload spec means one set of spares and one training syllabus; buying three specialised units for three missions triples both. That trade often dominates the unit-price difference entirely — see one multi-sensor payload or several cameras.
How to Read a Quote
Request full configuration on every quote, and compare configurations rather than headline resolutions.
- Thermal core: resolution, pixel pitch, published NETD with measurement conditions, spectral band.
- Radiometric: thermometry version or not, and the published measurement accuracy figure.
- Optics: optical zoom factor stated separately from any hybrid number, and thermal lens focal length.
- Stabilisation: pointing accuracy, axis count, and the controllable range that decides where it can point.
- Ranging and sealing: LRF range and accuracy if fitted, ingress protection rating, laser class.
- Support: integration manual availability, spares stocking, repair turnaround, firmware support period.
Two quotes filled out against that list are comparable. Two quotes saying “640 thermal, 30x zoom” are not, and the gap between them is exactly where the price difference lives. How to read a thermal payload spec sheet covers where padding hides, and choosing a payload without spec noise covers keeping the decision anchored to the mission.
Related reading
- Thermal Infrared Optics: Germanium Lenses and f/1.0 Apertures
- Continuous Zoom Optical Blocks
- Radiometric vs Non-Radiometric Thermal
- 256×192 vs 640×512: Choosing Drone Thermal Camera Resolution
- Hybrid Zoom on UAV Cameras
- What ±0.01° Gimbal Accuracy Means at 1,000 Metres
- An Integrator’s Checklist for Evaluating UAV Payload Suppliers
- The Five-Year Cost of a Thermal Drone Payload
- MV-2P — 640×512 thermal at 130 g
- OP-90A — zoom, thermal and ranging
- OP-125A — 30x optical multi-sensor pod
- Solar Farm Thermal Inspection with UAV Payloads
- Zenmuse H30T alternatives: DJI airframe or independent payload
- Multi-sensor payloads: do you use every module?
FAQ
Why do quotes vary so much for “640 thermal”?
Core grade, radiometry, lens choice and sealing are all invisible in the headline resolution. Two 640×512 cores can differ in pixel pitch, NETD grade and spectral band; one may be a thermometry configuration and the other not; the lenses may differ by focal length and aperture speed. Request the full configuration on every quote — the price gap between two apparently identical payloads is almost always sitting in the lines nobody printed.
Is a multi-sensor payload cheaper than two payloads?
Usually yes in total cost of operation: one integration, one gimbal, one aircraft slot, and no swap downtime between missions. The unit price may be higher than a single-sensor pod, but the comparison that matters includes the second integration you did not have to do and the mission you did not have to abort to change payloads. For fleets it also halves the spares and training burden.
Where is the most common expensive mistake?
Economising on stabilisation. A gimbal that cannot hold its pointing specification wastes every dollar spent on the resolution and zoom above it, because at magnification the mount rather than the optics becomes the limit. The second most common is buying non-radiometric for a programme that later needs measurements — that one costs a whole payload, because calibration cannot be added after manufacture.
Questions about the technology? Talk to our engineers — we reply within 2 business days.