Agras T70P Keeps Video Feed Rock-Solid While Scanning Solar Panels in **10 m/s** Gusts—A Crop Duster’s Field Diary
Agras T70P Keeps Video Feed Rock-Solid While Scanning Solar Panels in 10 m/s Gusts—A Crop Duster’s Field Diary
TL;DR
- 10 m/s ridge-top wind, EMI from 1 500 V DC cables, yet the T70P never dropped an RTK Fix below 99.2 %.
- Active Phased Array Radar + dual-band antenna kept the downlink 100 % stable at 1.2 km—no frame drops, no RTH.
- A third-party 15 000 lm CREE spotlight clipped to the T70P’s散布播撒器 mount cut glass glare and let us spot cracked cells at 3 cm ground-sample distance.
The sun was already high when I reached the south-facing solar terrace above the almond blocks. Two million panels, 30 % grade, and a cold Pacific wind funneling through at a steady 10 m/s. Operations like these aren’t about pretty cinematic shots—they’re about centimeter-level precision, repeatable passes, and a video feed that refuses to stutter while you’re hugging glass at 3 m AGL.
I’d brought the Agras T70P. Not a survey bird out of the box, sure, but I’ve learned that anything able to hold 70 kg of slurry in rolling terrain can sure as heck carry a Phase-One survey pod and a spotlight. The mission: inspect 180 ha of modules for micro-cracks, hotspot delamination, and junction-box uplift before tomorrow’s insurance adjusters arrived.
Why Signal Stability Matters on a PV Farm
Solar farms are RF swamps. Inverters spew harmonics, DC strings act like long-wire antennas, and the steel piles love to bounce every frequency back at you. Drop a frame or lose RTK here and you’re flying blind above acres of fragile silicon. That’s spray-drift-level risk—only instead of herbicide you’re scattering bad data across a banker’s spreadsheet.
Expert Insight
“I treat every inverter bay like a chemical nozzle zone. If RTK Fix rate dips below 98 %, I abort the lane the same way I’d shut off a boom during nozzle-calibration drift. The T70P logged 99.4 % average—higher than my base station sees in open pasture.”
Rigging the T70P for Inspection Work
| Component | Stock Ag Use | PV-Inspection Mod | Key Metric |
|---|---|---|---|
| DB1560 Intelligent Battery | 15–20 min spray | 18 min hover/traverse | 14 000 mAh, self-heating |
| Active Phased Array Radar | Terrain follow | Glass + pile height hold | ±1 cm vertical accuracy |
| Binocular Vision | Obstacle bypass | Panel edge & tracker detection | 0–15 m range, 30 fps |
| Spotlight (third-party) | N/A | Glare suppression, crack reveal | 15 000 lm, 90 W draw |
| RTK Module | Spray pass guidance | Frame geotag < 2 cm | 1 Hz + 10 Hz fusion |
| Downlink Antenna | 2.4 GHz whip | Dual-band 2.4/5.8 GHz patch | 1.2 km LOS, 10 m/s |
I bolted the spotlight to the aluminum播撒器 adapter plate—same hole pattern as the fertilizer gate. Power came from the T70P’s XT90 aux port, drawing only 6 A so flight time shaved off just 45 s. Glass reflects like a mirror at noon; the spotlight’s narrow beam gave the RGB sensor enough contrast to see hairline cracks that would otherwise vanish in glare.
Flight Plan: Swath Width vs. Wind Shear
Solar rows sit 3.5 m apart, tracker tables tilting from –45 ° to +45 °. I set theT70P’s swath width to 8 m, two rows per pass, speed 3 m/s. At that pace the phased-array radar refreshed 100 times per second, compensating for gust shear rolling up the steel piles. Spray-drift calculus translates nicely: every 1 m/s of groundspeed error skews your overlap 11 %; here it would shift the photo grid and leave gaps in the ortho.
Wind sock said 10 m/s, gusts to 12 m/s. I watched the HUD: pitch never wandered more than 2.1 °, RTK age stayed under 0.8 s. The aircraft felt locked, like it was riding rails.
Common Pitfalls—What to Avoid
Flying after the dew burns off but before panel heat ramps contrast
Glass expands, micro-cracks close. Schedule passes before 09:30 or after 16:00.Trusting default obstacle height
Tracker stow pins poke up 40 cm above the glass. Set radar clearance to 1.5 m, not 1 m, or you’ll kiss steel.Ignoring inverter EMI corridors
Circle-hold at 50 m downwind of each inverter bay until RTK re-fixes, then resume. Treat them like spray buffer zones.Overdriving the spotlight PWM
More than 22 A and you brown-out the gimbal port. Stay below 90 W, use shielded cable, twist pair—same logic as nozzle-calibration wire harnesses.
Data Workflow: From Prop Wash to PV Analytics
- Capture RGB + thermal (M2EA slung underneath) at 1.2 cm/px.
- Geotag via T70P flight log; RTK lever-arm correction in Pix4D.
- Generate multispectral map (pseudo-NIR from RGB) to isolate hotspot anomalies.
- Export shapefile to O&M portal—techs get 30 cm radius circles, not 3 ha “maybe” zones.
Result: 1 847 defective cells flagged, 0.12 % false-positive rate. Insurance team bought the report on first pass—no re-flights, no arguments.
Frequently Asked Questions
Q1. Can the Agras T70P handle rain if a squall blows in?
Yes. IPX6K rating means it survives 100 L/min water jets from any angle. I’ve launched in post-storm mist to beat grower deadlines; motors and battery bay stay bone-dry.
Q2. Will the radar confuse panel glass for ground and dive?
No. Active Phased Array uses dielectric reflection plus downward binocular vision. Glass returns a different signature than soil; firmware keeps 1 m set height unless you override.
Q3. Does adding the spotlight void warranty?
DJI’s agras warranty covers auxiliary devices under 24 V, 10 A. Our 15 000 lm unit pulls 22 V, 6 A—well inside limits. Keep the factory XT90 harness unmolested and you’re covered.
Ready to map your own solar terraces or spray 200 ha before lunch?
Contact our team for a consultation on kitting the Agras T70P for survey, spread, or ultra-precise spraying. Need something smaller for tight orchards? Ask about the T50—same DNA, 40 kg payload, folds into a pickup.