18 s
keys: R replay · Space pause · H record mode · arrows step · URL: ?rec&dur=18&paused&loop

Seventeen claims, checked one by one.

From the listing and the box · asterisks are soft passes, detail in the master · the hardware keeps passing; what fails is what was written
Pass
0
measured or verified on this unit
Fail
0
every one of them something written
Could not verify
0
below what this bench or this room resolves
What they builtmeasured on this unit
✓8000 Hz pollingwired and 2.4 GHz, bInterval=1
✓Five-layer damping *least 2 to 8 kHz ring-out of the three
✓Over 1.2 kg1.28 kg on the scale
✓108 keys in calibrationboth extremes, no phantoms
✓NKROten keys, all registered
✓Pre-lubed switcheslube visible on the stabs
✓Numpad stabilisedall three 2u keys
✓Fn layer without the driverevery shortcut as printed
✓10000 mAh50.7 Wh in, two cells
✓0.005 mm steps *in the software, hardware unverified
✓0.1 to 3.3 mm range *3.3 needs the bottom deadzone at 0
✓32 h with RGB on *consistent with a three-day run
What they wrotedoes not survive checking
✗500 h with RGB off50 to 70 h at the measured draw
✗Shine-through legendsopaque; they block the light
✗Double-shot legendsprinted on, you can feel the ink
Could not verifyfrom outside
?128 kHz scan rateinside the firmware
?0.2 ms latencybelow the bench; consistent with 125 µs polling
EVERYTHING MEASURED CAME BACK CLEAN EVERYTHING THAT FAILED IS SOMETHING WRITTEN, NOT SOMETHING BUILT
Soft passes: 0.005 mm verified at the software level only; 3.3 mm reachable only with the bottom deadzone at 0, which the app warns against; 32 h consistent with a three-day run at full brightness, not measured to the hour.
Damping is a soft pass: across the whole spectrum all three boards are into the room noise within 50 ms, but in the 2 to 8 kHz band the HE108 is back at the room floor by 69 ms while the Razer is still 6 dB above it at 80 ms. Unverifiable from outside: 128 kHz scan rate, 0.2 ms latency. Full evidence per row in the master script’s claim scorecard · CK Tech Check