handheld sniffer comparison  /  4 september 2026

In the hand,which one hears more?

These are handheld tools, so the comparison that counts is both of them held. A WLANPi Go against a NanoPi Zero2 with Osprey antennas, same Intel BE200, same channel plan, minutes apart. A free-standing sweep of each was captured as well, which is what lets the hand be separated from the hardware.

Primaryboth hand-held
Metricp95 RSSI, beacons only
Paired72 access points
Networks37, names obscured
RadioIntel BE200 on both
Positive numbers mean the NanoPi heard it louder. SSIDs are replaced by stable pseudonyms; every “Network 07” is the same real network throughout.

Held against held

the comparison that matters

As actually used — in a hand, phone attached — the NanoPi with Osprey antennas is +3.6 dB on 2.4 GHz and level on 5 and 6. Overall median across 72 paired access points: +1.1 dB.

2.4 GHz
+3.6dB
median, NanoPi minus Go
20 NanoPi1 tie4 Go
25 paired · spread -5 to +10 dB
5 GHz
+0.0dB
median, NanoPi minus Go
14 NanoPi11 tie14 Go
39 paired · spread -8 to +8 dB
6 GHz
-0.2dB
median, NanoPi minus Go
2 NanoPi4 tie2 Go
8 paired · spread -5 to +5 dB

The 2.4 GHz result is the decisive one: 20 of 25 access points louder on the NanoPi, only 4 on the Go. On 5 GHz the two are genuinely interchangeable — 14 to 14 with 11 ties.

Across the spectrum

median p95 per frequency, both held
2.4 GHz 5 GHz 6 GHz -25 -30 -35 -40 -45 -50 -55 -60 -65 -70 -75 -80 -85 -90 p95 dBm 2412 2432 2437 2447 2462 5180 5220 5240 5280 5300 5580 5600 5660 5700 5745 5785 5975 6215 6295 6615 6775 6935 MHz →
 WLANPi Go  NanoPi + Osprey marker size = access points behind the point

Where the two lines sit apart, one device is genuinely hearing more at that frequency. The overall up-and-down is the neighbourhood, not the hardware — different access points live on different channels — so read the gap between the lines rather than the height of either. The smallest markers rest on a single access point and should not be read as a trend.

Every access point, both held

hollow = Go · filled = NanoPi
-90 -80 -70 -60 -50 -40 -30 p95 RSSI (dBm) — right is stronger 2.4 GHz Network 10 · ch1 +3.0 Network 11 · ch1 +3.3 Network 15 · ch1 +3.8 Network 23 · ch1 +3.5 Network 12 · ch5 +4.8 Network 10 · ch6 +9.0 Network 15 · ch6 +9.0 Network 23 · ch6 +8.0 Network 25 · ch6 +3.6 Network 27 · ch6 -2.0 Network 27 · ch6 +6.0 Network 27 · ch6 +6.9 Network 28 · ch6 +1.3 Network 31 · ch6 +7.2 Network 34 · ch6 +7.1 Network 35 · ch6 +9.0 Network 33 · ch8 +10.0 Network 10 · ch11 -5.0 Network 10 · ch11 +3.0 Network 15 · ch11 +3.0 Network 15 · ch11 -5.0 Network 20 · ch11 +0.0 Network 23 · ch11 -5.0 Network 23 · ch11 +3.0 Network 26 · ch11 +4.8 5 GHz Network 01 · ch36 -8.3 Network 01 · ch36 -3.0 Network 35 · ch36 +1.0 Network 11 · ch44 -1.0 Network 28 · ch44 +0.8 Network 19 · ch48 +2.0 Network 09 · ch56 -2.2 Network 15 · ch56 -2.2 Network 16 · ch56 -2.2 Network 23 · ch56 -2.2 Network 36 · ch56 -2.2 Network 37 · ch56 -2.2 Network 09 · ch60 +0.0 Network 15 · ch60 +0.2 Network 16 · ch60 +0.0 Network 23 · ch60 +0.0 Network 36 · ch60 +0.0 Network 37 · ch60 +0.0 Network 03 · ch116 +8.0 Network 09 · ch120 -2.6 Network 15 · ch120 -2.6 Network 16 · ch120 -2.6 Network 23 · ch120 -2.6 Network 36 · ch120 -3.0 Network 37 · ch120 -3.0 Network 03 · ch132 +2.0 Network 09 · ch140 +2.0 Network 15 · ch140 +2.0 Network 16 · ch140 +2.0 Network 23 · ch140 +2.0 Network 36 · ch140 +2.0 Network 37 · ch140 +2.0 Network 25 · ch149 +0.0 Network 28 · ch149 -0.7 Network 20 · ch157 +5.2 Network 21 · ch157 +5.2 Network 22 · ch157 +5.2 Network 28 · ch157 +1.3 Network 31 · ch157 +1.3 6 GHz Network 02 · ch5 -5.2 Network 31 · ch53 -3.2 Network 24 · ch69 +4.8 Network 24 · ch133 +1.0 Network 24 · ch165 +3.0 Network 03 · ch197 -0.2 Network 04 · ch197 -0.2 Network 24 · ch197 -0.1
 WLANPi Go  NanoPi + Osprey green line = NanoPi louder orange = Go louder grey = within 1 dB

What the hand costs each device

same device, free-standing minus held

The free-standing sweeps exist to answer this: how much does each device lose when you pick it up? Comparing a device only against itself removes every other variable.

Device2.4 GHz5 GHz6 GHzOverall
WLANPi Go+2.0+1.3+5.0+2.1
NanoPi + Osprey+1.5+1.0+1.4+1.3
A hand costs the Go roughly twice what it costs the NanoPi

+2.1 dB against +1.3 dB overall, and on 6 GHz +5.0 against +1.4. The Osprey antennas stand proud of the case on their own connectors; the Go’s are films bonded to the inside of the shell, under the palm.

This is why the held comparison is the fair one for a handheld tool rather than a handicap against the Go. Being harder to detune is part of the hardware.

A WLANPi Go opened up and laid out as it is held, lid on the left and board on the right, with annotations. Two thin coax cables leave the Intel BE200's MAIN and AUX connectors and cross into the lid, one ending at the top edge of the case and one at the side edge. Three discs in the middle of the lid are labelled as mounting magnets.
Where the Go’s antennas actually are. Two coax runs leave the BE200’s MAIN and AUX connectors, cross into the lid, and terminate against the case wall — one at the top edge, one at the side. Held normally, the top element stays clear of you and the side element sits directly under your palm, which is a plausible mechanism for the Go losing roughly twice what the NanoPi does when picked up. The Osprey antennas screw onto external connectors and stand clear of the case in any grip, so there is no equivalent element to cover. The three discs in the middle of the lid are mounting magnets and radiate nothing. The photo resolves the coax terminations rather than the elements themselves, which are bonded flat to the inside of the wall. Card identifiers are blurred; the labels are ours, nothing else is retouched.

Free-standing, for reference

supporting data

With both devices set down and untouched the gap narrows, exactly as the hand-cost table predicts: 2.4 GHz +2.1 · 5 GHz +0.6 · 6 GHz +0.5, overall +1.0 dB across 81 paired access points. The NanoPi keeps a real 2.4 GHz lead; roughly a third of its advantage when held is the hand rather than the antennas.

Condition2.4 GHz5 GHz6 GHzOverall
Both held+3.6+0.0-0.2+1.1
Both free-standing+2.1+0.6+0.5+1.0

How many each one found

APs with 5+ beacons on their home channel
ConditionWLANPi GoNanoPi + Osprey
Held8277
Free-standing8982

Against our own interest, and worth stating plainly: the Go counts more access points in both conditions even while reading slightly quieter on the ones both hear. Signal strength and how many APs clear the detection threshold are different measurements, and this run does not disentangle them.

The full table, both held

p95 dBm
NetworkBandChGoNanoPiΔbeacons
Network 102.41-65.0-62.0+3.026/26
Network 112.41-78.3-75.0+3.327/19
Network 152.41-65.0-61.2+3.825/26
Network 232.41-65.0-61.5+3.525/31
Network 122.45-82.8-78.0+4.819/21
Network 102.46-60.0-51.0+9.028/25
Network 152.46-60.0-51.0+9.022/29
Network 232.46-59.0-51.0+8.029/28
Network 252.46-59.0-55.4+3.627/28
Network 272.46-51.0-53.0-2.030/24
Network 272.46-66.0-60.0+6.027/13
Network 272.46-57.0-50.1+6.929/23
Network 282.46-57.3-56.0+1.327/28
Network 312.46-67.2-60.0+7.224/25
Network 342.46-70.5-63.4+7.126/28
Network 352.46-86.0-77.0+9.024/19
Network 332.48-78.0-68.0+10.018/17
Network 102.411-64.0-69.0-5.022/22
Network 102.411-39.0-36.0+3.029/26
Network 152.411-39.0-36.0+3.029/25
Network 152.411-64.0-69.0-5.022/25
Network 202.411-54.0-54.0+0.023/29
Network 232.411-64.0-69.0-5.022/27
Network 232.411-39.0-36.0+3.027/27
Network 262.411-79.0-74.2+4.821/26
Network 01536-52.0-60.3-8.325/27
Network 01536-67.0-70.0-3.026/28
Network 35536-90.0-89.0+1.023/32
Network 11544-84.0-85.0-1.026/26
Network 28544-85.0-84.2+0.826/26
Network 19548-88.0-86.0+2.028/24
Network 09556-57.0-59.2-2.226/25
Network 15556-57.0-59.2-2.226/25
Network 16556-57.0-59.2-2.226/25
Network 23556-57.0-59.2-2.227/25
Network 36556-57.0-59.2-2.226/25
Network 37556-57.0-59.2-2.227/25
Network 09560-53.0-53.0+0.026/24
Network 15560-53.2-53.0+0.226/25
Network 16560-53.2-53.2+0.026/25
Network 23560-53.2-53.2+0.026/25
Network 36560-53.0-53.0+0.026/25
Network 37560-53.0-53.0+0.026/24
Network 035116-36.0-28.0+8.024/25
Network 095120-38.4-41.0-2.629/24
Network 155120-38.4-41.0-2.629/25
Network 165120-38.4-41.0-2.629/25
Network 235120-38.4-41.0-2.629/25
Network 365120-38.0-41.0-3.029/25
Network 375120-38.0-41.0-3.029/23
Network 035132-62.0-60.0+2.024/25
Network 095140-65.0-63.0+2.028/25
Network 155140-65.0-63.0+2.028/25
Network 165140-65.0-63.0+2.028/24
Network 235140-65.0-63.0+2.028/25
Network 365140-65.0-63.0+2.025/25
Network 375140-65.0-63.0+2.028/25
Network 255149-56.0-56.0+0.027/27
Network 285149-55.3-56.0-0.727/27
Network 205157-61.2-56.0+5.226/27
Network 215157-61.2-56.0+5.226/27
Network 225157-61.2-56.0+5.226/27
Network 285157-70.3-69.0+1.328/28
Network 315157-70.3-69.0+1.328/28
Network 0265-54.0-59.2-5.220/25
Network 31653-85.0-88.2-3.221/25
Network 24669-48.0-43.2+4.821/26
Network 246133-69.0-68.0+1.027/26
Network 246165-59.0-56.0+3.020/26
Network 036197-49.0-49.2-0.225/26
Network 046197-49.0-49.2-0.225/26
Network 246197-61.9-62.0-0.120/25

How the numbers were made

method

Beacons only, p95 of RSSI in dBm. Access points are paired by BSSID and frequency, never by name, because multi-BSSID radios reuse one SSID across bands.

Each BSSID gets a home frequency — the modal one across all four captures, so every sweep is cut identically — and beacons heard while the radio was tuned elsewhere are discarded. A pair needs 5+ beacons on both sides. Hidden and malformed SSIDs are excluded.

All four captures were checked to be single-radio and tri-band before anything was compared: one interface block each, roughly 50 frequencies, reaching 7095 MHz. That check is not optional. A device that splits its sweep across two radios — a BE200 and a built-in Broadcom, say — pools into a figure that compares one vendor's radio against another's on a third of the channels.