Naked Secondary
Do we at least get double insulation on the secondary? As you may have assumed after seeing the price, the answer is an all too predictable no.
Since I have yet to see a transformer without double insulation pass my isolation withstand test, I have no reason to believe this transformer will fare any better.
Who Left Out The Lights?
While the USB ports feature slightly diffuse transparent plastic components, and the LED footprints on both sides (bottom corner) clearly indicate that these were meant to be illuminated, the LEDs themselves were omitted from this variant. Considering that the contact support structure for the ports is surrounded by a metal shield (except on the back), the LED footprints’ locations do not make much sense.
In the unlikely event that this adapter survives my testing, I may add LEDs to see how they look.
Board Bottom
On the bottom, we find an attempt at separating the primary side from the output by removing some of the solder mask along the boundary, a diode bridge, a 330kΩ bootstrap resistor for the ND5773, a 1.9Ω primary-side current-sensing resistor, and miscellaneous other resistors. I had no luck finding datasheets for the ND5773, meaning it's likely a custom-marked chip.
Separation between primary and secondary appears to be at least one whole millimeter, which makes it double that of the A1265 look-alikes. This is better, but still not good enough.
Charger ID
What differences are there between the 2.1A “iPad” port and the 1A port? None. The two ports’ +5V pins (first horizontal pair at the top) are tied together, both ports’ D+ and D- pins are tied together in the middle, sharing a single 470kΩ resistor to ground, and both of their ground pins are tied together at the bottom.
While the soldering quality looks generally decent, there are tiny solder beads all over the place, one of the largest painted over in red marker left from the middle.
In case you are wondering, the unpopulated footprint to the right is for the omitted LEDs’ current-limiting resistor.
Standby Power
At 115V AC input, the AR-18’s standby power barely registers above my measurement's noise floor at 27mW, dropping to an even lower 14mW under 230V input. I had to check that my sense amplifier was actually turned on and repeat the measurement a few times to convince myself that this was actually real.
With noise accounting for most of my signal, I have no real accuracy on those two readings. Since there is a 1kΩ bleed resistor on the output, the absolute minimum possible power consumption is 25mW.
Efficiency
Results under load confirm that the AR-18 is actually capable of decent efficiency by beating Level VI requirements, which is exactly what I would expect from an adapter based on a monolithic flyback controller. Far less impressive is that the AR-18 only manages 960mA of output current on 115V input before its output trails off below 4.75V, making it a 1A output adapter at best.
Output Noise Waveform
What happens when you take a flyback converter and give it only one ordinary electrolytic output capacitor for output filtering? You get a wall of switching transients dominating the noise landscape. Thankfully, this is much better than usual for cheap adapters, as you’ll see when I bring in some past results for comparison.
Peak-to-Peak Output Noise
Where peak-to-peak noise is concerned, the AR-18 fares considerably better than its ~$3 competition. I’m going to guess that a large chunk of this is due to its single-board construction significantly shortening the distance between the transformer, output diode, and output capacitor.
RMS Output Noise
The AR-18 easily outperforms its competition in the RMS department thanks to a monolithic flyback regulator doing a far better job of regulating output voltage against ripples across its under-sized input capacitor than the other adapters’ simple two-transistor oscillator circuits.
Output Voltage Regulation
From how the AR-18’s controller raises its output voltage with load, we believe that it provides up to about 150mV of cable drop compensation. No problem; cable losses will easily knock that 5.3V maximum down to less than 5.25V.