RMS Output Noise
The SPS8038B does not break new ground in terms of RMS noise either, mostly because its two large electrolytic capacitors have much higher ESR than Aukey's and SilverStone’s polymer caps.
Why did the peak-to-peak performance get worse below 1.25A when the overall RMS performance improved? Peak-to-peak noise is dominated by switching transients. If the input capacitors' ESR improved from reconditioning, it would cause harsher transients, which explains both noise results and the disappearance of noise bursts timed with AC peaks.
MORE: Tom's Teardown: SilverStone UC01 USB Power Adapter
MORE: How (And Why) We Test USB Power Adapters
Scrutinizing Noise
Noise caused by ESR is the extra voltage that appears while either shoving current into a capacitor or drawing current from it due to impedance. The line in red approximates an ideal capacitor’s voltage, while the yellow zone is extra voltage from the SPS8038B capacitors’ resistance.
What does this tell us about the KF capacitors’ health? There are two of them in parallel rated at 76mΩ feeding a 1.5A load. Since the transformer is directly powering the load ~45% of the time, average current going into the capacitors must exceed 1.5A and the peak be greater still. Sixty millivolts at 1.5A is already 40mΩ. Conclusion? The capacitors’ combined ESR is well under 40mΩ, meaning they’re still well within their passable specs.
MORE: Tom's Teardown: SilverStone UC01 USB Power Adapter
MORE: How (And Why) We Test USB Power Adapters
Transient Response
Transient response looks fairly typical with a 180mV dip upon application and 150mV overshoot upon removal. For comparison’s sake, Aukey scored a 150mV dip and 150mV overshoot.
While both units settled after the dip within about two milliseconds, Philips' SPS8038S beats the PA-U32 by having its overshoot corrected within three milliseconds versus the better part of 10 for Aukey’s, likely because it has three times the total bulk output capacitance to absorb excess energy before the controller detects the load change.
MORE: Tom's Teardown: SilverStone UC01 USB Power Adapter
MORE: How (And Why) We Test USB Power Adapters
Short-Circuit Response
Upon applying a short circuit, current peaks at about 7.5A as the capacitors dump their charge into the short, promptly drops to less than 1.25A, slowly drifts down to 800mA, and cuts off after 50ms. Once over-current protection is tripped, the controller waits approximately one second before restarting.
MORE: Tom's Teardown: SilverStone UC01 USB Power Adapter
MORE: How (And Why) We Test USB Power Adapters
The Replacement Test
What do you do when you have a growing inventory of spare parts and wonder if a capacitor swap would do any good?
Bye bye 76mΩ CapXon KF with 1ARMS ripple rating; hello 45mΩ Rubycon ZLS with 1.2ARMS ripple rating.
MORE: Tom's Teardown: SilverStone UC01 USB Power Adapter
MORE: How (And Why) We Test USB Power Adapters
Moment Of Truth
Well, 45mΩ is 60% of 76mΩ, and 40mVpk is 66% of 60mVpk, so the result is within what you’d expect from specified tolerances. Including cycle-to-cycle noise instead of focusing on a single cycle’s ESR hump, the capacitor mod reduces total noise by 32mVPP (22%) and 8mVRMS (25%), which is about 15% short of perfect scaling and a reminder that output capacitors can only do so much about mitigating noise.
Curiously enough, between my initial results and post-mod ones, standby power increased from 93mW to 134mW. I wonder what else changed?
MORE: Tom's Teardown: SilverStone UC01 USB Power Adapter
MORE: How (And Why) We Test USB Power Adapters
Revisiting Efficiency
What sort of efficiency gains do you get from reducing output capacitor losses by approximately 30%? Apparently nothing compared to the original capacitors post-reconditioning. If anything, I was expecting efficiency to improve. But it actually got as much as 5% worse, with both the burn-in and modded results remaining nearly flat at 63%. These are disappointing outcomes.
Also, the maximum sustainable load dropped from 1.95A in my initial tests to 1.9A post burn-in, and again down to 1.85A after the re-cap.
MORE: Tom's Teardown: SilverStone UC01 USB Power Adapter
MORE: How (And Why) We Test USB Power Adapters
Isolation Withstand Test
Does the SPS8038B’s transformer (the only thing across the isolation boundary) manage to pass my 3.5kVAC withstand test? As you should expect from a major brand, it does indeed.
I do see a possible issue, though: the spark gap should be the weakest point across the boundary, but I swear I could hear something within the transformer about to yield. The spark gap may have needed to be a millimeter or so narrower, hugging the hole instead of being some distance away from it.
MORE: Tom's Teardown: SilverStone UC01 USB Power Adapter
MORE: How (And Why) We Test USB Power Adapters
The Verdict?
The SPS8038B delivered 1.89A out of its 2.1A claimed output rating on its worst run. That's disappointing, but not too far off the mark. Although its output noise is the worst among AC-USB power adapters that have survived my testing, at least it remained within acceptable limits for USB after the burn-in reconditioning.
My main issues with the SPS8038B are poor efficiency at no more than 63% after burn-in, when even Level IV efficiency requires 64% for a 10W adapter, plus those random metal bits on the transformer’s tape.
Is the convenience of a device tray worth picking one of these up? Maybe, if you are into these sorts of things and don’t mind the ~$2/year impact on power and outdated resistor-based charger identification.
MORE: Tom's Teardown: SilverStone UC01 USB Power Adapter
MORE: How (And Why) We Test USB Power Adapters