Testing: Charger Doctor and USB Detector Accuracy

Ever since I got my Charger Doctors/USB detector, I have been fascinated with how they have been able to provide such a useful tool for a low price. After taking each of them apart and analyzing their circuitry and connections, I had some idea of their underpinnings.

In my brief review, my experience was that for most things, it was within about 2-3 digits of the real value making it useful for diagnostics but not for precision measurements. But I also noted issues with burden voltage in both units, and I didn’t really check their accuracy over the whole range.

Seeing as I had some equipment at my disposal, and the curiosity was still raging inside me, I decided I might as well give it a good test against some really pricey (and accurate) gear.

Testing Methodology

If you want to check the accuracy of something, it pays to use something more accurate and assured. Testing a 3-digit device like these against a 3.5-digit $10 multimeter isn’t likely to yield any really useful information, as they are likely to have errors of a similar magnitude (i.e. several digits). With this in mind, I decided to run the following quick checks.

Open Circuit Voltage Indication

One of the usage scenarios of the device is just to measure the voltage from a port or power supply without any load on the downstream ports. In order to test the accuracy, I used a Manson HCS-3102 switch-mode benchtop power supply, hooked up to a header-to-USB-A-F port I stole from an old computer case. The charger doctor/USB detector was plugged into this port.

The power supply was set to its lowest voltage (0.8V) and the voltage was increased by 0.1V until it begins to display. Recordings were taken of the displayed voltage value, and the actual voltage value measured by a Tektronix PA1000 Power Analyzer using probes directly touching the back of the USB-A-F port itself. The Tektronix provides a 4-digit readout at this range, with a 0.05% basic accuracy.

The voltage was incremented by 0.1V and the process repeated until the unit was demonstrably saturated (i.e. indicated voltage does not increase with increasing input voltage).

Current and Voltage Error at Load

The main usage scenario involves using the charger doctor/USB detector to measure the voltage and current during charging.

In order to best simulate that, the same equipment above was used, with the USB power bank testing rig configured as a load. The voltage was fixed at 5v on the power supply, with the actual voltage measured at the input terminals to the test device compared with the displayed voltage shown by the tester. The actual current was measured by the PA1000 using the 20A internal shunt in series with the load, and this was recorded with the displayed current for comparison.

The load currents were cycled through all combinations for anywhere between about 0.5A to 3A.

Current Consumption of Testing Device

Knowing the parasitic consumption of the test device can come in handy in some cases. The device was supplied with 5v (+/-0.05v) from a Manson HCS-3102 benchtop power supply, hooked up to a header-to-USB-A-F port. The Tektronix PA1000 was put in series using the A1A shunt to measure the current consumed.

Results

Testing of the devices resulted in some interesting results and some unusual behaviour. It was found that the blue USB charger doctor tends to show jumpy voltage numbers below about 5v, making testing a little bit annoying. The black USB detector instead hangs at certain voltage levels and requires some time or power cycling to read the correct value in the case the voltage is changed during testing.

Current Consumption of Testing Device

It was found the blue USB charger doctor consumes 26.200mA, with the black USB detector consuming 15.566mA at a supply voltage of 5V.

Open Circuit Voltage Indication

The USB charger doctor has a stated voltage operation range of 3.5v to 7v, and was found to operate (show a display) starting at 3.2v, reaching saturation at an indicated 7.46v.

The USB detector is specified to operate between 3.2v to 10v, and was found to operate (show a display) at 1.9v, reaching saturation at an indicated 9.99v.

This shows that the black USB detector unit has a much wider operational voltage, but both unit span the critical 4-6v that you’d expect your ports to show.

A plot of the delta (measured minus displayed voltage) shows the following rather interesting trend.

Voltage-Error-at-No-Load

The USB detector shows a very consistent and low amount of error with a positive bias (offset error) throughout a large part of its range. When checked, for most readings it was pretty much 2 to 3 digits as I had observed earlier. Above 9.9v, it was rising due to saturation, and below 2.2v, it was likely due to the internal voltage reference “sagging” or not reaching its regulated value. It’s actually a pretty good performance.

The blue charger doctor was much more limited in its voltage range, and starts showing errors below 4.6v, where the displayed voltage overestimates the supply voltage. This is also probably due to the voltage reference not being stable, but it’s important to know. Above 7.4v, the error rises due to the unit saturating.

Voltage-Error-at-No-Load-Percentage

When plotted as a percentage error (absolute value), you can see the USB detector does a great job staying under 1% of error for a large portion of the voltage range. The line itself has a slightly negative gradient because a fixed offset error is a smaller portion of the input voltage for a greater voltage, thus the error appears to reduce when taken as a percentage.

The charger doctor is almost at 3% of error at 4v, which is a bit high, but it too settles to about 1% for the main “flat” portion of its range.

Current and Voltage Error at Load

Both units are specified for reading loads from 0-3A, and thus, they were tested at a load of about 500mA to 3A. Actual measured currents and voltages were used for the plot, and due to the resistive nature of the load, the actual currents measured for the two charger doctors are not identical but close.

Current-Error

Overall, the actual error on the blue USB charger doctor when it comes to current is quite low. It’s actually pretty much “on the money” with some variations mainly due to the granularity in the 3-digit display.

The black USB detector, on the other hand, seems to have a case of over-reporting current at higher current levels above 2A. This suggests it may be some non-linearity in the amplifier or measuring additional voltage drop due to traces (i.e. not only the shunt). The fact it doesn’t seem to be a consistent line suggests a non-linear effect.

Another big issue is the voltage burden. This is the voltage “lost” through the charger doctor/USB detector. I could have determined this better if I had disassembled the units again and measured the voltage on the output terminal versus the input terminal, but I was lazy so this is only an approximate value. This was measured from the input USB A-F terminal, versus the indicated voltage value which has some error.

Burden-Voltage

We notice that the charger doctor seems to have a voltage burden which levels off at higher currents. This is probably an artifact due to the over-reporting of voltages when it falls below 4.6v. But the striking problem is the USB detector which shows an alarming voltage burden – with a source voltage within a hair of 5v, the indicated voltage was about 4.3v at 3A. The voltage loss is probably due to the thin input cable and connectors, which kind of defeats the purpose of being an accurate diagnostic tool and a “doubling” adapter for a charger. This is disappointing, but it is close to my estimated value of about 1v burden due to the thin cables with my earlier observation.

Current-and-Voltage-Error-at-Load-Percentage

In this, we can see the combined graph which shows the voltage and current error (i.e. comparing measured voltage or current value vs displayed voltage or current value). It seems the Charger Doctor is superior, showing currents which are mostly within 1% of the actual value, and voltages which are within about 3.5% for an input of 5v (+/-0.05v) and currents between 500mA to 3A.

The USB detector suffers badly in voltage due to the loss of voltage in its short pigtail lead (and possibly even in the PCB) and suffers in current due to over-estimation at higher currents. It’s a bit night-and-day!

Conclusion

It’s not everyday when pieces of gear worth less than AU$2 or AU$8 get tested with a $3000 power analyzer. It seems that when I first wrote my review, noting that each had it’s strength and weaknesses, I was more accurate than I had intended.

The black USB detector is superior at voltage readings at no load current, showing a small fixed offset and a wide operating voltage range. The blue charger doctor has a smaller operating voltage range and overestimates voltages below 4.6v, which may be of significance when it comes to USB voltages.

However, when we look at voltages and currents under loading, it can be seen that the black USB detector falls far short for accuracy, with a high loss of voltage through the unit itself (probably due to wire and PCB design) and over-estimation of current at higher loadings which makes the double-adapter nature of the unit much less appealing. The blue charger doctor seems to handle these operating conditions much better with a higher accuracy and less voltage loss.

Looking at current consumption, it seems the black unit consumes slightly less power from the source, consuming 15.566mA where the blue charger doctor consumes 26.200mA.

But when it comes to accuracy, I suppose you get what you pay for. Under the right circumstances (e.g. 2A load), it may even mislead some users into thinking their chargers are putting out too little voltage, or force devices to charge slower due to voltage burden and thus report a lower charging current than the user would expect. Users should probably be aware of this.

About lui_gough

I’m a bit of a nut for electronics, computing, photography, radio, satellite and other technical hobbies. Click for more about me!

This entry was posted in Electronics and tagged , , , , . Bookmark the permalink.

6 Responses to Testing: Charger Doctor and USB Detector Accuracy

  1. JT says:

    This was incredibly useful – I bought the blue one based on your careful, thoughtful review. Thank you Gough.

  2. A bit late to the party I know, but I agree this is very interesting thanks. I just bought a clear one which otherwise looks the same as your blue one.

    I just fired it up for the first time on my phone charger, and the indicated current draw (at 5.05v) is 0.74A. However, my Ampere app on the phone suggests it’s charging at 1.4A (and indeed, that is the current I expected base on my kernel settings). For reference, the charger (a Belkin) is rated at 2.1A max. Can you think what might be causing this discrepancy please?

    • lui_gough says:

      I don’t have time to answer all questions in detail, but quick points would be:
      – Adding the charger doctor in line does cause additional voltage drop in sense resistor and contacts, which may reduce charging current due to loss of voltage.
      – Some charger doctors (but not the one I tore apart here) modify the signalling lines (D- and D+) and may cause a phone to misdetect the power supply as not a fast charger, and likewise, a non-official charger can be mis-signalling and thus not all the power capability is used.
      – Ampere app is not highly accurate, reading offsets of 0.3A are not unusual and highly dependent on the power management ICs used in your phone.
      – Many phones don’t take more than 1.2A under ideal situations – the only devices that do take more than that are generally tablets.
      – Testing with different remaining battery capacity and different application load will result in different consumption, as different charge rates are experienced (trickle at the beginning, largest in the bulk charge phase, tapering off once completed).
      – Charge current waveform could be significantly “peaky” and the charger doctor may not be sampling those “peaks” very well, resulting in under-reporting or fluctuating displays.

      That being said, the unit is decently accurate, but having ANYTHING measuring in series causes a burden voltage due to the introduced resistance – thus is the art of optimizing burden voltage vs noise in the measurement to have the least impact on the device under testing while producing an accurate enough reading. As a result, plugging in a highly accurate multimeter in series might make things worse, as they have considerably larger burden resistances, thus current measurement is difficult to get right.

      – Gough

  3. Postscript – with genuine Samsung cable, same charger charges phone at 1.27A. Same result at home with Apple charger. None of the many other micro USB cables I have lying around can charge it at >1A, so I will definitely be getting another Samsung cable or two!

Error: Comment is Missing!