Designing a proper USB powered device is no simple task, this is because:

1. If the device needs more the 100mA current, it is only allowed to draw that amount after requesting from host during USB enumeration. Some USB hosts, for example the down stream port on a bus-powered USB hub, is only able to provide maximum 100mA. So the device must be intelligent enough to differentiate the host types.
2. USB sleep mode current limitation must be considered as well.
3. I also wish the device to operate autonomously without a PC. That means it must accept power from “dummy” USB hosts such as a USB charger.

Clearly point (3) conflicts with points (1) and (2) because USB charger cannot be enumerated and does not have sleep mode. And this is why I’m interested in FT230XS’s “USB charger detection” function.

A thorough test is still needed. According to the datasheet, FT230XS supports detection of battery charger which conforms to the USB BC1.2 specification. AFAIK many USB chargers, notably those made by Apple, are not BC1.2 compatible. Whether FT230XS can detect those chargers is not specified in the datasheet.

A quick background about USB power supply and USB charging: the original USB specification limits the USB downstream current to 500mA (after enumeration).  However 500mA is too low to charge battery powered devices such as a smartphone. BC1.2 was then specified, which introduced “Charging Downstream Port (CDP)” and “Decicated Charging Port (DCP)”. A DCP internally shorts “D+” and “D-” together. When DCP is detected, the device may draw as much as 1.8A from the port. Identification for CDP is much complicated and outside the scope of my interest  (because I have no CDP compliant host), more readings can be found here.

While looks simple on paper, the actual USB charging situation is much more complicated. Reason being that USB charging specification came too late and vendors have been inventing their own charging protocols. Notably:

1. Apple has been inventing USB charging protocol since iPod days (check out the hardwork here). The charger’s capability is identified by the voltages on D+ and D- lines, as:
   • (D+,D-) = (2.0V, 2.0V) -> 500mA (iPod charger)
   • (D+,D-) = (2.8V, 2.0V) -> 1A (iPhone charger)
   • (D+,D-) = (2.0V, 2.8V) -> 2A (iPad charger)
2. Chinese standard YD/T 1591-2006. Which is basically the same as BC 1.2, shorting D+ and D-.
3. European smartphone charging standards says a 200Ω resister must be connected in-between D+ and D- to identify a charger.

FT230XS is very easy to use. I made a simple testing board using a SSOP-16 breakout adaptor. The contact was was plucked from a USB header and glued to the breakout adaptor. Chip under test is FT230XS, Rev D, date code 1243.

I further mount the adaptor onto a breadboard and connect 3 LEDs onto the CBUS pins. The configuration of the colors are:

• Yellow – BCD#, should light up when charger is detected.
• Red – PWEN#, indicates required power negotiated.
• Green – SLEEP#, should light up when USB enters sleep mode (host suspend).

The chip is programmed using official FT_Prog utility. Power required is set to 500mA. I would skip all the testing procedures and note down the results here:

Testing with PC USB port

Scenario 1. Connect to a PC USB port when PC is running: Green LED (SLEEP#) lit up then all LEDs turn off shortly. Windows start to find driver. Red LED (PWEN#) lit up when Windows starts to download driver from Windows Update. Red LED stays on after driver installed.

Scenario 2. While FT230XS still connected, put Windows into sleep mode: Red LED (PWEN#) off and Green LED (SLEEP#) lit up, indicating sleep mode.

Scenario 3. Resume windows: Green LED off, Red LED lit up, indicating power enable.

Scenario 4. Unplug FT230XS, put Windows into sleep mode and plug the FT230XS: Red LED off, Green LED lit up, indicating sleep mode.

Scenario 5. While FT230XS is still connected, shutdown PC: Green LED still on, indicating sleep mode.

Scenario 6. Unplug FT230xS, shutdown PC and plug FT230XS back: Green LED lit up, indicating sleep mode.

Testing with PC USB through a self-powered USB hub:

Same as previous results.

Testing with PC USB through a bus-powered USB hub:

Unfortunately all my USB hubs report themselves as self-powered even if they are not. I had to modify a hub into bus-powered. Bus-powered hub can be identified in Windows Device Manager as:

Scenario 1. Connect to bus-powered hub: Green LED (SLEEP#) lit up then all LEDs turn off shortly, Windows show the following message. All the LEDs stays off. It shows that Windows failed to configure the device because the USB port cannot supply enough current (500mA).

Scenario 2. While FT230XS still connected, put Windows into sleep mode: Green LED (SLEEP#) lit up, indicating sleep mode.

Scenario 3. Resume windows: All LEDs turn off.

Scenario 4. Unplug FT230XS, put Windows into sleep mode and plug the FT230XS: Green LED lit up, indicating sleep mode.

Scenario 5. While FT230XS still connected, shutdown PC: Green LED stays up, indicating sleep mode.

Scenario 6. Unplug FT230xS, shutdown PC and plug FT230XS back: Green LED lit up, indicating sleep mode.

Test with USB chargers

I tested 4 different chargers, as shown in the picture

Scenario A, plug FT230XS into a USB power adaptor from TPLink router (D+/D- shorted). Yellow LED (BCD#) lit up, indicating charger detected.

Scenario B, plug FT230XS into a USB charger for amazon Kindle (D+/D- shorted). Yellow LED (BCD#) lit up, indicating charger detected.

Scenario C, plug FT230XS into Apple iPhone charger (5V/1A), Yellow LED (BCD#) lit up, indicating charger detected.

Scenario D, plug FT230XS into Apple iPad charger (5V/2A), Green LED (SLEEP#) lit up, indicating sleep.

Scenario E, since I do not have a Apple iPod charger (my iPod charger is Firewire charger), I connect FT230XS to 5V supply and supply D+/D- with 2V from voltage divider. Green LED (SLEEP#) lit up, indicating sleep.

Scenario F, Connect FT230XS to 5V power supply (floating D+, D-), Green LED (SLEEP#) lit up, indicating sleep.

Summarize the findings, here is the result matrix: (On means asserted)

So the verdict? I need more intelligence from FT230X  for proper USB power management

The post FT230X charger detection investigation appeared first on Digital Me.

• First thing of course is the battery. For modern chargers it is usually a 3.6/3.7V lithium-ion or lithium-polymer type. The capacity of the battery directly determines the price of the charger. Overrating capacity is not unusual. As my personal guideline, for a charger weights about the same of an iPhone (120g), the capacity should be around 5000-6000mAH. So if you encounter a charger weights less but rated as much as 10,000mAH, avoid it at all costs. It is not because of the capacity, but the manufacture’s integrity and the confidence over other parts of the charger.
• Since the battery is 3.6/3.7V and USB specification is 5V, there should be a DC-DC boost converter inside the charger. DC-DC circuit introduces noise and ripple. Although noise and ripple are not of concern for charging mobile phones, they are very important for lab use.
• The battery needs to be charged. Lithium battery usually charges at 4.2V and it must be a CC/CV charging process. So there will be a charging circuit.
• Using lithium batteries also implies the need of protection circuit to prevent over charging or discharging.

Too much for the theory. Here comes the victim of today’s tear down.

The most useful tool to test battery is a constant current dummy load. Unfortunately I do not have one. So I’ll use a high power LED light as load. The light draws 414mA from USB port. It should be close to the charging current of mobile phones.

Charger “A”) This is one of the most popular chargers in the market today. The label says “Power Bank for iPad/iPhone”, with out brand or manufacture whatsoever. Here in Singapore it costs S$20 online. But in China you can get it as low as RMB29. The weight is 120g and capacity rating is 5000mAH, which does not seems to be overrated.

The first test is to measure its output ripple voltage under load. Just a side note: to measure ripple voltage (usually at several mV to several hundred mV), the ground link of your oscilloscope test lead will function like an antenna that picks up higher EMI noise than the actual ripple. So you must take out the ground alligator clip and use the spring header instead, as the photo below:

As a low cost charger I did not expect too much quality from the charger. The result confirms it: the ripple is as high as 768mV peak-to-peak. Base ripple frequency is about 100kHz but mixed with higher frequency spikes. I don’t think it is anything good for lab use. Even use this to charge mobile phone makes me feel uncomfortable.

So what’s inside this charger? Let’s have a look inside:

As I’ve expected, it contains a lipo battery without marking

Here is the DC-DC boost converter. The chip is marked “PCN TSA2Q”, corresponding part number PAM2421, by “Power Analog Microelectronics Inc”, datasheet here. The rated efficiency is about 87% for 3.6-5V boost at 400mA load. Interestingly my measurement didn’t show the rated 520kHz switching frequency. The datasheet does not mention anything about ripple, seems the manufacture is purposely trying to avoid the topic.

The charging circuit is based on “AP5056” chip. Googling 5056 resolves several Chinese webpage, like this and this. It seems the chip does implement CC/CV charging mode and automatic charging termination.

Chip “DW01” and the two MOSFET 8205A are the lithium battery protection circuit. DW01 datasheet can be found here.

There is a 14-pin chip in the upper side of the PCB. Interestingly it is not a chip with markings being scratched off (like many other Chinese products). This chip is manufactured without marking! My guess is that this is a micro controller in charge of blinking LED lights, as the charger actually has a 3-segment LED to indicate battery level. But why is it without marking? To prevent being copied?

Verdict: This is a typical Chinese ShanZhai(山寨) product. I’m not sure if it is copied from somewhere else or the engineer roll out their own. The parts used in the charger is sub-prime, hence it is low cost. With huge ripple voltage I’ll not take it for lab use. It is potentially good to light some LED lights but use it to charge expensive mobile device is the last thing I’ll do.

Charger “B”) Charger “B” is a typical iPhone charging case. I got it from a friend who “upgraded” his iPhone to a Samsung Glaxy Note. It weights 70 gram and rated 1900mAH, seems realistic for this type of charger. The battery is pretty weak when I received it. My intention is to take out the PCB and see if it can be used for other purposes. So I measure ripple voltage:

Ripple is 212mV peak-to-peak. Compare with Charger “A” this waveform is much clearer. DC-DC switching frequency is about 416kHz.

Taking the charger apart and check the battery, again we found an unlabeled lipo cell

Surrounding the bulky inductor should be DC-DC converter circuit.

The converter chip is marked “AA2C”. But Google cannot find its actual part number. “AA2C” could just be the manufacture’s internal code. Anyway giving its ripple voltage I do not believe it is anything worth investigating.

The charging circuit is based on the chip marked “2YL6”. Again it is some internal code. But this time Google resolves some post at Yahoo groups which leads to this datasheet. The actual part number is BL4045, by Shanghai Belling Co. Ltd. It does CC/CV charging and automatic termination. Everything looks fine here.

The protection is using the same chip “DW01” as we seen in charger “A”. Only difference here is that charger “B” uses one 8205A MOSFET instead of two in Charger “A”. Since 8205A is dual MOSFET, why charger “A” needs two of them?

Charger “B” also has a MCU to blink the LEDs. This time it is a Haier HR6P61P2S4L. Before today I thought Haier only makes washing machines and fridges HR6P61 is a 8-bit, 8Mhz, Harvard architecture micro controller. The word “Harvard” sounds familiar to me … and no surprise Microchip is using the same architecture, and no no surprise Microchip has filed lawsuit against Haier for copying there microcode. I have no authority to give  judgement here, but this kind of thing is not uncommon, shame! (after thoughts: maybe it explains why charger A has the chip blanked?)

Verdict: The build quality of this charger is decent. It functions well as a charger, but to use this as a lab power supply is a bit shaky. Use it for LED lamps perhaps.

Charger “C”) This charger is special. The model is “ML-102”, originates from a design by members in a Chinese forum. I paid RMB 32 from taobao.com. This charger does not comes with battery, you have to supply your own 18650 lithium battery. The advantage of using external battery is that the capacity will never be over rated. How is the ripple? This is the result:

73.6mV! Or 1.5% of full 5V range. Not too bad at all. Ripple frequency about 500kHz with some high frequency spikes.

Here is the DC-DC converter, based on “PT1301”, datasheet here. Rate efficiency is 88% flat at loads from 50-300mA, which is hard to believe. The maximum output current at 5V is 300mA, so do not expect this charger to charge quickly.

Protection circuit is “VGP4” and “8205”. Apparently 8205 is the MOSFET. VGP4 is? Google has no idea, let alone me.

Charging is supported by two TP4057 and one Fairchild MOSFET FDS4953 at 1A charging current. It takes about 5 hours to fully charge my Panasonic NCR18650A 3100mAH cell. On the discharge side, it is enough to fully charge iPhone twice.

Verdict: The design of ML-102 is unique. It can be used as mobile charger or directly charge 18650 cells. However 18650 battery is not available on the consumer market (at least here in Singapore), so the charger is more suitable for the professionals. Overall the charger performs pretty well. May fat hop was that the schematic for this charger is open. But as a Chinese design I would rather not expect too much. As the output ripple voltage is considerably low, this charger is good for lab use to power digital circuits. But for analog circuits I would incorporate a linear LDO in series.

Charger “D”) This is the first mobile charger I bought 3 years ago, long before the market is saturated with cheapo chargers. As it is from the famous “Energizer” brand I’m having good faith in it. And in fact after 3 years of extensive use it still stands strong, better than many laptop batteries.

The charger weights about 175g, rated at 4000mAH, I’d rather think it is under rated. The model number is “XP4001”, manufactured by “TennRich Electronics (Shen Zhen) Co., Ltd.”. At least the manufacture dare to put their name on it!

We check the ripple first. And Wah!  38mV peak-to-peak, and this 38mV is only contributed by spikes.

I was a bit hesitated to pry the case open, but I really want to see what’s inside. And in fact during the process I broke two plastic latches  Not surprisingly this time the battery has a label, although the label gives no information about the actual capacity. There is a “3M” double side tape on top but only used as “single side” tape. It would be harder for me to tear it apart if the top side also sticks to the case.

Here is the DC-DC part. Again a mysterious chip “L3K3”.

Interestingly just beside the DC-DC converter there is another inductor. Tracing the circuit, I realize it is actually an LC Π filter sits in-between the DC-DC converter and output socket. It explains why the ripple is so low in this charger. They really spare no effort to keep the waveform clean.

The XP4001 charging circuit is based on “VA720?” IC, manufactured by “VIMicro”, date code looks like 51 week of 2007. I have looked into VIMicro website but unable to find any information.

The battery protection circuit is similar to all the others, using “312F” chip. No further information available.

Verdict: The XP4001 charger is among the best in terms of life span, output stability and build quality. Although expensive (S$99 three years ago), it really worth every cents. The output ripple is good enough for most lab use, adding some more filtering will make it even better. However this charger has a annoying feature that it disconnects its power if the load is too low. So I have to frequently press the button on the unit to keep it “awake”.

Conclusion:  The electronics market is full of gimmicks. While consumers focuses more on appearance and advertised features, the internal quality of the product is often neglected. In the event that one cannot tear the product apart and check the internals, trusting the brand name seems to be the last insurance. And my advise to those who are constantly looking for bang-for-buck deals: never ever trust a Chinese product without looking inside of it.

The post Mobile charger mini review appeared first on Digital Me.