Sick of Apple chargers? Solution: SupperCharger+, the Next Generation

[SharpEars]

This was from a previous design, without the meters.
I must say it is much better with the meters.

Nice! Really nice! Is that all just power supply circuity or is there other circuitry mixed in?

Oh, I see... You separated the tan power supply board from the computer and added the daughter card.

So, what's the current limit on this thing? If I connect four iPads to it, will it put out 8+ Amps?

[JacquesBBB]

Oh, I see... You separated the tan power supply board from the computer and added the daughter card.

So, what's the current limit on this thing? If I connect four iPads to it, will it put out 8+ Amps?

I have not push it to its limits,  but I believe its about 5A. I had never any limitations with the 4 outputs,
and I believe there is an internal protection as well,  but I  did not experimented with it.

[JacquesBBB]

Oh, I see... You separated the tan power supply board from the computer and added the daughter card.

Quite nicely, I retrieved the daughter board connector and the USB plugs from the main board.

[SharpEars]

The 2.7  were just obtained with voltage dividers with 2 resistors.

I've decided to use LM317 for the fixed USB data voltages (so I am going to mount at least two of them for 2.3 V and 2.7 V, respectively, allowing for compatibility with all Apple and many Android devices). I can use the output of the 2.7 V going to both D+ and D- for Apple devices. I can use a mixture of 2.3V and 2.7V for Android devices, depending on the charging level they support. I will mount dip switches to control the voltage that goes each of the four (because there are two USB ports) Data pins.

[SharpEars]

Oh, I see... You separated the tan power supply board from the computer and added the daughter card.

So, what's the current limit on this thing? If I connect four iPads to it, will it put out 8+ Amps?

I have not push it to its limits,  but I believe its about 5A. I had never any limitations with the 4 outputs,
and I believe there is an internal protection as well,  but I  did not experimented with it.

I am surprised that it has such good load regulation, because the PS circuit looks very, well, simplistic. Do you (or anyone else on the forum) know if it's a linear or switching circuit? I can't quite tell just by glancing at its pic from the top.

[JacquesBBB]

Its a switching power supply.

The file attached is an attempt to reverse enginer  the regulator part of the PSU.
I could then modify the voltage by changing R49.
As initially it was a little bit below 5V.  I shorted it to put it at 5.2V, but now, I am more thinking of
putting the voltage to 5.1 V . Maybe I should  put some switch or pot  to leave the possibility of
slight changes in the voltage.

[SharpEars]

Its a switching power supply.

The file attached is an attempt to reverse enginer  the regulator part of the PSU.
I could then modify the voltage by changing R49.
As initially it was a little bit below 5V.  I shorted it to put it at 5.2V, but now, I am more thinking of
putting the voltage to 5.1 V . Maybe I should  put some switch or pot  to leave the possibility of
slight changes in the voltage.

If there's anything I decided on up front it's to allow for voltage trim. So, I would definitely agree with your premise of adding a pot to be able to adjust the voltage. The downside is that I will have to use voltage regulators for the Data pin voltages instead of resistor ladders.

[JacquesBBB]

In fact, it is a samsung AD-3005 PSU  30W (6A).

The switching  component is a TOP247Y
https://www.westfloridacomponents.com/mm5/graphics/H02/TOP250YN-Power-Integrations.pdf

Compared to  an apple PSU,  it is much less cramped in a small place; Of course, it  is not portable,
but  the heat dissipation is much more efficient.
For my desk use, it is perfect. The improvement that could be done would be a way to shut it down completely in case of no load.
This would be some next project.
 

[SharpEars]

In fact, it is a samsung AD-3005 PSU  30W (6A).

The switching  component is a TOP247Y
https://www.westfloridacomponents.com/mm5/graphics/H02/TOP250YN-Power-Integrations.pdf

Compared to  an apple PSU,  it is much less cramped in a small place; Of course, it  is not portable,
but  the heat dissipation is much more efficient.
For my desk use, it is perfect. The improvement that could be done would be a way to shut it down completely in case of no load.
This would be some next project.

A guy in the UK is selling it used for $46.75 on eBay at the moment with unspecified shipping charges. I spent less than $25 on my power supply (all in including shipping, which accounted for almost a third of the price), which I will reveal this weekend. For now, I will say that it is also capable of sourcing at least 6A and that it is not one of these pieces of s..t being sold on eBay:


One shortcoming of that Samsung supply is that it lacks voltage sense capability and a voltage trim (without mods). A tried to find a spec sheet PDF for it, but I couldn't find it.

[JacquesBBB]

I dont say that  this is the best PSU, or that I would  buy it  at the price it is on the net.
I got a bunch of them for free, before they went to the trash.
I realized that they were very decent PSU for USB power, and found a way to reuse the original
case without too much work.

I could not find the schematics, but I am now getting familiar with them, and can consider making some
modifications,  although I still need to finish  one or two more in order to fill my needs.

[SharpEars]

I dont say that  this is the best PSU, or that I would  buy it  at the price it is on the net.
I got a bunch of them for free, before they went to the trash.
I realized that they were very decent PSU for USB power, and found a way to reuse the original
case without too much work.

I could not find the schematics, but I am now getting familiar with them, and can consider making some
modifications,  although I still need to finish  one or two more in order to fill my needs.

I do agree, you can't beat free.

[Monkeh]

I dont say that  this is the best PSU, or that I would  buy it  at the price it is on the net.
I got a bunch of them for free, before they went to the trash.
I realized that they were very decent PSU for USB power, and found a way to reuse the original
case without too much work.

I could not find the schematics, but I am now getting familiar with them, and can consider making some
modifications,  although I still need to finish  one or two more in order to fill my needs.

I do agree, you can't beat free.

You got a free charger with all your devices. Why do you keep trying to beat it?

[JacquesBBB]

You got a free charger with all your devices. Why do you keep trying to beat it?

Very simple,
1) I do not want to move the chargers with the devices.
2) with a 4 ports  USB PSU, I need only one plug.
3)  I had several of the original chargers who failed after some time.
4) Some chargers are 0.5 A, other 1 A others (not so many)  2 A,  so  they are not interchangeable,
while you can plug anything in a more powerful PSU, it will always charge at maximum possible speed.
5) Some devices do not come with chargers, as the power bank which I bought on ebay.

[amyk]

You're assuming that with a voltage of 5V at the charger, the phone is actually limited by that CC current. I dare say that's a completely unfounded assumption (and seemingly disproved by the OP's assertion that the current does increase) -- phones compliant with standard X (I forget the name) will look for voltage sag from the charger as an indication that the charger is at its limit, and will be current limited by the charger instead, i.e. charging current = min(LiPo CC, charger capability).

LiPo batteries have very low internal resistance, hence why they need current-limited charging. Any difference between Vbat and Vbus will be dissipated as heat in a linear charger. Assuming a battery that stays around 3.8-4V during most of its charge cycle, and LDO-type current regulator with a dropout voltage of ~1V, then a 5V input would be enough.

Even if the adapter is supplying 5V, the cable and contact resistance reduces this, which is why they're adjusted to 5.3V no load - the extra 0.3V is to compensate for resistive losses in the path so ~5V ultimately ends up at the current regulator. In light of this, you'll probably be charging as fast as possible if you just use 5V with remote sense at the Lightning connector's Vcc.

(Of course this is assuming they don't do something funky like increase the CC based on input voltage, which being Apple, wouldn't surprise me. But based on the numbers posted so far, it's just plain resistive losses you're fighting against.)

FWIW the USB forum recently (August 2014) increased the maximum official Vbus to 5.5V, likely because most if not all existing devices were fine with 5.5V anyway and this gives a wider 10% tolerance.

[rs20]

LiPo batteries have very low internal resistance, hence why they need current-limited charging. Any difference between Vbat and Vbus will be dissipated as heat in a linear charger. Assuming a battery that stays around 3.8-4V during most of its charge cycle, and LDO-type current regulator with a dropout voltage of ~1V, then a 5V input would be enough.

Even if the adapter is supplying 5V, the cable and contact resistance reduces this, which is why they're adjusted to 5.3V no load - the extra 0.3V is to compensate for resistive losses in the path so ~5V ultimately ends up at the current regulator. In light of this, you'll probably be charging as fast as possible if you just use 5V with remote sense at the Lightning connector's Vcc.

(Of course this is assuming they don't do something funky like increase the CC based on input voltage, which being Apple, wouldn't surprise me. But based on the numbers posted so far, it's just plain resistive losses you're fighting against.)

FWIW the USB forum recently (August 2014) increased the maximum official Vbus to 5.5V, likely because most if not all existing devices were fine with 5.5V anyway and this gives a wider 10% tolerance.

Agreed on all points -- but the end conclusion is that a charger that can hold up 5.3V at 2A is faster charging that sags to 5V at 2A, right? Because the latter would starting eating into the voltage budget at the phone once you factor in cable losses. Regardless, JacquesBBB's figures demonstrate this; the current going into the phone increases rapidly up to 5.1V, before tailing off at 5.2V. So charger delivering, say, 5.15V + loop losses for 1.5A, at 1.5A is ideal there, and it seems bold to claim that any randomly chosen charger will achieve this.

To clarify my point, I of course agree that going arbitrarily high on voltage is obviously pointless, at some point you just start burning up more in the linear reg. What's not so clear is whether your typical phone plugged into your typical charger is actually operating in that regime.

[JacquesBBB]

I have measured the output of my genuine 10 W  Apple Ipad charger 

I found
   
         0 A           1.6A

V+ : 5.15 V       5.02 V
D+ : 2.76 V       2.70 V
D-  : 2.06 V       2.02 V

So the difference with my PSU is not the voltage at no load, but the fact that  it is much better regulated.

[Richard Crowley]

So the difference with my PSU is not the voltage at no load, but the fact that  it is much better regulated.

Please remind us why that is important?
Considering that the built-in charging circuit does the ACTUAL CURRENT REGULATION and cares nothing about the incoming voltage.

[rs20]

So the difference with my PSU is not the voltage at no load, but the fact that  it is much better regulated.

Please remind us why that is important?
Considering that the built-in charging circuit does the ACTUAL CURRENT REGULATION and cares nothing about the incoming voltage.

A fact that is proven false by the numbers posted by JacquesBBB a few pages ago (current rising as voltage rises). Evidently the in-phone regulator is not getting sufficient voltage if the supply voltage sags much near/below 5V.

[Siwastaja]

It makes a lot of sense to limit charging current if the supply voltage sags below 5V, so charging current control based on input voltage has long been reality in laptops, and why not phones, too. It's not very surprising that this principle extends to a bit over 5.0V, too.

It is true that higher current charging does more damage to the battery (via lithium plating), but it's not a simple equation at all. And I understand that some people might want quicker charging, even if they needed to replace the battery more often. But don't try to do it when the battery is cold. Cooling the battery too much may be detrimental. Optimum cell temperature for quick charging might well be around 40-45 deg C.

Quick charge damage can happen in a surprising way; the cell may work fine for 50 cycles, then degrade quickly over the next 20.

Most of the quick-charging damage is done at high state of charge (over 4.0 volts), so if you only charge your phone to approx. 70%, probably no extra damage is done, even if the current was 20-30% higher than normal. Very quick charging requires an additional current curve in addition to CC-CV, as too much damage is done near the CV voltage at high rates. For example, a cell can be fine up to 1.5C charge until 3.9V, then linearly derating to 0.5C until 4.2V, then CV.

[SharpEars]

I dont say that  this is the best PSU, or that I would  buy it  at the price it is on the net.
I got a bunch of them for free, before they went to the trash.
I realized that they were very decent PSU for USB power, and found a way to reuse the original
case without too much work.

I could not find the schematics, but I am now getting familiar with them, and can consider making some
modifications,  although I still need to finish  one or two more in order to fill my needs.

I do agree, you can't beat free.

You got a free charger with all your devices. Why do you keep trying to beat it?

Because you get a 1 A charger with phones (even though some can charge at 2 A) and a 2 A charger with tablets. Apple tries to sell you a so called 2.4 A charger for $19 (even though in reality it charges iPads at only 1.8 A). I want something that charges at MAX A, within the device's safety parameters. This MAX A will always be more than any Apple (or other) charger will supply, because it will be caused by charging at the highest reasonable (and safe) voltage before the device OVP kicks in.

[SharpEars]

You're assuming that with a voltage of 5V at the charger, the phone is actually limited by that CC current. I dare say that's a completely unfounded assumption (and seemingly disproved by the OP's assertion that the current does increase) -- phones compliant with standard X (I forget the name) will look for voltage sag from the charger as an indication that the charger is at its limit, and will be current limited by the charger instead, i.e. charging current = min(LiPo CC, charger capability).

LiPo batteries have very low internal resistance, hence why they need current-limited charging. Any difference between Vbat and Vbus will be dissipated as heat in a linear charger. Assuming a battery that stays around 3.8-4V during most of its charge cycle, and LDO-type current regulator with a dropout voltage of ~1V, then a 5V input would be enough.

Even if the adapter is supplying 5V, the cable and contact resistance reduces this, which is why they're adjusted to 5.3V no load - the extra 0.3V is to compensate for resistive losses in the path so ~5V ultimately ends up at the current regulator. In light of this, you'll probably be charging as fast as possible if you just use 5V with remote sense at the Lightning connector's Vcc.

It is nearly impossible to build a device that can remote sense at the Lightning connector's V_cc. This is one of the challenges I will be facing.

(Of course this is assuming they don't do something funky like increase the CC based on input voltage, which being Apple, wouldn't surprise me. But based on the numbers posted so far, it's just plain resistive losses you're fighting against.)

But, my claim is that this is exactly what they are doing, increasing current as the voltage increases up to the OVP.

FWIW the USB forum recently (August 2014) increased the maximum official Vbus to 5.5V, likely because most if not all existing devices were fine with 5.5V anyway and this gives a wider 10% tolerance.

Awesome, so if I can get 5.5V into the device, I would charge it even faster and no wall wart (5 V) charger puts out 5.5 V.

[SharpEars]

I have measured the output of my genuine 10 W  Apple Ipad charger 

I found
   
         0 A           1.6A

V+ : 5.15 V       5.02 V
D+ : 2.76 V       2.70 V
D-  : 2.06 V       2.02 V

So the difference with my PSU is not the voltage at no load, but the fact that  it is much better regulated.

The 12 W charger does not do much better, maxing out at (IIRC) 1.8 A at (from memory):

        1.8A

V+:  5.06V - 5.08V (at load)
D+:  2.75V (Note: 2.75 D+/2.75 D- is the 12 W charger indicator config)
D- :  2.75V

[SharpEars]

So the difference with my PSU is not the voltage at no load, but the fact that  it is much better regulated.

Please remind us why that is important?
Considering that the built-in charging circuit does the ACTUAL CURRENT REGULATION and cares nothing about the incoming voltage.

So you mean to say that the built-in charging circuit, rather than intelligently reducing the current load as voltage increases to maintain a constant output voltage (and current), burns it off instead (e.g., like a Zener)?

[eas]

I got tired of reading the idle speculation, so apologies if I missed this, but I went digging through teardowns and checking part numbers last year because I wondered about Apple's choice of charge controller. From that research, I learned that it looks very much like Apple iPhone and iPads made in the last 3+ years use switch-mode chargers.

My own idle speculation is that they use switch-mode charge regulators in service of speeding charging time. With switch mode regulators, they can devote more of the thermal budget to heating due to battery internal resistance, giving them more headroom to use higher charging currents.

I also speculate, based on anecdotal observation of charging currents on my iDevices, that Apple is not using the standard two-phase CC/CV charging algorithm. I haven't collected datapoint through a whole charging cycle though.

The thing about Apple products, which inspires both their admirers and haters, is that they are designed as full system that actually extends across multiple devices over the lifetime of the user. From this perspective, one conclusion I reach is that there isn't going to be a lot of headroom for you to improve charging speed over what you see with Apple's cable's and power supplies, not without tweaking firmware and potentially internal components.

That said, I'm interested to see how you do. I bet there is 10-15% in there somewhere, though probably the cheapest way to get there is with a fat power leads to a short lightning cable, or if you want something a little more fun and elegant, sense leads the splice points on a short lightning cable.

A few more, thoughts:

• "Clean" power matters to the extent that too much noise can wreak havoc with capacitive touch sensors if you are using the device while charging.
• Some quick-chargers, like the multi-port Anker I have, step up the delivered voltage as current rises. I assume this is done to compensate for resistive losses in the USB cable at higher currents. I think this may actually be a features of some USB fast-charge ICs from TI or similar.
• I think that with the iPhone 6, and the wall-warts that shipped with it, Apple may have moved away from their proprietary charger sensing to just taking as much juice as they needed or until voltage drops too much.

[Siwastaja]

Awesome, so if I can get 5.5V into the device, I would charge it even faster and no wall wart (5 V) charger puts out 5.5 V.

That may be a wrong conclusion, however - nothing physical nor any sane design principle dictates that kind of behavior.

You have experimentally found that the charge controllers you tested do indeed increase charging current with 5.3V supply compared to 5.0V supply, but the expectation that it would go on still beyond 5.3V is unfounded. Going too far, however, will increase heat losses in devices that utilize linear charge controllers, and at some point, they may start thermal limiting. There will be a sweet spot, and you may have found it at 5.3V. In any case, it will be highly device-dependent and the only way to know is to actually test as you have  done.

Do the modern iDevices with high charging currents really use linear controllers? Anyone? If you think about charging an empty cell (3.5V) at 2A from a 5.0V supply, it would be 3W loss in charge controller only - the heat sinking would make it bigger than a comparable buck converter. If it's really a DC/DC, then the actual charging current increases even more than the OP has measured. Edit: eas has confirmed (above) the use of a switch mode controller, as you would expect.