Polarity agnostic power supply

[Boogami]

Hello,
I want to build a charging dock using an existing 5V USB wall charger. The charging dock will have two pads on which the device that needs to be charged sits with the other two pads. How can I make the charger deliver power no matter the position of the device on those pads? I've used a full bridge rectifier but the voltage drop is too big. Do you know any other circuit that can do the polarity inversion without a voltage drop? I have seen a solution with MOSFETS but I dont know how to choose them for this application. 
Thank you.

[LapTop006]

Either use a boost convertor, or perhaps switch from contacts to a Qi charger module (which will probably have such a boost convertor on the output).

[SiliconWizard]

First thought is that the solution would have to be integrated in the device to be charged and not in the charger itself, unless it has some means of detecting how the device is plugged? Is that what you had in mind?

[ArthurDent]

We could use some more information on the device and charger and how much or little you want to modify them but IsoTip uses a mechanical feature so the device to be charged can only go in one way. Can you design the charger so the device can only contact the pins if it is inserted the correct way?

[james_s]

Schottkey diode bridge may have a low enough drop. Synchronous rectifier with low resistance mosfets is an option too.

[capt bullshot]

One MOSFET solution is here:
https://www.analog.com/media/en/reference-design-documentation/design-notes/dn519f.pdf
https://www.analog.com/media/en/technical-documentation/data-sheets/4320fb.pdf
To be integrated into the charged device.

If you want to automatically reverse the charge pod polarity depending on how the device is put into, you'd want some kind of H-Brigde controlled by some kind of "try and error" algorithm, cautiously avoiding damage to the device.

[Boogami]

We could use some more information on the device and charger and how much or little you want to modify them but IsoTip uses a mechanical feature so the device to be charged can only go in one way. Can you design the charger so the device can only contact the pins if it is inserted the correct way?

Thank you all for the answers. I am trying to charge a bluetooth speaker like this and I want to put a sticker behind the speaker so I can charge it with the charging dock. I don't really want to modify the speaker, only the charging dock so it can detect if there is something on it and charge it regardless of its position.

[SiliconWizard]

I don't really want to modify the speaker, only the charging dock so it can detect if there is something on it and charge it regardless of its position.

Then, as I hinted, I still don't know how a full bridge on the charging dock would help if you're not modifying the contacts. It would only work if you put the bridge inside the speaker I guess?

[james_s]

In order for the charging dock to automatically set the polarity, it has to have a way of identifying which way the load is plugged in. I don't see how this is going to be accomplished without modifying the speaker.

[NiHaoMike]

How does the stock charging dock solve that problem? I would expect that there's a mechanical feature (might be subtle) to only allow it to connect one way or the internal circuit is reverse polarity protected and the dock simply alternates polarity until it detects a current draw.

[mikerj]

How does the stock charging dock solve that problem? I would expect that there's a mechanical feature (might be subtle) to only allow it to connect one way or the internal circuit is reverse polarity protected and the dock simply alternates polarity until it detects a current draw.

In the absence of any other information I was assuming this was a DIY add-on rather than a replacement for the existing dock i.e. the pads would be connected to the existing micro USB charging port on the speaker.  If so then the only chance is to do something like you suggest, and if no reverse polarity protection then it may be possible to use a very low current to determine polarity before switching on the full charging voltage.

[james_s]

If it's possible to put a schottky diode across the pads on the device then you could feed a small current and check for the relatively predictable voltage drop of the diode indicating that the polarity is backwards and reverse it. Of course there is also the issue of the device getting moved around on the pad once charging has already started, which could cause a sudden polarity reversal. The shunt diode may provide adequate protection though and you could detect the sudden voltage drop and reverse the polarity.

[NiHaoMike]

How about a Qi upgrade kit for smartphones? There would be a small efficiency loss but it would eliminate contacts that could become dirty.

[IconicPCB]

A quad MOSFET synchronous rectifier will replace the diode bridge with losses governed by drain source on resistance.

[Boogami]

How does the stock charging dock solve that problem? I would expect that there's a mechanical feature (might be subtle) to only allow it to connect one way or the internal circuit is reverse polarity protected and the dock simply alternates polarity until it detects a current draw.

In the absence of any other information I was assuming this was a DIY add-on rather than a replacement for the existing dock i.e. the pads would be connected to the existing micro USB charging port on the speaker.  If so then the only chance is to do something like you suggest, and if no reverse polarity protection then it may be possible to use a very low current to determine polarity before switching on the full charging voltage.

Yes this is a DIY add-on rather a replacement for an existing dock.  I've attached a diagram, I cannot take a picture with the setup at the moment but this is basically what I have right now. In block 2 I have the bridge rectifier but I would like to have the circuit in block 1, if it's not possible than ok. I am thinking of a flexible circuit that can attach like those QI charging stickers placed on the back of the phone. 

[SiliconWizard]

OK, I thought the speaker directly had two pads like the charger? Wasn't clear, now it is. The speaker has an USB connector for charging, then your solution shown on the diagram will work. Yes, you can use a simple diode bridge - use Schottky diodes and make sure they can handle the charging current. You'll get a voltage drop, then whether it's still within specs for the charging circuit inside the speaker - you'd have to test that.

There are ideal full bridge controllers, such as the LT4320, but they usually require 4 external mosfets, besides, this one isn't for voltages under 9V, so that wouldn't work here. I haven't looked up thoroughly, so if someone has found something that can work at 5V and ideally without the need for external transistors, I'd be interested in seeing that myself. You could also use so-called "ideal diodes" (eg: http://www.ti.com/product/LM66100 ), but you'd need 4 of them? No big deal but unless you know for sure you can't charge your device with a full bridge made of Schottky diodes, I'm not sure this is worth it.

As mentioned above, a QI adapter would also be a solution and more elegant. Then I wouldn't bother keeping the charging dock. Just buy a QI-USB adapter (you can find them for a few bucks, look up "qi usb adapter") and a QI charging dock (can be found around $15, maybe less).

[NiHaoMike]

The losses in a diode bridge at such a low voltage would be comparable to the losses in a Qi charger (maybe even more), in which case you might as well just use Qi.