DIY USB charger for MP3 player

[savidis]

I recently tried (and failed) to make a battery charger for my Monolith MX-7010 MP3 player. The player normally charges via it's USB connection and I wanted a standalone charger for use when I am not near a computer. The MP3 player reports to my computer that it requires 90mA when it is connected so I built a circuit which had a 5V dc output limited to 100mA. I connected the output of my circuit to a USB plug and tested the output voltage and current limit. When I connect the USB cable to the MP3 player it behaves as if it is charging (red LED on) but doesn't charge (battery still flat after many hours) and only 5mA flows (probably only due to the MP3 player LED coming on)

Anyone have any ideas? Maybe the device has to initiate some kind of communication with the host computer to tell it how much current it will be pulling before it will start to work or maybe I need to terminate the USB data connection pins with a pull-up or pull-down resistor? Has anyone ever tried charging such a device using a powered USB hub which is not connected to a computer? Thanks for any help.

Alex

[Set Square]

It probably needs to 'talk' to the computer before switching itself into charge mode. It could be quite difficult to simulate the dialogue, using a dumb device!

Otherwise, if it's just logic signals, you need to hold the other lines in the right state.

Do you have a powered hub? If so, and if it charges when the hub is connected to a computer but not when it isn't, investigate what happens to each line when you plug and unplug it.

[savidis]

Thanks for your help. I don't have a powered hub to check with but I have now made some progress: There isn't any data communication required. I confirmed this by cutting the data wires in a spare USB cable and checking that the unit charges properly from a PC. I also found that when charging it draws around 130mA (not the 90mA that it reports to the host computer). I upped the current limit in my home made regulator circuit but this also failed to work. The problem seems to be that the output voltage of my supposedly regulated charger circuit was dropping from 5.0Vdc to 4.5Vdc with the MP3 player connected even though only a few milliamps are being drawn. The circuit works fine into any resistive test load but the MP3 player seems to upset it. I can only guess that some internal switching regulation is taking place in the MP3 player itself and this is upsetting the stability of my voltage regulator and current limiting circuitry. I'll redesign the circuit for improved stability and give it another try.

[Chris Hodges]

For a test you might try running off 4 AA cells with a diode in series for ~5.3V. Batteries can't get confused and if there's an internal regulator it won't notice the extra .3V. (I've just measured my USB voltage to 5.1).

[Christian McArdle]

Why all the current limiting circuitry? Personally, I'd use a fuse! If not, set it to half an amp or something. The required current draw might be average, rather than peak.

Christian.

[savidis]

Thanks for the good ideas.

Last night I rebuilt the circuit using just a LM317 regulator set to output 5.0Vdc without current limiting cicuitry and everything appeared to work OK. The MP3 player drew around 150mA when "charging". I noticed after about 15 minutes the current was about the same but the output voltage had fallen to 4.5V or so. I find this surprising since the regulator should be able to maintain the output upto at least half an amp (it is mounted on a decent sized heatsink). The supply input to the regulator circuit is an unregulated 12Vdc @ 500mA max wall-wart type thing so should be quite suitable. I left it "charging" like this for a couple of hours but when I tested the MP3 player I found that the battery had actually discharged completely! Could it be that the output voltage dropped so low that the regulator started to sink current from the battery? I think the problem may be a defective LM317 since I used it in both versions of the circuit (and it was pretty much the only common component). For rev 3 I plan to use a new regulator and install an output diode to prevent any possible current leakage back to the charger, but it might have to wait until next week before I get the chance...

[Dave Plowman (News)]

You don't need constant voltage charging for a Ni-Cad - that's for lead acid. You need constant current. The actual voltage doesn't matter provided it is adequate - within reason.

I could send you a very simple cc circuit for a 14 hour charge.

[Ian Stirling]

You do have capacitors on the input/output>?

[Stefek Zaba]

I'm sure you could (one transistor and a couple of resistors, right?). But once you're into building, I'd find it hard not to reach for a Maxim charger IC, which will do the Right Thing for a Ni-Cad (fast charge, detection of voltage levelling off/dipping to stop the fast charge, and all the rest of the NiCad voodoo) - one-off prices around a fiver, if that...

Stefek

[Ian Stirling]

It's not a battery. It's a USB device, that needs powered, that has it's own internal charger.

[Dave Plowman (News)]

Depends on whether it includes LED indictors. But the basic circuit is one transistor, two resistors and two diodes. And if the DC supply is near enough the ideal voltage you'd probably not even need a heatsink for the transistor.

[Christian McArdle]

But the MP3 device will internally manage this. I don't think the USB spec is to provide a constant current power supply.

Christian.

[Christian McArdle]

Just bung in a 7805. Very simple. Works every time.

Christian.

[Ian Stirling]

Pity it's not a bit more flexible.

The option to have a higher voltage (12/30V?) would be nice. Only for devices that request it of course. So that (for example) instead of 2.5W, 15W would be available, while keeping within the current limits of the connector.

[savidis]

Thanks for all the replies. Problem was indeed a defective LM317. I found that the peak current demand could be in excess of 500mA (so strictly speaking it actually exceeds the limits of the USB electrical specification). Whilst searching for a new voltage regulator IC in my garage I came across a compact DIN rail 5V 1.5A switch mode power supply. This works perfectly so I've ditched all the nasty bits of Veroboard.