E-bike kludge opinion?

[conundrum]

Hi guys and gals.

I acquired 12 "new" phone charger Li-Ion packs, containing a 5.2Ah LiPoly cell, charge state LEDs and a push button USB charger.

The problem is that lacking this many discrete chargers, connecting them in series to charge is a BAD idea, because there is no isolation built in.
So my plan is to make a small SMPS which generates 12 isolated 5.2V supplies limited to 2A, that can then power the existing charger circuitry and overcome any cell imbalance allowing any old 12-36V >3A power supply to be used for charging.

The output side is simple as the cell(s) have internal failsafes to prevent overdischarge.
In addition I plan to add two optocouplers to each board, to connect the enable switches and monitor the LEDs for a dash mounted display of charge states.
Retaining the discrete nature of the packs means that replacing a bad/weak cell should be easy if connector pins are used on each side.
Sound feasible?

[Alex555]

As far as charging batteries I do not know very much, however you may be able to use one lm317 current limiter circuit per battery, all attached to a 5.2 volt supply that has enough amperage to supply two amps per circuit continuously. One regulator per circuit is enough for charging at one amp, you will need two in parallel for two amps.

[Alex555]

Here is a schematic. You will have to adjust the resistor to get the desired current, and make sure it has a sufficient wattage. It is also worth noting that no matter the output voltage, your supply will always have to supply the same amperage as the output. So your cell can draw 1a at 2v and your 5.2v supply will still have a 1 amp load. The lm317's also have a thermal overload shutoff so they will not burn out. I believe it may also be possible to emit the resistor because they have a 1 or 1.5 amp current limit built in, but you would have to check the datasheet. if the datasheet says it has over current protection just make the resistor 0 ohms. Multiple circuits will let you charge several batteries at the same time on the same supply.

lm317.gif (7.45 KiB) Viewed 1093 times

[Alex555]

i forgot to mention the 1.7v drop across the regulator.

[conundrum]

That might work for one battery..

Just a thought, but isn't the LM317 going to dissipate a lot of power if it is dropping 29 volts at 2A?

This is why I was planning to use a common regulator with lots of outputs, and transistor switches on each to isolate the batteries once charged.

[Alex555]

Ye,s it would dissipate a lot of power with a twenty-nine volt drop. For this design you would need a low voltage high current supply with a bit of heat sinking.

[conundrum]

Possible idea here would be to use an MC34063 for each stage, and a 1:1 isolation transformer.
This would get around the problem of variable current, and ensure each battery gets an equal charge.

Seems an excessive waste of chips though, I haven't exactly got an unlimited supply of them.

[Alex555]

The idea of the lm317 is that they are cheap. And wouldn't a 1:1 transformer only regulate voltage, and give different currents for batteries that have different charge levels? Another option would be to use an ATX supply, which can supply many amps at 5 volts. To compensate for the voltage drop, you could add a power transistor on the output. this way you would only need an lm317, a power transistor, and 1 or two resistors per circuit. A nice safety feature and convenience would be a comparator on each charger to isolate the battery. Also, what voltage are the batteries, and what voltage do the commercial chargers charge them with?

[Alex555]

I just realized you were trying to power charging circuitry, not charge the batteries themselves. Why can the chargers not be connected in parallel if they are separate units?

[conundrum]

I don't think this will work because there is no isolation between the grounds.

The chargers have a 5 pin socket and an onboard controller which charges to 4.15V and then stops, all the lights then shut off.

-A

[Alex555]

I think I am missing something. Are they separate units? If so, why do they need isolated ground?

[matseng]

Don't even think about charging lipo or li-ion batteries with a LM317 or something similar if you care about your property and/or life.

Use a real charging ic like a mcp73811 http://www.farnell.com/datasheets/77596.pdf , they cost like 50 cents and don't require any external components except for a capacitor or two (just like the lm317) and they charge the battery in a proper way.

[Alex555]

All that he needed was isolated supplies to power proper charging circuitry. I don't know anything about charging batteries, he just said he needed a constant current supply and that's what I gave him. Although I am still not sure why the chargers need to be isolated? would a picture be possible?

[matseng]

If you have three non-isolated chargers connected when the batteries are in series they will be shorted by the ground wires from the chargers. (Several other problems are also occurring but this one is enough)

Each charger (yellow) circuit connected to power and ground and then the output of the charger unit is connected to the poles of its battery. As can be seen in the (very professional and nice) drawing the middle battery is short circuited by the ground wires (blue). So the power supplies for each battery charger needs to be completely isolated.

Charging three batteries in series non-isolated=BAD

[Alex555]

Now I see the problem, that picture helped a lot. You could use a relay to switch from series to parallel while charging, which would not consume battery power if done right. A 1:1 transformer would work, so would individually switching power and ground to capacitors at a very high frequency. I have had success with capacitor isolation, and it might be cheaper than transformers (I have no idea how much they cost, I never use them). a relay would probably be your best bet as far as cost, size, and convenience.