Step down constant current for LEDs and LiIon 18650 as source

[BloodySword]

Hi folks,

according to , I have some very beginner questions.
(how can I ensure that the video is not embedded? :/)

I'm searching for a step down converter which also is a constant current source for 3 Cree XM-L U2.

- Input should be 4 LiIon 18650 Cells in series, which will be charged individually (a complex switch will ensure this).
- Output current should be 620 mA (I just want to increase the lifespan of the LEDs, nominal current is 700 mA)
- The circuit should take care of the LiIon cells to not deep discharge
- The circuit should also take care of the LED's temperature and throttle the current or shut them off (with "temp. too high" LED indicator)
- The circuit should of course monitor its own temperature and also throttle or shot uff (with "temp. too high" circuit indicator)
- On normal temperature, it should remain on full brightness till the end and go off with a red indicator LED "Charge me!"

Goal is a small PCB with surface mount devices that fits in my stainless steel pipe.
- Inner diameter is of the pipe: 2.5 cm
- Length of the pipe: 36cm
- Could also be 2 PCBs stacked together

For charging, I have bought 4 of these circuits, that can fit into the pipe:
http://www.ebay.com/itm/New-1PCS-5V-Mini-USB-1A-Lithium-Battery-Charging-Board-Charger-Module-EP98-/121089857674

A switch will connect the cells in series when in ON state, and in OFF state it will connect the cells individually to the charger circuits that will be stacked. LED indicators goes with acrylic plastic rods to the side of the pipe, so that the charging status of each cell is visible.

What I need to know:

- What should I know to build such a circuit?
- How to create a PCB surface mount modell with free software?
- How to PLAN a compact PCB even when I have tested a circuit on a breadboard?
- Thermal issues?
- Which type of connection should I use to connect any 5 Volts, 4 Amps power supply for charging?
- I have a beautiful big 7.5 Volts/ 4 Amps power supply in europe plug style! Is it suitable to be used with a step down converter to 5V / 4 Amps?
- The charging current of each charger is not constant till the cell is full. The moment where 1 Amp is pulled is a very short time span and I don't think that the cells would be so empty that 1 Amp ever will flow for each cell.

Edit:
Forgotten to mention which cells I will use:
http://lygte-info.dk/review/batteries2012/LG%2018650%20D1%203000mAh%20%28Pink%29%20UK.html
So no china crap.
They will be charged till 4.2 volts, so 2.5 Ah are guaranteed I think.

Edit 2:
Found better ones:
http://lygte-info.dk/review/batteries2012/Panasonic%20NCR18650B%20Protected%203400mAh%20%28Green%29%20UK.html
Costs me 50 EUR for 4.
I have to decide....

Thank you in advance!

[BloodySword]

Don't know if a double post after a relatively long period of time is allowed, here. Sorry for that.

But I have a small Update:

Just bought these cells: http://lygte-info.dk/review/batteries2012/Panasonic%20NCR18650B%203400mAh%20%28Green%29%20UK.html
Siginificant is that the cut-off voltage should be 3.00 Volts to protect the cells against overdischarging.
So in summary the circuit should cut-off at 12 Volts remaining.

The officital datasheet tells 2.75 Volts. But since we have 4 cells in series, nobody knows which cell is at this limit when discharged to 12 Volts in summary. Perhaps I should even go to 3.1 Volts. So in summary the cut-off voltage should be at least 12.4 volts to protect the cells.

This is the disatvantage with unprotected cells. But unprotected, the self-discharge is much lower and the voltages are slightly better.

[peter.mitchell]

I doubt you will have any issues with those batteries and at your current levels; they are more prone to go "out of balance" at higher currents, but you're well within your limits, 12V is safe as houses. http://en.wiktionary.org/wiki/safe_as_houses

An idea for a cut off circuit is simple, tl431 with some resistors to make it a 3.0v~ voltage reference, high~ value resistive divider, make it a 3x divider, put output of tl431 + output of divider into a comparator, make comparator drive a mosfet that switches your stuff off, to prevent oscillation (due to loaded/unloaded battery voltage), use a p fet for high side switching, use your resistive divider from the switched side of the fet so that when the fet turns off, it stays off (because the voltage to the divider will be 0), and use a small momentary pushbutton with a resistor to pull the gate on when you want to start it.

Protected cells cell protection is "worst case" protection, it isn't designed to be your primary cutoff (generally), it's just there to save your behind incase your circuit doesn't cut off or your parasitics pull the battery low, or if your charger has a bit of a runaway and pushes the voltage right up, the main disadvantage of protected vs unprotected is they are longer, so they don't fit into some slots on devices that use unprotected ones.

[BloodySword]

I doubt you will have any issues with those batteries and at your current levels; they are more prone to go "out of balance" at higher currents, but you're well within your limits, 12V is safe as houses. http://en.wiktionary.org/wiki/safe_as_houses

Nice quote.
Looks like I thought in the right direction.
Thank you.

An idea for a cut off circuit is simple, tl431 with some resistors to make it a 3.0v~ voltage reference, high~ value resistive divider, make it a 3x divider, put output of tl431 + output of divider into a comparator, make comparator drive a mosfet that switches your stuff off, to prevent oscillation (due to loaded/unloaded battery voltage), use a p fet for high side switching, use your resistive divider from the switched side of the fet so that when the fet turns off, it stays off (because the voltage to the divider will be 0), and use a small momentary pushbutton with a resistor to pull the gate on when you want to start it.

Uff... No idea how this would look like as a schematic.

Protected cells cell protection is "worst case" protection, it isn't designed to be your primary cutoff (generally), it's just there to save your behind incase your circuit doesn't cut off or your parasitics pull the battery low, or if your charger has a bit of a runaway and pushes the voltage right up, the main disadvantage of protected vs unprotected is they are longer, so they don't fit into some slots on devices that use unprotected ones.

I think I would have bought the protected ones, but there was nothing on ebay which is situated in germany... I'm actually waiting more than one month for some neodyme magnets from china. :/

Is it safe enougth to solder a wire on the neodyme magnet and then put it on the cell? The magnet is 5mm diameter and 1mm thick. So it could not get shortet out.

Additionally I will put some insulation electical tape on it. Is it safe enough, or do I have to use battery holders? This will be very scarce with the space in the pipe for the circuits and other components...

Edit:

The pipe will get a huge massive stainless steel sphere which has a 10mm deep hohle and a flat end. The circuit should be cooled on that thing.
Since it are only 620 mA, I think this should be enough.

A smaller but longer through-hole should hold a very long switch (8x change/swap) for the logic which will wire the cells in series or individually to the cells own charger circuit shown in the first post.

I tested them and they're seem to be safe. Perhaps I will send one to Dave to check it on a mailbag monday, too. Would be interesting how much current flows when the 5V input is disconnected and the cell is still on that thing. Could not measure it because I have a Fluke 117C multimeter which has a range of minmum 1 mA. But I guess it is near zero. Okay, honestly I hope it to avoid a second switch which would disconnect all cells completely from everything. It's a pain in the ass, but there is no other way to do it. Pulse relais are too big, power transistors are too big and they need additional heatsinking and will drop the voltage... Ouahhh... Don't even think of it...

The charger circuit will be stacked and sticked with double sided thermal conductive tape on a round aluminum/aluminium core with slices in it. The USB-Jacks will be removed completely to save some height.

I noticed that the test LiIon cells are old and broken, so they did never consume 1 Amp more than 10 seconds... The new ones are sucking 1 Amp for 170 minutes! So I need any input jack that is capable of 4 Amps continously, no mather how long.

http://en.wikipedia.org/wiki/File:Photo-RoundPowerConnectors.jpg
The very right one would be okay, I thought... But I can not find ANY current rating at all.
Any idea?

Thanks!

[BloodySword]

I could use a Step-Down CC-Source. Does anybody know a small shielded module that is available for this?

The current should be adustable, because I don't have 700mA, I want 620 mA. Could also be 630 or so, no worries of inaccuracy but still a big step under the typical 700 mA rated for the LEDs to increase the lifespan as well as run time on battery.

I really doubt that there is something suitable for me... :/
But I kinda wish it would be...

What's missing then:

- The overdischarge protection at 12 Volts minimum, should only reset when it's more than 13.5 Volts. (charging)
- The max. current protection to protect the cells (has nothing to do with the LEDs)
- No LED output while charging and only 5 minutes after charging
- No charging when the cells are not under 13 Volts.
- The small 4 LED VU-Meter for the battery's charging level (when holding a small monumentary switch)
- A sollution for the 8 channel swap switch...

Still so much left. :/