Using a single cell LIPO to power an 8M2 or 14M2

I am planning to use an 8M2 or a 14M2 on a project to drive 8 UV LEDs. Considering using a 3.7V LIPO. What can I expect to happen when the voltage of the LIPO drops below 3V
 

PhilHornby

Senior Member
The Picaxe will keep running - potentially down to as low as 1.8V (I believe), depending on what internal hardware is configured. I'd assume you'd need to monitor the supply voltage and 'sleep' when you get down to some minimum safe voltage, to avoid damaging the LIPO.

It would probably stop being able to drive the LEDs at some point before this - whether this 'point' would be ok for the LIPO, I don't know.

(I'm not a LIPO expert ;) )

(This clashed with @inglewoodpete 's message - but, what the hell, I've typed it in now :) )
 

AllyCat

Senior Member
Hi,
What can I expect to happen when the voltage of the LIPO drops below 3V
As already said, the main "concerns" should be with the LiPO and the LEDs, not the PICaxe. As far as I can see, UV LEDs require a forward voltage of more than 3v (as do white and blue LEDs); also, LiPO cells should not be discharged below about 3 volts (and might/should contain a "protection" chip to switch off any output drain below this voltage). But an 08M2 will keep "working" down to about 2.2 volts and the 14M2 to about 1.8 volts. Note that the PICaxe can use its CALIBADC instruction to measure the (battery) supply voltage without any additional components or pin connections.

I don't know how you plan to drive 8 LEDs from a (4 output pins) 08M2 but don't expect the PICaxe to deliver anything near to 20 mA through each pin at 3 volts. The graphs (e.g. FIGURES 31-43/44) towards the end of Section 31 of the relevant base Microchip Data Sheet shows that at 3 volts supply, you can expect about 4 mA pull-up or 10 mA pull-down current for each pin, with a 0.5 volt drop.

Cheers, Alan.
 

geezer88

Senior Member
If you look on the popular online seller sites you can find tiny boards about the size of a postage stamp that provide a micro usb charging port and a cutoff to protect the lion cell. I've used these on several projects, and they are usually priced around $2.
tom
 

premelec

Senior Member
Look in Ebay for "USB Power Bank Case 18650 Battery Charger DIY Kit" - there are many varieties... usually holding one or 2 18650 cells; the key is that they are Do It Yourself "battery bank" units... have fun... If you want just the protection circuits rather than up converter part they are under cell protector..."BMS Protection Board " for one to many cells in series... [Battery Management System = BMS]
 
Last edited:

AllyCat

Senior Member
Hi,

For low currents you can do much the same with a few lines of code in the PICaxe, or try the following descriptions for your "local" website:



Overcharge detection voltage: 4.28 ± 0.05V
Overcharge release voltage: 4.26 ± 0.05V
Over discharge detection voltage: 2.75 ± 0.1V
Overcurrent detection current: 1-3A
The main function: Overcharge protection, over-discharge protection, short circuit protection, overcurrent protection.


Cheers, Alan.
 

bogbean

Well-known member
For low currents you can do much the same with a few lines of code in the PICaxe, or try the following descriptions for your "local" website:
I have been mulling over some battery protection to add to a 14500 battery powered (recharged by a solar panel approx 50mm x 50mm) circuit that flashes red and blue LEDs to tell me the temperature at the bottom of my garden. My existing circuit is identical to that of the attached schematic but has no MOSFET between the solar panel and battery (only a diode). It has been running well since last December but the 14500 battery does not now hold a charge. I wonder if I have damaged it by exposing it to the approx 5.5V the solar panel can generate (less a diode drop).

I would like to protect against discharging the battery below 3.3V and against over charging it above 4.2V; for both I would sense the battery voltage using calibadc10. To prevent over discharge I could disable flashing of LEDs and sleep in an attempt to save power. To prevent over charging of the battery I am thinking of switching off the MOSFET (Q1) shown in the schematic. The circuit consumes low power, I found it necessary to average less that 1mA because of the limited size of my solar panel (the panel and battery are part of a solar light I have modified, convenient because of the enclosure they come in).

I am still getting my head round MOSFETs and have never used one but it seemed a sensible component to use in this application. My next step would be to find a suitable MOSFET to try the circuit out but I wondered if anybody could spot any obvious problems before I do that?

I had considered buying protection boards like the ones suggested above but it would be nice to solve this with a picaxe.
 

Attachments

erco

Senior Member
Of course a Lipo is dead at ~3.4v and permanently damaged at 3.0v.

 

AllyCat

Senior Member
Hi,

Yes what you describe is perfectly possible, but probably overkill. It might even be possible to devise a "zero pin count, zero added hardware" solution, i.e. Software Only. ;)

Your "Low battery voltage" method is exactly that - CALIBADC10 needs no pins or external components and you "adjust" the load by switching off the LED(s) and SLEEPing the PICaxe. For the "Overvoltage" situation you could just increase the power consumption - known as Shunt Regulation. Normally that is considered "wasteful" of power, but with a PV panel you can't make the energy disappear! The series Transistor/FET would stop the energy getting to the Battery/PICaxe, so the Panel gets hotter instead! Probably not particularly damaging to the Panel, but it's usually better to dissipate power in a component designed to get hot, such as a resistor. :)

If the excess power isn't too great then you could use the opposite of SLEEP, i.e. SETFREQ M32, which will increase the PICaxe's supply current to about 4 mA. For higher currents, a resistor of about 150 ohms could be connected between an available output pin and the supply, which can drain up to 25 mA, or 100 mW. For higher currents still, you don't need much more than a simple shunt NPN transistor (e.g. BC337) or FET (2N7000 or BS170). An advantage of the PICaxe is that you can set your own limits and I'd probably try a conservative range of 3.6 v - 4.0 v.

A 5 cm square crystaline silicon solar cell might deliver up to 0.5 watt (90 mA) assuming 1000 watts/m2 and 20% efficiency. But the majority of "garden" lights use Amorphous Silicon with nearer to 10%, or even 5%, efficiency, and of course that light level only occurs with very bright sunlight striking the panel near to perpendicular. In practice you probably won't ever need to dump more than an average of 20 mA over any 24 hour period, so a single external resistor should be sufficient.

Cheers, Alan.
 

papaof2

Senior Member
If you can find a small enough LiFePO4 cell, they're often rated for 2000+ cycles to 20% SOC (over 5 years at one cycle/day).
I found some 1100maH 18650 LiFePO4 cells with an attached one-cell BMS so the overcharge/overdicharge concerns are handled. Try batteryhookup.com (in the US) for some good deals on surplus and overrun cells. Although an 18650 cell probably won't fit the existing enclosure :-(
 

AllyCat

Senior Member
Hi,

Yes, my preference is for LiFePO4 cells if their 3.2 volts is sufficient to drive the LEDs. They're certainly available in 14500 (AA) format; I used to buy them from Maplin and believe they're still available from battery specialists as a "Replacement to use in Solar Garden Lights". But check that they're not NiMH which are used with a boost converter in the cheaper lights, and both versions often have an unusually low capacity, e.g. 200 mAh or 400 mAh (sufficient for a daily cycle). Also they can be quite expensive, particularly if the supplier complies with the postal/shipping restrictions on Lithium cells. :(

Cheers, Alan.
 

bogbean

Well-known member
For higher currents, a resistor of about 150 ohms could be connected between an available output pin and the supply, which can drain up to 25 mA, or 100 mW. For higher currents still, you don't need much more than a simple shunt NPN transistor (e.g. BC337) or FET (2N7000 or BS170). An advantage of the PICaxe is that you can set your own limits and I'd probably try a conservative range of 3.6 v - 4.0 v.
I've implemented the 150 ohm shunt resistor solution for now with routines for high and low voltage protection. I have quite a few 18650 lithium ion batteries (I don't know the exact chemistry) available so will use those for now and monitor how the battery varies over the year (I have the device 'flashing' the battery voltage as well as temperature now). I see that LiFePO4 batteries seem to be capable of many more charge / discharge cycles so perhaps I'll try those in the future.
My small solar panel was generating 32mA at 16:50 BST yesterday in direct sun in London, which I was impressed by. I've estimated it should never generate more than 44mA (i.e. at noon on the summer solstice) and I think that the shunt regulation will be enough to protect from overvoltage. I'll have to wait 9 months to find out. Thanks for the suggestion, it's very satisfying to learn how to do this with picaxe instead of buying a separate controller board.
 

AllyCat

Senior Member
Hi,
My small solar panel was generating 32mA at 16:50 BST yesterday in direct sun .... it should never generate more than 44mA ...
That's probably enough to have prematurely "killed" an unprotected Lithium Ion cell, although it's normally over-discharge that causes the problems. It might be worthwhile to set the "overvoltage" level at (say) 4.0 volts (at least in summer) since you might not be able to pull any excess voltage rise back until the following night. Even running all the PICaxe's internal hardware "flat out", I can't see how to make it drain as much as 10 mA. :)

You didn't describe the "modified solar light", but another trick could be to mount the PV panel facing south at 45 or even 60 degrees to the horizontal. That would increase the energy generated in the winter and reduce it in the summer. ;) Typically, a PV panel in UK will generate 4 times more power in June compared with December.

EDIT: Using my favourite PV System calculator for London, a Horizontal panel generates 9 times more power in June compared with December. Maximum total yearly generation uses a slope of around 40 degrees, but 60 degrees to the horizontal reduces the June/December ratio to only 2. Or a completely vertical (south-facing) panel gives the most "constant" output over the year, about +/-20% on a nominal 60% of the PV rating. The data is clearer when using the "Grid Connected" calculator; the "Off Grid" calculator needs care with the units, but suggests that a 250 mW panel (e.g. 50 mA @ 5v) mounted vertically can deliver an average of at least 10 mW (2 mA @ 5v) throughout the year at a 70 degree slope and assuming an 18650 battery.

Cheers, Alan.
 
Last edited:
Top