Regulated 5v LiFePO4 supply

stevesmythe

Senior Member
I'm working on an improved version of my wireless AXE033 LCD weather station receiver/screen.With 3xAA zinc-carbon or 4xAA NiMH, when the LCD display goes dim, I know it's time to change the batteries.

I am drawn to the small size and low weight (and rechargeability) of a 2x3.2v LiFePO4 supply, regulated down to 5v. I'm thinking of using an LP2950.

My question is, what would happen in this scenario when the LiFePO4 voltage drops near to its recommended cut-off? The regulator minimum operating voltage is 5.5v (which would give me a bit of headroom before damaging the LiFePO4s), so will the output voltage collapse quickly below that - causing the display to dim? Is there any other way for me to tell, except by measuring directly with a voltmeter? I don't want to try it for fear of ruining a pair of expensive LiFePO4s.
 

premelec

Senior Member
Have you looked at the LP2950 data sheet? One of the usual characteristics of LDOs is they increase current at near dropout - I haven't looked at this sheet recently but it's worth actual testing as well as reading the DS - likely doesn't quit drawing entirely and that might be best taken care of by a relay or some such [other discussions on this forum]. In a recent discussion there was mention of using the units modelers use monitoring Li packs - these have alarm output when any cell drops below a set voltage and they are pretty cheap - the alarm signal could be wired to cut the power entirely...
 

stevesmythe

Senior Member
Thanks Premelec

It looks like the dropout voltage reduction is pretty linear - according to the DS, when the input voltage drops to 2.8V, the output voltage is in the same ballpark (depending on load).

OOPs - forgot that it's 2x2.5V that I'm worried about! I'm still none the wiser
 

Attachments

Last edited:

manuka

Senior Member
Stevesmythe: Are you using a dedicated charger or solar? Is this need just for the indoor receiver (which thus should be handy for audible or visual alerts) ? For the latter perhaps consider even a paralled white LED & simple switch that can be used to test the cell(s). Usefully white LED's dim & cease glowing at the LiFePO4 low voltage cell limit of 2.5V. See more on such skinflint alerts here.

You'll no doubt find -as I most certainly have- that 3.2V LiFePO4 are near bullet proof cells, and I've yet to have any problems with a swag of AA/AAA 14500 sized ones used here in diverse applications- many outdoors.

Several have been accidentally run to dead flat for some time, yet they've come back when smart charged. I certainly don't recommend this however ! Stan.
 

stevesmythe

Senior Member
Yes, this is the indoor receiver and I have a dedicated charger. The outdoor transmitter has 1x14500 sized LiFePO4 and I have programmed the transmitter Picaxe to go to sleep when the battery voltage drops below 2.9 and to pulse a warning LED every SLEEP cycle. Because the indoor unit has an LCD screen a single LiFePO4 isn't enough - hence the 2x3.2V need to regulate the supply down to 5v. I have managed to kill a couple of LiFePO4s that my charger doesn't seem able to bring back to life. I have followed your articles and saw your test LED idea, but I'm looking for something that doesn't require me to press a switch. If the supply to the Picaxe is regulated then I can't use the Picaxe to monitor the supply voltage. Or can I?
 

manuka

Senior Member
One imagines a voltage divided PICAXE ADC read would do it? I haven't tried but recall postings re reading 12V battery voltages. Alternatively, as a 5V LCD won't need much current, perhaps consider a simple charge pump?

But are you still using that 16x2 display from your earlier posting (& did you manage more 433 MHz range?) ? As the electronic age now is VERY 3.3V orientated why not just grab a "3V" LCD ? They've become cheap enough - check this RPi slanted 16x2 from UK ebay or this 16x2 from China - the latter post free! Stan.
 

stevesmythe

Senior Member
I like your thinking Stan. With a 3.3v display, I could use a single unregulated LiFePO4 and the display would dim before any battery damage ocurred. Unfortunately, I haven't come across a readily-available 3.3v 16x2 with the pin headers on the lower left of the screen (as you look at it). The ones you linked to would leave the AXE132 sticking out the side. I'll keep looking, as that would be a very neat solution.

Yes, I'm still using the same setup as per my earlier posting. It works well in some parts of the house. The reason I am working on a version 2 is that I didn't really give much thought to the placement of the Dorji RX module since they had given such good performance wherever I had put them before. I only soldered it where I did as it required the minimum effort on my part! I was a bit surprised when it didn't get good reception. This time round I'm determined to get better reception, so I'll experiment with different ways of adding the RX module to the AXE132. The other disadvantage of version 1 is that I am using one of the precious spare 18M2 ADC pins for the CS connection on the RX module. This could be used for another (indoor) sensor, so I'm going to swap it for C5 (which would otherwise be used as the serial input if the display was driven in serial mode).

When I get it working properly, I think it will be quite a neat little "wireless serial display module".
 

manuka

Senior Member
- pin headers on the lower left of the screen-
Why not solder fine wires (perhaps ribbon cable sourced?) at the top to make flying leads terminated at header pins?

Indeed a near wireless serial display, but maybe consider "stroll by" smart phone approaches! Bluetooth or WiFi modules built into the receiver unit can resend data for display on handy & versatile smartphones - Blueterm is well thought of for Android use. If your outdoor weather station unit is close enough you may well be able to receive such 2.4GHz signals directly, & thus do away with 433 MHz.
 

Attachments

nick12ab

Senior Member
I like your thinking Stan. With a 3.3v display, I could use a single unregulated LiFePO4 and the display would dim before any battery damage ocurred. Unfortunately, I haven't come across a readily-available 3.3v 16x2 with the pin headers on the lower left of the screen (as you look at it). The ones you linked to would leave the AXE132 sticking out the side. I'll keep looking, as that would be a very neat solution.
Some LCDs including the Winstar ones Rev-Ed supply have pads where you can fit a low cost voltage inverter IC (7660) to generate the negative voltage that "5V LCDs" require on their contrast pins in order to work on a power supply of 3V.

Look here:
Modify Winstar HD44780 LCDs (like those supplied with AXE133) for operation on 3V

If your LCD does not have the required pads, building the 7660 circuit on a piece of stripboard will also work.
 

stevesmythe

Senior Member
Stan
Yes, I've used BT/Blueterm before on robot projects but I hadn't thought of the "stroll by" approach for connecting to a weather station. That'll make a nice project!

I'd just about given up looking for 3v LCDs with the pin headers in the right place when nick12ab's post appeared. Thank you Nick, that's an interesting discovery (n.b. the data sheet link in your blog post isn't working any more). However, I've never done SMD soldering before and I haven't got the right equipment, I don't think. I wonder whether there's an SMD soldering evening class I could go to???

At least in the short term, I'm going to go with the Chinese 3V LCD with the pin header in the wrong place and use a flexible ribbon cable to connect to the AXE132. I didn't much like the look of the white on blue UK one you found, Stan. I'll report back when the slow boat from China arrives.
 

manuka

Senior Member
I wonder whether there's an SMD soldering evening class I could go to???
No doubt they'd match the micro topic & hold it in a broom cupboard by torch light!

OK on the 3V LCD order - at such bargain prices & free shipping it should be worth a punt. Versatile smart phone apps. seem to be reducing basic LCD demand, so there are heaps of similar offerings from Alibaba, Deal Extreme etc- GOOGLE! Stan.
 

nick12ab

Senior Member
I'd just about given up looking for 3v LCDs with the pin headers in the right place when nick12ab's post appeared. Thank you Nick, that's an interesting discovery (n.b. the data sheet link in your blog post isn't working any more). However, I've never done SMD soldering before and I haven't got the right equipment, I don't think. I wonder whether there's an SMD soldering evening class I could go to???
Thanks, datasheet link updated.

The required surface mount parts are quite large so they can easily be soldered with a normal soldering iron tip, like they were in my pictures. Alternatively you can use the DIP version of the 7660 on external stripboard.

It's also possible to make a negative voltage generator with two diodes and two capacitors (no 'special' components required) but this requires a square wave signal such as a spare PWM pin on a PICAXE.
 

stevesmythe

Senior Member
OK Nick, you've convinced me - I'll give it a go. If I make a mess of it, I can probably take the bits off again and still end up with a working 5v display.

Please can you confirm that these are the right components to fit the boards?

7660

Capacitors

What value/type of capacitor would you suggest for C3?

Taking advantage of their free delivery/no minimum order value at the moment. HOWEVER, you have to buy at least 100 SMD capacitors, so that make £3.60 just for capacitors. You can buy a whole 16x2 5v display for that price!

[added later]
I just had a look at my "official" Picaxe LCD. It looks different from the first picture in your blog post as the space for the ICL7660 is at the opposite end from yours. Next to the IC pads it only seems to have R7, C3 and C2.

How do these correspond with C1, C2 and C3 in the data sheet you linked to?
 
Last edited:

nick12ab

Senior Member
OK Nick, you've convinced me - I'll give it a go. If I make a mess of it, I can probably take the bits off again and still end up with a working 5v display.

Please can you confirm that these are the right components to fit the boards?

7660

Capacitors

What value/type of capacitor would you suggest for C3?

Taking advantage of their free delivery/no minimum order value at the moment. HOWEVER, you have to buy at least 100 SMD capacitors, so that make £3.60 just for capacitors. You can buy a whole 16x2 5v display for that price!
Sorry for the late reply!

The IC you chose is correct.

0805 size surface mount capacitors are very small - only 2mm long - so go for larger ones. The LCD supports quite big surface mount capacitors and in fact, you could connect through hole capacitors there if necessary.

For C3, use the same as the other two.

I just had a look at my "official" Picaxe LCD. It looks different from the first picture in your blog post as the space for the ICL7660 is at the opposite end from yours. Next to the IC pads it only seems to have R7, C3 and C2.

How do these correspond with C1, C2 and C3 in the data sheet you linked to?
Sorry I don't have an official PICAXE LCD. A quick test with a multimeter will reveal how they are connected to the 7660.
 

julianE

Senior Member
Nick,
the 7660 booster has me thinking of another use, cheap 433 Mhz transmitter modules will give a longer range with a higher power supply voltage. I think the current from the 7660 is 10mA which might be enough especially if i charge a larger value capacitor. Do you think it will work in that application?
 

fernando_g

Senior Member
Stan
However, I've never done SMD soldering before and I haven't got the right equipment, I don't think. I wonder whether there's an SMD soldering evening class I could go to???
.
For SMD work you require essentially these 5 items:

-The finest pointed solder tip you can find.
-Fine gage solder. I use a 0.7 mm diameter one.
-Tweezers.
-A GOOD, illuminated magnifying glass.
-Liquid flux

A few practice runs with a kit like the one in the link below, and you are good to go.

http://www.dx.com/p/smd-soldering-practice-pcb-board-kit-green-149752#.VYLfREZpLRI
 

stevesmythe

Senior Member
Thanks Nick

Fernando You forgot "a steady hand" from your list! I have seen those kits on eBay and wondered whether they are worth getting.
 

manuka

Senior Member
cheap 433 Mhz transmitter modules will give a longer range with a higher power supply voltage.
.
JulianE: There's more to it than this!

1. Overdriving the transmitter (TX) may introduce distortion

2. It takes a boost of 6dB (= x4 the power) to double range under the same conditions.

3. Local power limits are usually a mere 10 - 25mW. If you want more TX power than the few mW of dirt cheap ASK modules consider Dorji's that are rated as 25mW at 3V supply. Prices are usually much the same.

4. Better receivers (RX) greatly help pull weak signals out of the noise. Some are deaf & can barely work across a room, while others are up for 100s of metres with the same TX. ( Receiver supply voltage can be an issue too - some insist on 5V)

5. Superior modulation schemes (FM or GFSK rather than ASK) can give interference & noise immunity.

6. Height can be everything at UHF - elevate the TX above obstacles.

7. TX gain antenna are usually not allowed, but RX ones certainly are - even a simple Yagi will give 6dB gain.

8. Lower data rates usually give improved reliability at extended range - try 1200bps.

9. If you really want RANGE then consider LoRa™, which has at least an magnitude improvement over regular modulation!

HTH - Stan.
 

premelec

Senior Member
For soldering SMD the stuff required is minimal - really - a tiny hot tip :) And to correct the mistakes you'll make decades of experience have proved to me that one of the most useful tools for clearing holes and chasing shorting solder bridges to clear them is.... TaDa! a wooden toothpick slightly charred and impregnated with flux... heat the solder and poke it around with the dirty toothpick... [I use flat picks as they don't jam when poking into holes]. You can renew or shape them with a sharp blade... or whittle your own from a local good smelling wood... :) You can also hold the component in place while initially tacking one pin - with a toothpick...
 

julianE

Senior Member
Stan,
Thanks for the detailed info. I just received the better modules and will be learning to use them. I have the cheapo ones working well at 1200 baud, about 7 meters now with simple wire antennas. I found an instructable (might be yours) on how to make yagi antennas for the cheapo modules out of tape measure metal, i may give it a try, if i can get a 6 db boost i should be golden, all i need is about a 12 meter range and I have 7 now.

I best not add to this thread much more since it's about a totally different topic.

Julian
 

manuka

Senior Member
Yes- guilty as charged on the tape measure Yagi. But that 7 metres range you mwntion must surely be thru' the likes of solid rock or lead shielding? You otherwise are doing something seriously wonky for such a pathetic LoS (line of sight) range at 433 MHz ! Send me a PM & we'll discuss it. Stan.
 

julianE

Senior Member
The 7 meters is far from line of site. There is a thick brick wall and also a cinder block partition between the two 433 Mhz modules. The modules are the $1 Ebay ones. I do have a regulated 5 volt power supply on the receiver. I'm rather impressed by the modules, getting them to work is a trifle compared the RFM22. Provided I keep bluetooth modules away the link is solid with very little spurious data and I live in a large city with a lot of rf noise. Of course to get longer ranges it's worth the extra effort of the better modules.
 

eggdweather

Senior Member
The problem with LiFePO4 batteries is the very sharp knee between their nominal 3.2v and being empty and the capacity difference is a few mahr meaning they transition from 3.2v to typical 2v in a very short time, there is little to no advanced warning. I've tried many techniques of voltage monitoring but most are futile unless your actually there to see the transition take place. You might get at best 1-hour notice of the transition to empty.
 
Top