AAA Battery High Discharge Rate - NiMH - Troubling problem needs solution
JBWolf
Posts: 405
Hello again everyone!
I have a problem with supplying power to led's that is vexing me.
The project is a portable LED strip. I have spent several months in designing display patterns with great success, but am having a bit of trouble with powering them.
The LED strip is a standard RGB ws2811 available on ebay. Each LED on the strip is a 5vdc RGB diode. For designing the patterns, I have been using a 5vdc bench power supply.
The bench supply can deliver 3A, which has proven more than adequate for the needs of the strip which contains exactly 200 led's.
The display patterns range from only a single led on at one time only using one element (green or red or blue), to all 200 led's on the strip set to white (all led's with all 3 elements running full brightness). The 3amp bench supply does a great job.
I am using an LM1084 to change 5vdc to 3.3vdc for the prop, everything powered off the same 5vdc supply (LM1084 and 200 led strip). The prop has nothing extra, just the minimum needs... a 5mhz crystal, a 256k eeprom and a single pin running to the ws2811 led's.
The problem I am having is almost certainly caused by current dropout from not having enough amperage on the batteries.
Now a serious restriction that determined the type of battery I could use was diameter. I need 5vdc for the led's and so four AA's or AAA's in series will provide this, however due to case size I am limited to 10mm diameter batteries which make AAA's in series the only viable solution.
Here is where things start getting confusing and weird.... choosing the batteries: I cannot find any Li-ion batteries that have a high current output rating, so I have only tried using Ni-MH batteries. I am still wondering if there is an Li-ion solution and have not tried any, please let me know if you have a recommendation after reading the following.
I first started with trying to find a AAA battery with a high discharge rating, which energizer 'power plus' 900mah Ni-MH batteries where readily available and after testing, worked fine for smaller strip lengths of 125 led's... but for lengths of 200 led's cause the prop to lockup and hang on some display programs (most likely from amperage dropout since the 3A bench supply works just fine and the batteries also work fine for a smaller strip length). Some of the displays that only light a few led's at a time worked just fine, but locks up on displays that light up all 200 led's.
So I ordered 1300mah AAA NiMH batteries off ebay... but these proved to be a total bust... did not work at all, period. Even for a strip length of 100 led's, these batteries proved to be completely worthless. I'm guessing that they are designed for a slow discharge rate as they do not work at all compared to the 900mah energizer 'power plus' batteries. Actually the 900mah energizer AAA 'power plus' batteries will run all 200 led's set to white for a few minutes after a fresh recharge... but then starts locking up the prop shortly after. I can then re-recharge the batteries and get the same result. To make absolutely sure, I then connected the 3A bench supply and find it runs the exact same display with no problems... everything wired identically.
So theres where i am currently, confused and trying to find a solution. I need to power 200 led's and dont know what battery solution to choose. I'm sure AA's would do it as I have found 'power plus' energizers rated to 2500mah, but they are just too big for the casing.
High output AAA's from ebay rated to 1350mah dont work worth squat... the 900mah 'power plus' energizer batteries work for just a short while.... and I dont know what to do.
Seems I need a rechargeable, high discharge AAA battery rated above 1A. but since the 1300mah ebay batteries didnt work, I dont know what to look for.
Li-ion would certainly be preferable since they output full power until discharged, but they seem to have a much lower output rating than Ni-MH.
I have already spent nearly $200 in batteries... hoping someone can help me out with a possible solution
I need a AAA that can dump lots of power regardless of lifetime.
I have a problem with supplying power to led's that is vexing me.
The project is a portable LED strip. I have spent several months in designing display patterns with great success, but am having a bit of trouble with powering them.
The LED strip is a standard RGB ws2811 available on ebay. Each LED on the strip is a 5vdc RGB diode. For designing the patterns, I have been using a 5vdc bench power supply.
The bench supply can deliver 3A, which has proven more than adequate for the needs of the strip which contains exactly 200 led's.
The display patterns range from only a single led on at one time only using one element (green or red or blue), to all 200 led's on the strip set to white (all led's with all 3 elements running full brightness). The 3amp bench supply does a great job.
I am using an LM1084 to change 5vdc to 3.3vdc for the prop, everything powered off the same 5vdc supply (LM1084 and 200 led strip). The prop has nothing extra, just the minimum needs... a 5mhz crystal, a 256k eeprom and a single pin running to the ws2811 led's.
The problem I am having is almost certainly caused by current dropout from not having enough amperage on the batteries.
Now a serious restriction that determined the type of battery I could use was diameter. I need 5vdc for the led's and so four AA's or AAA's in series will provide this, however due to case size I am limited to 10mm diameter batteries which make AAA's in series the only viable solution.
Here is where things start getting confusing and weird.... choosing the batteries: I cannot find any Li-ion batteries that have a high current output rating, so I have only tried using Ni-MH batteries. I am still wondering if there is an Li-ion solution and have not tried any, please let me know if you have a recommendation after reading the following.
I first started with trying to find a AAA battery with a high discharge rating, which energizer 'power plus' 900mah Ni-MH batteries where readily available and after testing, worked fine for smaller strip lengths of 125 led's... but for lengths of 200 led's cause the prop to lockup and hang on some display programs (most likely from amperage dropout since the 3A bench supply works just fine and the batteries also work fine for a smaller strip length). Some of the displays that only light a few led's at a time worked just fine, but locks up on displays that light up all 200 led's.
So I ordered 1300mah AAA NiMH batteries off ebay... but these proved to be a total bust... did not work at all, period. Even for a strip length of 100 led's, these batteries proved to be completely worthless. I'm guessing that they are designed for a slow discharge rate as they do not work at all compared to the 900mah energizer 'power plus' batteries. Actually the 900mah energizer AAA 'power plus' batteries will run all 200 led's set to white for a few minutes after a fresh recharge... but then starts locking up the prop shortly after. I can then re-recharge the batteries and get the same result. To make absolutely sure, I then connected the 3A bench supply and find it runs the exact same display with no problems... everything wired identically.
So theres where i am currently, confused and trying to find a solution. I need to power 200 led's and dont know what battery solution to choose. I'm sure AA's would do it as I have found 'power plus' energizers rated to 2500mah, but they are just too big for the casing.
High output AAA's from ebay rated to 1350mah dont work worth squat... the 900mah 'power plus' energizer batteries work for just a short while.... and I dont know what to do.
Seems I need a rechargeable, high discharge AAA battery rated above 1A. but since the 1300mah ebay batteries didnt work, I dont know what to look for.
Li-ion would certainly be preferable since they output full power until discharged, but they seem to have a much lower output rating than Ni-MH.
I have already spent nearly $200 in batteries... hoping someone can help me out with a possible solution
I need a AAA that can dump lots of power regardless of lifetime.
Comments
http://data.energizer.com/PDFs/nh12-900.pdf
I need a AAA battery with AA power
'Not cheap, though.
You may be better off with LiPOs in flat packs, if thickness is an issue.
-Phil
http://www.amazon.com/BPI-High-Voltage-Rechargeable-Batteries-Original/dp/B008J7RNFG
I've got a set of Varta "Charge & Go" that are spec'd for 15 minutes for a full charge. And can be discharged just as quickly. They do get hot but they've certainly worked. The charger came as part of the pack and actually takes about 20 minutes from totally flat to charged for 4x2000mAH AA cells. Haven't ever looked for AAA so wouldn't know if they exist in this form.
Of course the other way to do it is to have a boost convertor either for the Prop or just for 5V.
EDIT: Here's a TO92 part if you are after through-hole, it's cheap and it will be around 50mv drop.
I will probably try out that MCP1700 sometime though, thanks!
I do need rechargeable batteries as the casing is sealed permanently. I have been connecting 4x AAA NiMh in series and charging together with this: http://www.all-battery.com/smartuniversalchargerfornimhbatterypackRed.aspx
At full charge they put out 5.4v tested on my MM. I have worked out a way to get them all balanced in charge before permanently connecting in series as a battery pack.
I have looked into NiZn previously, but read about terrible discharge degradation and reliability. Also the current rating is less than NiMh... 1.6v @ 1000mWh = 1 / 1.6 = 625mAh rating. I have 900mAh NiMh that work for just a short while after a fresh charge.
I have tried 5 batteries to make sure voltage is always above 5v, but this got the exact same results... runs for few minutes then starts locking up which is why I am almost sure the amperage is the problem. Once again I can get it to work fine with a 5v 3A bench supply.
They are the highest discharge rating I have found on a reliable brand... praying they work
Can I suggest that you could try running your Prop at a slower clock speed (even RCFAST) and disabling the brown-out (BOE pin)? This will lessen the load on the regulator which could probably also do with better output filtering, using a 22uF tantalum and 0.1uFs. Normally I would isolate the input of the regulator with a diode but even a Schottky will drop too much. The idea here is not to get rid of any supply problems completely but to make the system more tolerant to it.
Focus more on simple things you can do easily first and you may find that the battery solution is unnecessary.
seems like a LiPoly battery might be found with the right capacity and size?
I would by far prefer to use Li-ION as the output is stable until the battery is discharged, means no loss of brightness over time. But I could not find any Li-ION batteries with 10mm diameter and an amperage output of around 1000mah.
With LiPo, I could find many with sufficient amperage, but none with a 10mm diameter. Most of these seem to be square/rectangular packs.
If you know of any possible product that can fit a 10mm diameter and has a high amperage output, please let me know, I would greatly appreciate it.
I am using an object that relies on a fixed clock speed. So unfortunately this doesnt look to be an option.
However after looking at your regulator suggestion I am wondering if that MCP1700 is a better choice than the LM2937. The LM2937 has an output of 400ma, whereas the MCP you suggested outputs 250ma. Assuming that the regulators provide this current despite usage, this means your suggestion would save 150ma consistently correct?
I am using almost no extra power off the prop.... no led's or anything, just a pwm signal off of a single pin to the WS2811 led controllers.
So im not sure what the prop uses, but i dont need anything extra. For the controller wiring & components, I used the exact schematic provided in the PEkit book. The BOEn pin is directly connected to ground, I believe this is bypass mode correct?
I am using two .1uf capacitors on the prop power inputs (using QFP chip), after reading the LM2937 schematic, I believe I should add a 10uf capacitor to the Vout pins for stability.
I doubt very much that this object could not work with slower clock rates, it's a very easy thing to change and most objects will adapt to the settings.
Now, about the regulator, the current rating is a maximum that the device will handle but there are other limiting factors as well, so it won't "save" power, it doesn't mean that. The important figures are in the datasheet and they have to do with "dropout voltage" etc. The MCP1700s are one type that are readily available, cheap, and have good figures in relation to dropout voltage and also I use them because they are stable with 1uff ceramic caps rather than larger tantalums. In your case you would want some extra capacitance on the load side anyway. Also while the LM2937 says it can handle more current than the MCP1700 this is sometimes a misleading figure as linear regulators are limited by how much heat they can handle which also depends upon other factors too.
EDIT: I've had a look at the LM2937 data and it's still not a good LDO, the dropout is 250mv (typ) @50ma. Although this device is suitable for working with higher input voltages (which would severely limit the current) there are better ones for your application, the MCP1700 will have about 25mv dropout for your application.
I would like to get the MCP in a surface mount package... will that be ok or should i go with through hole because of the high current being pulled by the led's? I've had similar circuit setups where an SOT lm317 burned out even though it wasnt powering the high current device.... had to go with a TO220 package to handle the heat.
Any other recommendations?
I am also planning on switching to an SMC version of the eeprom: 24LC256
and I have found surface mount 5mhz crystals, but its the same size as through hole, I would like to find a smaller one.
hoping the new 950mah AAA batteries arrive tomorrow.
if they dont work, i am going to have to figure out a way to get AA's to fit.
Anyone know of a high power Li-ion in AA (prob 14500 size) that might do the job for this many led's?
I have looked at a few 14500's, but they were all rated at only 600mah and such... I know li-ion doesnt have the energy density that ni-mh does, but this is considerably less than a 950mah nimh AAA. was hoping someone might know where i can find >1200mah AA's in li-ion... I have found energizer makes nimh AA's up to 2500mah, so there is definitely a nimh solution if li-ion isnt viable. It's impossible to fit 18650's unfortunately.
I have no problem with nimh, but would prefer li-ion just for the advantage of full output power until battery is near fully drained.
I use the SOT-89 version of the MCP1700 and the reason you may have had to go to a larger package before has more to do with how much power (volts*current) it has to dissipate. For 4.8V in to 3.3V out it is only 1.5V that it has to drop on average times the current, say 100ma --> 150mw, barely warm at all. If now you wanted to power this from 24V then it would be 20.7V * 100ma --> 2W which would be very hot and need a large package and heat sink, but then in that case a switch-mode would be far better. BTW, the MCP1700 only needs 0.1uF ceramic caps although as I mentioned earlier your app might benefit from a 22uF tantalum or so for the 3.3V side esp if you can run the Prop at a slower speed.
For the EEPROM you can use the 24LC256 but you might benefit from a 64K one for a few cents more or this one which I use which costs less and runs faster!.
I have been using 10MHz crystals in my designs for quite a few years and they work perfectly so I recommend either a cylindrical type or something like these cheap little 5x3 guys. Look at this post here to see a pcb layout with it.
If the batteries and LDO etc (esp the slower clock speed) doesn't help then remember what I said about boost regulators. This may mean you can get by with another type of battery or even 2 AAs.
BTW, What did you mean by " because of the high current being pulled by the led's" ? Aren't they run directly from the batteries and not the regulator?
Nope, NiMH are 1.35V, dropping below 1.25V only when nearly flat. Fully charged you will get 5.4V from 4 cells.
High discharge rates will alter this of course, but they are definitely higher voltage than the NiCd cells they replace.
To the original poster - connect cells in parallel for more current?
I believe i read in the object that i am using, that the clock speed has to be fixed because of the ASM code: jm_ws2812_demo
Unless i misinterpreted, this means i cannot change clock speed.
Regulator SOT:
yes the led's are running directly off the batteries/source, only the prop is running off a regulator.
Come to think of it, the other project did use a 12vdc source @ 5A... that was the project that was burning out the SOT lm317.
That was dropping voltage even more than this one... down to <2v @ 250ma actually.
While the SOT lm317 only used a little power... there was another device running off that power source that used 12vdc @ 4A... was afraid that was causing it to burn out the lm317 like it was getting caught in the wake of the higher power device since they were both being powered from the same source.
Was afraid something similar might happen with this project. I am designing and cutting my own circuit boards with a CNC, so any design changes take a large investment of time, just wanted to make sure before going from TO220 to SOT.
EEPROM:
You think it would be a better option to go with a 64k eeprom instead of 256k? would that save any significant amount of power?
Actually I'm not sure if my program is too large for a 64k eeprom, the code is actually just over 100 pages long and is 204kb on my pc.
I would be interested to see what kind of visual improvements might happen by going with a faster eeprom.
I have made 75 display routines, took a few months
Crystal:
I had a question in another thread about going with am SMC crystal.
In that discussion they said that it was important to keep load capacitance at around 20pf... i dont fully understand load capacitance so I'm not sure why.
But that 5x3 would be absolutely perfect in size! I see it's 10mhz, so just halve the multiplier correct?
Battery voltage:
With 4x fully charged AAA batteries, I typically get about 5.4v.
On lower number of led's (like 100 instead of 203 im using now), they run for more than an hour with no problems.
Unfortunately the LED seller didnt have a datasheet available, so I dont know what the minimum voltage is before dropout, but i would guess 4.8v or possibly less which seems to match the battery discharge well. Didnt have any trouble until doubling the amount of led's
All crystals have two resonance modes: series and parallel, and the oscillation frequencies of the two are slightly different. Crystals sold to be used with parallel-resonant oscillation circuits are calibrated by the manufacturer to be used with a certain load capacitance. A different capacitance than what the manufacturer specifies will result in a slightly different oscillation frequency. The Propeller uses parallel-resonant crystals, and it provides the load capacitance internally. The value of that capacitance depends upon which XTAL mode you select in your program. XTAL1, the most common, provides about 18pF per pin. So when you buy a crystal to use with the Prop, choose one calibrated for parallel resonance with a load capacitance between 18pF and 20pF in order to get the most accurate manifestation of the nominal frequency.
Although 10 MHz is outside the recommended 4-8 MHz range for use with a PLL, several Parallax products have used 10 MHz crystals quite successfully, and I would not hesitate to spec one again if board real estate is at a premium. Just be sure to use PLL8x instead of PLL16x.
-Phil
That would only affect the animation speed. Generally there is a limit to the maximum clock rate but you can run slower.
EDIT: I'm looking at the WS2812 datasheet now and I will get back to you about this.
It does look like it has a minimum speed but the only reason we wanted to slow down was simply to make the Prop draw less current so that the supply might hold up a bit better. Did you look at disabling brown-out by disconnecting the BOE pin from ground. This will enable the Prop to continue running at a lower voltage assuming it's not too low.
Maybe you need to read up about regulators, how they work. When you say " only used a little power" it doesn't matter what load there is, what is important is the heat load the linear regulator has to handle, power is volts*current. If you have a load running at 2V @250ma then that is only 500mw but the 12V supply will still have to supply 250ma, that's the nature of linear regulators. But 12V * 250ma = 3W yet the load only accounts for 0.5W so there is 2.5W being used up somewhere and that somewhere is the linear regulator. An SOT regulator might be spec'd at 250ma or more but you CANNOT use it at this current if you exceed the power rating of the component, SIMPLE. Even a 5A TO220 running at 250ma will burn up at 2.5W without a heatsink, and it would have to be a good heatsink. Linear regulators are the wrong type to use in these instances and switching regulators are a far better choice.
Never make the mistake of reading 256kb and think that that is more than 64kB. One is in kilobits and the other in kiloBytes. Having an extra 32k bytes with the 24LC512 is handy sometimes and esp. if it's cheaper than a 32k byte 24LC256 EEPROM. Don't get confused either with 24LC or 24F etc as the letters are just used by the manufacturer to describe the variation in the silicon.
For someone who glosses over linear regulator specs you seem to be very picky about the crystal without knowing why. Just trust that one you've been pointed to is fine and work out the rest at a more convenient time, like after you get everything working first. The recommended 4-8MHz range is for when you are running it at the extremes of temperature and voltage etc and I don't know of any product that gets close to that.
Checking dropout is easy, just vary the voltage down until it drops out. But anything near the dropout won't be very useful unless the LEDs have proper current regulation as resistor current limiting will not be able to compensate.
Unfortunately the PE education book doesnt give any details on the BOE pin and how to use, just a standard wiring diagram to follow... wasnt sure until now if running straight to ground was to disable it or if there was a voltage divider method to enable it or what.
I did see that the eeprom you recommended is actually 512kb in the spec vs the 256kb one i am currently using... unfortunately it is near twice as expensive in cost, so I might hold off on this until the program grows some more.
I cannot get the prop to run stably off off batteries at all.
I am using the basic schematic provided in the PEKit for the prop circuit, then running the WS2811 leds off a single pin.
I can plug it into a 5v switched supply from a 110v source, and run that 5v to the LED strip and then through an LM2937 to the prop which results in flawless operation through every LED pattern I have designed.
if I use the exact same setup but instead of the 5v switched supply, use 4x AAA batteries (950mah nimh in series), it runs for a short time then locks up the prop chip.... the led's stay lit but stuck at the last pattern, and I cannot communicate to the prop any more through the propclip unless I disconnect and reconnect power. But then I have to work fast as it will then lock up again within 60 sec.
So I tried using 4x AA's (2500mah rated nimh in series).
exact same results.
So I tried using 2x 18650 Li-ion batteries in series and add a 5v regulator since they give almost 8v out.
exact same problem.
so I figured maybe since the switched supply works fine, maybe amperage was the problem?
So I reduced the length of the led strip by over 1/2, then reprogrammed the brightness settings to 1/2 what they were.
I tried every battery configuration again and got the exact same results.
I just dont get it.
why is the prop locking up with batteries?
I have also tried adding all kinds of capacitors (1kuf + 47uf + .1uf)... tried them in every configuration, from all 3 together to just one or the other.
Just seems like the prop doesnt want to work with batteries
Can anyone please help.
I have built 4 different controllers and used 3 different led strips in testing... all give exact same results.
I know its not a problem with specific led display programming as they all work perfectly with the 5v wall supply.
Edit: nevermind just saw how many LEDs you're working with.
What about a small SLA battery like used in alarm panel backup power?
I only have 72 LEDs and have run them off 4 alkaline AA successfully, but 200 and full white, that is a lot. What if your bench supply is maxxing out at 3a and it's really pulling more? My strip of 72 at full white is close to 2a.
Also are you using a logic level shifter to get to 5v for the strip?
Anyway it looks like you have some other problems too. Have you looked at the output of the LM2937 on an oscilloscope? Regulators can be quite picky about output capacitors and I've had a few oscillate on me. It's especially important to check, as the Prop is quite tolerant of dirty supplies. One regulator was outputing about a ~7KHz 300mVp-p triangle wave that averaged 3.3 volts. The prop ran fine for years before I caught that problem.
Marty
I don't think SLA is an option for the OP
That's a small area to pack the power into.
can you give an example of the li-ion pack you mentioned? 20c+ should do the trick.
I've reprogrammed the brightness levels to about 1/2 max brightness, just enough to look good still but low as possible to reduce power consumption.
it does work fine on some modes where just 10 or so led's are on at one time. but there are many modes where every led is on (only one mode has them all white and i run that at 20% output brightness).
to get 5v, im just using 4x eneloop 950mah Nimh batteries in series which gives about 5.4vdc.
The KAN batteries are enticing... i need to learn a bit more about the output ratings... does 10c refer to charge rate or output rate? cause a 10c output rate froma 400mah battery would be 4amps, my wall supply gives 3.5amps (which works beautifully) so that means this battery would do the job right? I would use 4 of them in series to get 5v naturally... so i just run 4 in series right to the led strip for 5v, and then use a regulator on the prop to change 5v into 3.3v.
I dont see anything like that for output details on the eneloop batteries.... but either way I already invested in a bunch of nimh chargers, so i do need a nimh solution.
The KAN battery is 9.9mm diam which is perfect - would you recommend these for high output projects over the super expensive name brand eneloop batteries? they were the highest quality battery i could find that actually outputs what it advertises.
they work for about a minute but then lockup.
I have ordered some of the MCP1700 3.3v regulators to lower the dropout voltage... it also reduces power to the prop to 250ma vs 500ma from the LM2937.
I also ordered a LP2950 that has a super low drop out voltage and only 100ma power output to prop.
I sent a message to the guys at cheapbatterypacks.com to see if they can offer a solution.
I did not know about the difference between high mah rating and high discharge burst... I thought they were the same thing.
I wish I had a way to compare the eneloop 950mah batteries with those KAN 2/3AAA's... I also tried energizer 900mah AAA's with the same results.
Thanks for the info, hopefully they can provide a solution.
Are you using a level shifter with the WS2812 LEDs? Apparently using a switching regulator with controlling WS2812 LEDs can interfere with the WS2811 chips' ability to recognize 3.3V logic. I was getting a lot of glitches in my WS2812 displays when using a switching regulator. Adding a level shifter solved the problem.
You're correct about JonnyMac's code requiring certain clock speeds. There's code limiting the minimum speed to 80MHz. I think the fast clock speed is needed to generated the precise control pulses the WS2811 chip needs.
Edit: I see Xanadu asked about level shifters above. I didn't see an answer to this question. IMO, this is a very very important issue (possibly the most important in this application).
can you provide a little more info about the 'level shifter'.
I havent heard of them before and am probably not using one.
I just placed an order with digikey, if you can recommend a part asap i can get it added to my order before it ships
Thank you for the help