Hi I Thought: I'd do a video looking at how much energy is left in Double-a and Triple-a batteries under Naught Point 8 volts. I Mentioned this in a previous live video because I think there's much, if any real data out there on just that energy under Naught Point 8 volts. Because as I've mentioned many times in many videos over the years, the manufacturers data sheets their characteristic discharge curves a stop at Point a volts. It's like the industry standard cutoff voltage level where everyone pretty much agrees that there's basically bugger-all energy left in these batteries once they hit and point Eight volts.

And that's true in most cases, of course. But of course, with circuits like the Joule Faith and other really ultra-low power boost converters, there is actually some energy under there. But how much so the data sheets are typically stop it. You know, like a hundred milli watts or something like that.

So I thought like, what for really low power drains? Is there anything in there? Is there anything usable? and really, it's I Know there's something usable for some applications, but how much as a percentage of the total battery at very low discharge levels. But you know things like products that work for a year or something like that on a set of batteries? Is it worthwhile actually designing your products to go less than point eight volts down? To say 0.5 volts or 0.6 volts is like a typical figure. So I thought we'd take a look at it because it's hard to get data on this sort of stuff. so this is gonna be some really long-term testing.

but I just wanted to do this part one even though we won't get real, you know any real results here. We need to do long-term stuff. Just show you the setup that I'm going to use and some just an initial overnight test I just did here I said that last night and just did a quick test to make sure all my setups working. I'll show you exactly what I'm doing here and then I'll get some.

I'll start out with some triple-a batteries and then I'll get some double-a batteries and only one brand at the moment yourself. Maybe I can do more brands and compare them things like that. but really? I wanted to discharge them at get multiple characteristic curves that maybe a hundred milli watts, 50 milli watts. you know, 10 milliwatts things like that.

really low power levels. So I'll show you the setup, well have a look. Let's go now. I Started out just wanting to use my beak a Precision 8500 electronic load here and I've done a previous video on this where I've done some discharge testing of lithium polymer batteries.

so click here if you want to see that video. So this actually has a battery test mode building. You might be able to see battery but just under there and it's got some real crappy software that comes with it that also does our characteristic discharge art curves of batteries and that's all fantastic. But unfortunately the battery discharge test mode on this thing only supports constant current and I wanted to do constant power just to make things easier.

and you know, maybe a bit more valid for modern products that use DC to DC converters for example. So I could have don't use the test mode and just use constant current. But yeah, I really wanted a constant power type thing and this, of course does constant power. I can enter in 100 milliwatt so I can go to 1 milli watts resolution I think it's actually 0.1 milli watts resolution.

Not sure if it's actually capable of that, but it at least allows me to enter that anyway. Still, haven't you know? Actually check the performance verification of this thing right down at the low levels. But anyway, I can set constant power discharge so it'll just act as a constant power load all the way down to hopefully zero, which we'll take a look at in that today's video. But unfortunately, that stupid software.

The only way you can log data out of this thing. It's got an Rs-232 port on the back and I've got the isolated cable for it and everything. The software that comes with it does not allow you to do constant power either data logging or battery discharge testing. It's ridiculous.

and then even if you did, the software is so buggy that doesn't even allow you to export the data to an Excel file. It's got that capability. or a CSV file. It's got that capability.

but it doesn't work. it's just raw. I Hate horrible software I'm bloody. This isn't Otherwise really good product, but how the software is just really is.

David - he said oh yeah, look I'll just you know it should be fairly easy I'll just write some software for it. So yeah, I said. you know it's only our it's Rs-232 interface. It's probably like a serial command structure.

You know. Center to command like you know, please read voltage or set constant power mode all in ASCII text and things like that and within like five minutes you realize mwah, nope, it. it doesn't use regular ASCII You know text serial commands. It's actually got a 26 by 2 I think or 24 byte packet structure with check sums and all sorts of commands.

and it's all in hex. And it's horrible for an Rs-232 interface. Why it can't just use bloody serious strings commands? Anyway, so he abandoned that and yeah, I've just got my own interface now using my HP bench meter. Now because I Want to use constant power work mode.

We can just use this as a constant power load and the only thing we need to measure is the voltage across the battery. Hence why we only need a multimeter to actually log the data. We don't have to log the current because it doesn't matter as long as we are confident in this load that it gives constant power all the way down to zero or whatever cutoff voltage we want to use. it'll be zero in this case.

Then you know everything's hunky-dory We only need to measure the voltage. Perfect. and yet you can use like a handheld data logging multimeter for example. But often because this some of this testing could take a week or more, you know it may be.

you know, a thousand hours or something or things like that. So the actual if you use the battery power of Bodie meter, it would actually run out of battery before you could actually test this single triple-a or double-a battery at very low power levels. So of course you use a proper bench multimeter. I'm using Maya Very nice.

Three, double four, seven, zero a seven half digit meter here which actually has a data logging mode. So we'd actually go in there and set up the interval time and everything like that. Absolutely perfect. Now here's the setup.

Just got a triple-a battery holder. also have one for way as well. Now the important thing to do is actually tap the voltage directly off the pins itself. Um, it doesn't necessarily matter so much at these ridiculously low currents because these huge beefy wires, it's at these currents and it's not gonna drop anything, right? But hey, you do it properly.

So I've tapped off out there and they've just got some pins going into these heavy-duty connections into there. so if you wanted to do you know, large currents your that'd be really important. So that's all we need to do. and then we can go into here and dialogue and then start it and then start our load here.

Now I've actually got this data overnight. I just used like a Fuji battery here. it's just when I had lying around and already been discharged a little bit. so this isn't a real test.

I just wanted to test that the whole thing ran overnight and I got the data out and I could graph it and everything else just to test the yup methodology. So I'll take a look at that in a second. and I've been discharging overnight at a hundred milli watts here and there is a hundred milli watt curve in the duracell datasheet as well, so we can actually compare that. So if we if we go out of there, we can see that it's doing a weird thing here.

it's like jumping up and down in voltage. It like this is like fully discharged to zero overnight. So it's doing something weird here. so that's most likely like the constant power mode trying to compensate and it sort of looks like it's oscillating or doing something weird like that.

If we actually switch it off, we'll see the voltage in the battery actually start to recover a little bit. look. So I've actually had that effectively pretty much shorted overnight with the load. Anyway, a relative constant power load of 100 milliwatts and you see it's ramping back up.

But of course, if we there's no energy when you remove the load, you're not really recovering the energy in the battery. Yes, the voltage goes back up due to the chemistry, the internal ESR and other you know, complex things that are happening inside the battery. But you're not really recovering any energy. So if we put that load back on that hundred milli watt load, it's just gonna drop instantly right down to zero.

There's just nothing there. But if we set that, if we did a constant current for example of say, 10 milliamps 0.01 Okay, so put that in there and we'll switch on constant current mode. We'll actually see it. We'll be able to deliver us some energy because constant powers are different.

The thing that's got to do math and then compensate constant current is probably used as a different thing internally. So if we switch that on, you can see that we can't actually get. you know, I Continuous 10 milliamps out of this still flat battery at point. Five six volts and it's going to drop and drop and drop.

We can actually maybe increase that. see if we can get can we get 50 milliamps out of the thing, but it's very low current. Remember that if you want the power, just multiply it. There we go.

It's drop-in so we can still get 50 milliamps out of it at point 4 volts. but it's slowly. there's not much energy left there at all why it's gone. And of course, the BK precision stood showing that it's actually still delivering 20 milliamps there at no voltage at Bugger-all Oh no, look, it just vanished.

So yeah, there's like something really at the low end with this thing, so we need to check that. But anyway, let's go have a look at the data. So what? I actually had running overnight here I Had the I'm using the data logging mode Here I am at the sample interval 60 seconds. So I'm taking one sample every 60 seconds and that one sample is like just your regular DC volts mode.

I've got a fixed 10 volt range here. 10 power line cycles just to do that's. just standard to do a little bit of averaging to get a reasonably stable reading. and so it's taken one sample every 60 seconds and you can actually a duration I put in I'm doing number of readings so I calculated that our 2,000 because unfortunately that's the problem with this method.

Using the multimeter and as opposed to the one with using this allows us to program. If we just use the electronic load, we could program in a cut-off voltage where it would actually stop logging at. whereas this one with just the multimeter doesn't allow us to do that, so you have to do a little bit of math up front. No, we know what the capacity here batter is, know how many readings you're going to want or you can do it in that duration as well.

But I decided to just calculate the based on sample interview interval and the number of samples roughly. you know how long I'd need overnight and then I doubled it or something. I even figured I'd need like a thousand samples or something and there was no no delay. We just started it straight away and then I logged it to an external weigh USB key on the thing.

So into a CSV file which we cannot take a look at. So all you need to do to get to do this thing is to what start your data. log in here, just start it and then turn on. You load at whatever power load that you want and bingo You just leave it there for a day a week.

However long it takes to just discharge a thing and you've got the CSV data on this stick. Let's go take a look at it. so please excuse the crudity of this data and graph. I Didn't have time to build it to scale or to paint it.

Now the CSV export in from the Agilent meter is a keysight meter sorry is actually quite good and there we go. It's very simple. the reading number here and then all the readings there in voltage. If you didn't fix it, to say the ten point, the 10 volt range, for example, it might switch to like exponent mode.

So instead of doing you know naught point. if we go right down here, you know no point. Oh, I eat and things like that, it would, you know, give you e to the minus 2 and stuff like that. So anyway, here's our graph: I Just graphed that very quickly and you can actually see it look drop off like an practically brick wall response at that naught point.

Eight volts? Anything Actually, you know, under one volt. you can argue there's not much there. So we're actually discharging this. Remember, at a hundred milli watts.

So not a huge amount of our power. In fact, the the Duracell datasheet I mentioned before, it only has a characteristic curve to 250 milli watts for constant power. The Energizer one does I'll show you that in a second. But look, it's just like you know, even below a volt, there's not much left, let alone not eight.

And if we go and actually have a look at the data in here and actually see our individual data points, even though we've got 642 data points by the time it drops out. Look, we've only got one data point there under naught point eight volts. So that was with that sixty second sample in a interval which we actually got up here. So you know, really, we actually have to sample a lot more data than that if we wanted increase the resolution on this.

Drop off here. Having one sample per minute just didn't cut the mustard there. So you actually see there that even if you design your product that drew 100 milliwatts to work down to naught Point: 5 volts. you're so proud of yourself.

Yeah, it's extracting every last drop of energy in there. It's only this is a one-minute time interval. You're only gonna get maybe an extra minute or two tops out of the thing before it dies. You went to all that trouble and expense and you might have had to use it much more.

Expensive Materials Cost: decent DC Converter: you might have to change it. Architecture: I Don't know. Whatever. Do work down 2.5 volts and you're just.

you're just wasting your time at a hundred milli volt product discharged. Not worth it? Yes, of course I Could actually speed up all this testing by like you know, really heavily pre discharging the battery down to a vault under load, for example, and then just get you know, the fine data right at the end. But there's you know, technically that's not the best way to do it, and you know people might complain that you know the proper scientific way to do it is. hey, we've got a product.

It draws a constant 100 milli watts or whatever you know over it's over the life of its product. by the time you put it in. Let's you know, not talk about the you know efficiency curves, the DC to DC converters, and things like that As the battery voltage drops, it can change Anyway, let's assume that a product draws a constant power all the time, which is a product design with the DC to DC converter already in it and it, you know. So we want to get some real data, a real characteristic curve of what it looks like.

Even though this series of videos I'm going to do I'm only worried I'm only interested in this little tiny little part under Naught Point 8 volts here. and clearly 100 milliwatts. We're getting bugger-all data, and even if we actually did sample it once per second instead of once per 60-second and we might have been able to get you know, an extra, you know, half a dozen sample points in there. Let's look at the area.

it's just nothing compared to the bulk of the rest of it's not even 1% its bugger-all But the idea of these videos is that we can go down to really low powers. you know, 50 milliwatts, 10 milli watts, you know, maybe even lower and see what the characteristic discharge curves look like. In theory, they should be a little bit better than that. They might like extend out and drop off more gently.

for example, under Naught Point 8 volts, but hey, that's but it may not. It may still drop off like a brick wall, so only one way to find out. Actually do the long term testing. Now you could actually see the weirdness happen in here right at the end of the like, after its fully discharge that discharged set at zero for a while and then maybe due to the battery chemistry at the site at all.

You know I'm gonna start doing something weird. and then the constant power mode as you saw before when you could actually see the readings just jumping around, it's doing some sort of oscillation. They're doing something weird right down at that level. So I actually now just want to do a quick test on the 8500 electronic load and see what its performer.

It's constant power. performance is like right down at low voltages. Let's check it out so this is a real easy test to do. We get our Reigle watt power supply here.

like a nice precision power supply that can go all the way down to zero volts and Oodles power supplies can, but a good bench laboratory power supply can go down to zero volts. And by the way, this isn't like a point. Oh five percent our class instruments so is the RAI goal here. And of course the Agilent 7 RF digit meter high is the Ducks guts.

so you know, no problems with any sorts of our precision in our measurement. But and you know it doesn't matter. we're looking at the characteristic discharge curve. We could you know 1% absolute accuracy Now voltage measurements would be fine, you know, you know, and 1% in our power for example would be you know a constant power load.

It'd be just fine. In fact, I'll check the data sheet on the and we'll have a look at the data sheet for the electronic load see how accurate is in constant power mode. I Think it's going to be less than what it is in constant current mode and things like that say cuz it has to do some math so you've got some additional errors in introduced there Anyway, so we've got our one. We've got it hooked up here.

We remove our battery. of course you don't, you're better in there and we just plug the power supply straight in. Easy-peasy So we're outputting a volt at the moment it's on. So we've got one volt and of course we're measuring a volt on here.

but I haven't turned on the load yet. We're also measuring a volt. We don't need this. We can just rely on pay for these anyway.

So it's going to constant power mode. Our naught point One: What's a 100 milliwatts Which were doing before? So we switched that on okay and we now join. There we go. 100 milli watts at least.

significant digit? Okay, right? They basically totally agree. Okay, but that's at 1 volt. Now if we start winding our wick down and Ola zoom in for this one just so you can see it a bit better. Okay, so if we wind our wick down, I've got my cursor on the second decimal place there, so can point 99 and we can see it drop.

Ok, now what we're looking for now is that this stays at a constant hundred milli watts. Okay, near enough to 100 milliwatts. Once again, we don't care if it's in a few least significant digits out and you'll notice that our current is going to increase because voltage times current the equals power. So if our voltage drops, our battery voltage drops, then our current must increase to compensate.

And that's one of the disadvantages of DC to DC converters down at low voltage. As I explained in previous live video, it's like a snowball In effect, the lower the voltage you get, and you'll notice that it's um, it's still hanging in there at 100 milliwatts. So we're just testing the ability this 8500 electronic load here to actually still maintain constant power performance right down at low voltages. So it still gets dropping and it's still maintaining our hundred milli watts there.

No problems whatsoever. So yep, down at point Six. Still not a problem at point five volts which is really what we care about I Guess it's still fine. So yep, this is load as more than suitable.

but we go down right now. No, it's still 100 milliwatts. Let's keep going. Might bore you to death point 2 volts.

It's still a hundred. Oh, there we go. We're getting some error. So down it.

Let's say at point two volts here. So yeah, yeah, or is that yeah? Maybe because we're looking at half an amp? Look, we try to draw half and a half at point once and Bob that's a lot of current, so you've got a, you know, like we might. This is where the drop in these leads could be quite significant, but it's not because look, we're getting still getting seven. This is where our A Julep meter up here can still help because we're actually tapping that point there.

but these are really thick beefy cables. If they weren't these were just really thin wires. Half an amp would get significant voltage drop across there and this wouldn't be accurate as you can see. No problems whatsoever.

But we're still like we're ten percent out there. Now here's actually where we get to the limitation of our test setup. Yeah, we might be shown an error here, but that's not actually an error caused by this low. We're now getting the error caused by the voltage drop here because look, multiply 0.15 volts times 0.6 Six amps you get basically precisely the hundred milli watts which we've programmed into the thing.

So this load is still doing the business down at point One, Five down at point one. It's hard to the resolution of this thing there. We go exactly one amp there, right? So that's it's still exactly 100 milli watts. So this thing is actually still working absolutely perfectly.

even though we're getting like point like forty percent error over here. That's due to even though we're using huge big thick beefy cables here, these things are monsters. Big binding post terminals. this wire is really thick as well and these massive binding posts.

We're still getting that drop on there because we're at an amp. It's amazing. The drop. You can get a nap and especially in this sort of situation that can cause a 40% error.

but this sucker still works perfectly fine so it's no problems whatsoever. So the hiccup in we must be seen in this thing must be like the battery chemistry. like you know it, causing an effect where okay it draws would be the current. Look, the voltage drops were it rises back and this thing is trying to has it has a software loop in it which calculates the power and then effects the load which backs off the load a bit and then our the battery come Rises back up in voltage a bit and then drops back down.

a bump bump bump. So then it just it just hits the bottom and this thing just starts bloody well. oscillating battery oscillation beauty. But hey, I'll just check another power mode I'll do 10 milli watts now.

okay and just see how good it is in 10 milliwatts. There we go, we're down at 10. Our resolution is right down there. You know we can.

you know, get better instruments to measure this more and there's better techniques to measure this more accurately. But yeah, that's good enough. Once again, yeah. 10 milli watts down 2.2 4.2 1.15 volts.

We're still getting our 10 milli watts near enough, so I'm happy with that setup. It passes the test. so let's go and measure some real batteries. Sorry I won't be able to give you the data today.

This is gonna take a long time. We have to do it in part two. But anyway, I've got my sauce a brand new our jurist I'll copy Top Guru Lak ones just got them from the local shopping centre. They're factory fresh and as you can see, the Duracell one I only use constant power down to 250 milli watts here.

It doesn't have anything better than that. but you know, if we jump over to an equivalent energizer triple-a battery, it does actually have a characteristic curve down 500 milli watts. which is what I just did overnight on that Fuji battery. And yep, it's a similar sort of our you know, 11 12 hours something like that to discharge.

So I will do now. 100 milli watts. we'll start. that is our baseline.

so I'll discharge one overnight. I'll get the data tomorrow. Then I'll do another one that maybe out of a 50, then the one at 10:00 and have a look at the data and then I might do them. you know, in between that or whatever.

So yeah, we'll give it a ball. So how many data points do we have to do? well if we go back to our graph that we had before as I said, like we just didn't have the resolution down here, only had that single data point. So we want you know like we want better than that. this is once per minute.

so let's do it once per second. Okay, let's really go to town and this took out 642 samples at once per minute. So let's just say we had a thousand samples because this wasn't a full battery. Let's say we've got a full one.

Maybe you know it's a thousand, Should be A thousand samples should be plenty at once per minute. Well if we want once per second, well let's set our a julep meter to say 60,000 samples at once per second. All right. So let's do this.

We got the Giroux sort of do a lot thing guaranteed for 10 years in storage and don't leak. Yeah, right. Anyway, they were manufactured 10th month 2014 so they're pretty fresh. That's the freshest one.

I can find a local supermarket so we'll laughs whack that open. So let's whack that in there. and tada one point 600 volts. We've got our power set to a hundred milli watts.

Everything's hunky-dory We're ready to go and our we're got data logging once per second here. 60,000 art readings should be plenty I Can stop it I don't have to go to 60,000 I can just press stop there. No problems whatsoever so we can sample, interview, data, log. everything's everything's fine.

I've got my manual our 10 volt range. everything's ready to go right Here we go. So we want to actually start our data log in. There we go: one point 601 volts.

It's fine. Okay, you want to start at first before you start your load. We can always chop out a couple of samples in our data set and here we go: Tada Bingo! One point Six three and we'll just leave it running. And here's where: I Nice Seven, half digital, six-and-a-half digit or even five enough digit meter comes in handy.

You can actually even at low discharge levels. this is a razor I Mean this is 64 milliamps. You know? it's a reasonable amount of discharge 100 milliwatts. but even at low levels with a a high-resolution multimeter, you can actually see the drop in the thing so you can just see it slowly counting down there.

And if we went down to ten millivolts, we'd see it be. you know, like one tenth of that speed for example. So yeah, very handy to have a high-resolution meter like this. So there you go I Hope you found this useful, even though it was just setting up a test.

Rikka like this. There's actually you know a lot to it if you haven't done this before. Hopefully that starts some useful info in there for doing a battery discharge testing like this. So yeah, I'll leave this run overnight, get the hundred milli watts, then I'll probably do 50, then I'll do 10 and you know, have a look at the data, analyze it, and then decide if I want to do more or less.

and then of course this is only for the triple. A's I'll also do the double-a ones as well if people want me to do different brands, but there's nothing you know. this is not a brain comparison thing. This is a test to see if there's any useful data and 0.8 volts and well, I think we're gonna see much.

You certainly saw Bugger-all at 100 milliwatts. it was like Ii and I'm not even point 1 percent. it was half a B's dick. Anyway, hope you liked the video.

If you did, please give it a big thumbs up because that always helps. A lot. Comments and all that stuff down below links to everything. If you want the shirt, I'll probably link in the Teespring store where I where I crowdfund.

It's sort of like a Yep, it's not really a crowdfunding thing, but anyway, you can buy the shirt. it'll link in down below somewhere. Catch you next time you you.