Is Peukert misunderstood?

[booster]

I tend to think so, at least for me. I had always taken at face value the normal statement about if you take power out of a battery faster, you get less total power out of it. I think I was just assuming the chemical reaction couldn't keep up with the rate at high discharges, so you were out of reaction earlier. Today, with all the talk of big alternators, large banks, fast discharges, etc, I started to wonder about what happens if you draw down the first half of battery fast and the second half slow. Do you still come up short because of the fast first half? According to my chemical reaction idea it would be likely the battery would recover a bit as the chemistry got a chance to stabilize at the lower output, so you might get the capacity back.

When I went looking for information, I remembered that Bogart Engineering had an explanation of why they don't use Peukert, so I tracked that down. It was very interesting to me and it blew up just about everything I has had assumed.

Bottom line, according to Bogart, is that the battery will always be able to give up exactly the same amount of power, regardless of discharge rate, as long as the end of the discharge is at a low rate. The explanation is so simple it is borderline silly.

http://bogartengineering.com/sites/default/files/docs/PeukertsComments.pdf

The only reason a battery gives up less power at a high discharge rate is because the definition of discharged is getting to 10.5 volts under the load being tested and the bigger the load, the more voltage is dropped because of that load. The battery isn't less full near the end of fast discharge, it just has more voltage drop during the entire discharge than if it was slow, so it reaches 10.5v easier. It actually be more full than one drained slowly.

This would totally explain why AGMs have much lower Peukert coefficient than wet cells. It is only because they drop less voltage under load, which has nothing to do with capacity at all. It is often said that lithium does not have a Peukert effect, which makes sense because they give up power even easier and drop even less voltage under load.

I think I now have a much better idea of how it works in the real world, and that way is to just ignore it. I had wondered if we ran the batteries to 50% of rated AH running something that was a big load, how much would we really have left? I had assumed that it would be under 50% of AH, but that is not right. It will be just as it appears. Just go be the number of AH you have used to figure where you are in discharge, and you will be correct no matter how you discharged the batteries.

Nice to know, at least for me

 

[avanti]

Thanks for the observation. I am a firm believer in ignoring anything that Nature will let me get away with.

 

[markopolo-ClassB]

Here's what a competitor says: SmartGauge Electronics - Peukert in depth - Advanced maths

 

[booster]

Here's what a competitor says: SmartGauge Electronics - Peukert in depth - Advanced maths

At a first look, I don't see anything wrong with what he says, or that it contradicts what Bogart said.

In all the equations and examples they are using a constant drawdown rate until they hit the rating empty point of 10.5v, I assume from what they say. And in that case, they will get less amp hours out of the battery to that point because the extra voltage drop makes 10.5v come sooner, when the battery is not empty. What they didn't address that Bogart talks about is that the "missing" amp hours are still in the battery after the high amp test, and if you went to a low rate, you could still get them out just as you would with a low rate test. A good test would be to do 20 hr rate drawdown and get the time to 10.5v and do it again at the 5 hr rate to 10.5v. Then look at how much voltage rebound you get once the load is off (or check specific gravity on wet cells). If Bogart is correct, the 20 hour rate test should rebound to a higher voltage, and thus remaining capacity.

Using Peukert in the real world of varying loads would be really tough because the same average load would give different run times based on when the high and low loads were run-like in big runs first or last. Your remaining run time would be determined by how big the load was, but even more so by how much voltage drop it caused and how close you were to 10.5v when you started.

Of course if you are using the battery in fixed load backup battery type situation, the information from Peukert would be great because you could calculate run time on the battery.

As I mentioned, I find this whole thing very interesting because Bogart's theory is so very simple, and lots of other folks make the whole thing very complicated, like all the equations in the competitor link. My inclination is that Bogart is the most correct, but I don't know if anyone has actually tested the idea or not. If Bogart is correct, our monitoring just boils down to how many amp hours we have used, regardless of how fast or slow they were used, and with a secondary concern being at the low end if the batteries are capable of holding voltage at the load we want to put on them. Wet cells would have trouble with the second part, I think, as they don't hold voltage under load well, AGMs probably can handle a pretty good load down to the 20% SOC most of us would consider absolute bottom. At 50% AGMs are probably fine for almost any load they can handle when full, at least for a while.

 

[markopolo-ClassB]

I get it.

Here's another link: Lee Hart on High Current Battery Use that makes the same point as Bogart Engineering but adding some capacity will have been lost due to heat at high discharge rates.

Peukert would matter if you need a specific current delivered for a specific amount of time. You would be able to appropriately size the battery bank by using the equation.

 

[booster]

I get it.

Here's another link: Lee Hart on High Current Battery Use that makes the same point as Bogart Engineering but adding some capacity will have been lost due to heat at high discharge rates.

Peukert would matter if you need a specific current delivered for a specific amount of time. You would be able to appropriately size the battery bank by using the equation.

Great link-good find. The tests he describes are just what I would do to prove, or disprove, so I tend to believe he is correct with it.

As you say certain load, certain time= Peukert counts, similar to my example of a backup battery running a fixed load.

Very nice to know that if the meter says we are down 100AH, there isn't anything else we need to know, or any need to adjust it.

 

[booster]

For those that like to play with numbers, this also brings up some other interesting scenarios, probably most applicable to folks that are limited in battery capacity and often use most of it.

All of the "rules" we hear about talk about % of capacity, particularly the "don't go below 50%" or " absolute minimum is 80%" discharge. What capacity are those rules applied to? Is it the AH that are determined by the 20 hour rating, or is it whatever capacity you get based on the size of you loads, particularly at the end of battery capacity, do to Peukert? Especially with wet cells, you could have significantly more AH available than the rating implied if your loads are significantly less than the 20 hour rating. Should you take that into account in determining your 20% absolute minimum, or is that built in to the manufactures recommendation? You would certainly be removing more energy from the battery at the lower rate scenario, so maybe it would be more damaging? The 50% point would also move based on load, but it would probably not be as big an influence as it is in the middle of the range of safe, not at an end.

On our new setup, we will have 440AH of AGm batteries, which seems to be getting pretty common based on what we are seeing folks doing here. To get to the 20 hour rating, we would have to be using 22 amps of power, but only at the end of the discharge cycle per this discussion, to get to 10.5 volts, or whatever other voltage chosen to limit discharge %. Our worst amperage might be 6-7 amps if the TV, Fantastic fan, frig, and lights are all running. If we are really in need, another 3-4 amps could go to charging laptops, tablets, phones, but we usually have that going on during the day when the solar is on, or we are driving. So 10 amps is really, really, high for us. If we have a low voltage cutoff, or do it manually off voltage, that would give us 80% discharge based on the 20 hour use rate of 22 amps, we would go way further discharged at our lower use rate as we would have less voltage drop do to load. The same would be true before we got to the cutoff, so we would seem to have more capacity left than we probably should use. Lifeline publishes a chart in their battery manual that is based on voltage to expect vs discharge rate based on % of 20 hr capacity. That chart now makes more sense and would allow a more accurate determination of your low voltage cutoff. Of course, if you set it to be safe at your normal usage amps, you would be much more likely to trip it with a big load like the microwave if the batteries were fairly low, but that might be a good reminder that you are nearing the end of your capacity and you shouldn't be using 100 amps.

I think when we get ours up and running, I will run the batteries down to 50% and turn the microwave on full power to see where the voltage runs. It may be very close to what the voltage would be at 20% SOC and our normal loads, and if it was I think I would set the inverter cutoff to that voltage. It would keep us from killing to much for the batteries at normal loads, and remind us not to run the micro without the engine running below 50%.

 

[booster]

I dug out a couple of charts on the Lifeline batteries to get a better handle on where to set the low voltage cutout, if you have one. I think most only have it in the inverter section, so for running on batteries, it would be an operator function for most of us when not running the inverter.

We have all seen lots of the SOC vs rested open circuit voltage. This is Lifeline's version.

This is our favorite, but basically pretty useless, chart, as it doesn't account for load and who has 4 hours to have the batteries sit unused, just to check SOC.

Here is the similar chart from Lifeline that takes into account load, so you can read your current load of an ammeter, or you may know based on what is on, and your voltage, and then know what your SOC is with everything still on. Based on the previous discussion about Peukert, I think this will be accurate no matter how you got to the point your testing, or if the load is your typical load or not.

Here is basically the same information on a graph so it is easier to determine areas between the data points on the chart.

The chart and the graph don't match perfectly, but for our typical use would be around the 75 hour rate, so somewhere in the 11.70 volt range would be what we would see with that load running if we were going to 20% SOC at our typical load. This is all really the same thing as the old "light bulb method" of checking battery SOC by putting on a fixed load to eliminate the rest time, except this is good for all loads, and I didn't have to build the chart by testing (which would take a long, long, time with the 4 hour waits).

The next question is how low could the batteries be and still be able to run the microwave, which pulls about 100 amps or about 4.5 hour rate. If we used the same 11.70 volts, the chart would say we could not run the microwave below about 55-60% depth of discharge, unless we had the van engine on. That DOD would allow our rule of 50% SOC or higher to use batteries only for the microwave to be just fine with the low voltage cutoff set at 11.70 volts.

For us, this is not really much of an issue, as we will base most of what we do on the amp hour counter, but it is an interesting exercise in finding out where to set the low voltage cutout in the inverter to be a reminder of where you really are in the SOC curve, which could be very useful if you don't have a battery monitor, and want to make sure you have enough power left to make it through the night for instance.

I can't say how these charts would be for other AGMs, but may be close. They would be way off for wet cells, as the voltage drops way down compared to AGMs. I was really surprised at the only .4 to .3 volt difference in voltage, at the same SOC with the vastly different loads.

 

[Boxster1971]

Thanks booster - great insight on how to use my upgraded battery pack.

 

[Mike-ClassB]

AGMs rock. Nuff said.