True 80% DOD or are we just guessing

[ccmdr]

Sunday night syndrome kicked in and what crossed my mind 'Cycle Life' and getting that real world Ah usage from a battery/pack.

There is alot of information on 80% 'Depth Of Discharge' (DOD) vs extending cycle life, but how are you measuring yours?

I'm no expert and could get some of this wrong, so please correct me if so .

Now for the question. This is for all cells, but I'll use my pack for a real world relation.

Say you have a 20Ah LiFePO4 pack and the manufacture recommended spec is:
Max charge Voltage 3.65v
Min Voltage 2.5v

So they 'claim' from the battery being fully charged to discharge it will output 20Ah at a specific Discharge rating, for this example 1C. We now want to extend the service life of our expensive batteries and have seen 80% DOD can extend Cycle Life. But looking at a standard Lithium discharge gragh we all know about the 'Cliff' where going below a certain voltage gives you diminishing returns at a sacrifice of cycle life.

The standard smooth Discharge graphs give a representation of voltage and amps used over time. Unlike some of the sawtooth gratphs when the cell is loaded and unloaded.

Background covered, if you were to discharge at higher C rates loaded vdroop becomes more apparent. This is where I'm getting lost, wanting 80% DOD your minimum voltage would be around the 2.7-3v, however this would have to be a 'loaded' voltage due to the resting voltage bounce-back. Running 2-3C the discharge rates the voltage drops below your 80% DOD voltage of 2.7-3v.

80% DOD is can measured by loaded voltage, but depending on discharge rate you can exceed it in a very short amount of time. 80% can be 16Ah, but realistically your taking from the top and bottom from the discharge curve so you'd lose more capacity from the top. Now take into consideration temperature and the vdroop involved with 10C or 0C ambient. Even on a mild discharge of 2C you'd be close to exceeding 80%.

This must mean anyone running a mild C discharge rating is already exceeding 80% DOD right?

With the Adaptto BMS you can set the voltage exactly for 80% DOD, but I find it frequently cuts out due to vdroop under load, so I can never truely keep the battery cycling at 80% DOD. Anyone else experience this?

 

[dogman dan]

As I understand it, it's unloaded voltage. But yeah, when you are using a battery that sags a lot, whether because they suck, or because you are hitting a good cell very hard, it will show a very low voltage under load. Pretty scary, if you are used to an idea that this pack should be over the cliff at 49v, and you are seeing 44v under load. Riding with a bms, you must have a better match of your battery and motor. Cant' have too much sag, or your bms will shut off premature all the damn time.

Then on top of that, whatever sag you have been seeing so far will increase as you near the cliff. So if you have this mental idea that sag should be 4v under load, and now it's 6v, you can get pretty worried. I like to ride in such a way that the sag you get as you near 100% discharged is not more than before. Go easy, once you are on the edge of that cliff.

So what do you do to accurately measure % DOD? Well, if you really want to stop at 80%, you need to stop guessing. ( the bms can only guess) Stop going by voltage and go by a watt meter. Use the watt meter to find out what your real world capacity is, or close to it. Figure your battery still has 4 or 5% after the cliff. So find the cliff to start with. The cliff you can find by voltage only. Ok, you have a "20 ah" But when you find the cliff, you have measured 19. Ok, call that 95% discharged, and now simple math can tell you that 80% dod on that pack should be about 14.25 ah. Stop at 15 ah close enough. 80% is not some magic number where all hell breaks loose if you discharge 81%. In fact, you can just guess at your capacity, call it 15 ah to 80%, and be plenty close enough.

So for your pack, if it has 19ah real world, just keep it down to 15 ah discharged as much as you can. But don't be an idiot, and push the bike home the last mile because you'd discharge it 19 ah to do it. If you do need to use that last bit, slow way down so the c rate is less when you use that last 20%.

Personally, I don't worry about 80%. If I need 100% I use it. But I do plan my rides, so 99% of rides do end up finishing well under 80% DOD. It's a matter of owning enough battery if you take long rides. I like to buy one battery at least 10 ah every year, so at any given time I tend to have 48v 30 ah on hand. 30 ah is enough for some very long rides, and if I need it all, I use it. But like I said, 99% of the time I can just pop a second battery on the bike and have more than plenty.

 

[ccmdr]

Cheers for the input Dogman Dan ,

It's just seems odd to measure DOD from unloaded voltage. In my minds eye I see low voltage/higher internal IR as the instigator of low cycle life. Maybe I've been overly skeptical of my pack, but for example I've just hit 36km at avg 56.7kph with a graded A123 26S pack. After settling the unloaded voltage is still 85v (cell avg 3.26v), however under 3C load two lazy cells drop to 2.7v the rest at 2.9-3v, but full pack voltage under load is 78-79v. Am I then to presume the battery is not even close to being fully discharged being 3.3v is nominal, which in my understanding is the avg voltage over the 20Ah discharge curve.

I've tried reading the watts, but my Adaptto on regen instantly tricks itself into thinking the battery has fully recharged...

Also, if internal resistance is the instigator of low cycle life are we fooling ourselves by first measuring Watts and low voltage limiting the battery, assuming you can program your BMS/LLV.

Pretty scary, if you are used to an idea that this pack should be over the cliff at 49v, and you are seeing 44v under load.

If you measured unloaded voltage it would 'appear' you've not 'cycled' your cells at a high IR right? This is why atm I presume that you need to measure loaded voltage, it would show your running higher IR over time. As the voltage drops, the IR goes up and heats your battery. Higher IR = Low cycle life right?

Depending where you take your 80% DOD from 5% from the top and 15% from the bottom vs 10% top 10% bottom. Wouldn't a watt count give an untrue guesstimation of IR, where voltage would give a more linear guess ie. HS3540 Constant speed 30 mph = 1000W@Motor which = Nominal 85v@15A or Minimum 65v@20? Both are give the same power output over time, but the Ah used would vary and the cell at 2.5V 65v would have higher IR than 3.3V nominal at 15A?

With that in mind surely running any discharge rate thats causing a vdroop below your 80% DOD minimum voltage be infact cutting into your assumption of cycling a battery at 80%.

For example:

LiPo = 4.2v Max 3.0v Min a 20% voltage range isolation for 80% DOD = 0.24v
Most LiPo users 80% DOD = 4.2v-0.1v = 4.1v Charge Max so vMin should be 3.14v

LiFeP04 = 3.65V Max 2.5v Min a 20% voltage range isolation for 80% DOD = 0.23
Most LiFe users 3.65v Charge max so vMin should be 2.73v

I'm not a LiPo user so I can't give real world experience, but I know my LiFeP04 vdroop at 1C = 83v HOC and 2.5C is 79V around 3v per cell HOC. But this is for frequent <15degC ambient usage. After looking here as a rough guide for Temp vs discharge rating anything much over 2C is would hit 80% DOD. People with larger Ah packs and a lower C running rate are going to get a much better cycle life, however we would have to adjust the DOD settings based on pack temperature aswell ie. it's 3degC atm so I 'know' the pack at 80% DOD wil be around 60% Ah at 2C discharge, but isn't there a more measurable value, getting the most out of the pack all the time without overstepping capacity requirements and DOD. Or would we have to individually put in our battery parameters at a given temp or C rating? Surely if we could get an IR measurment of our personal battery pack per cell with respect to Wh and impliment that into a 'capacity gauge' this would solve not only our 80% DOD guesstimations we would have a valid function of measuring actual capacity depending on temperature and discharge ratings?

For example, (these are just figures plucked from thin air btw):
6mOhm could represent 100% charged
12mohm could represent 20% charged
0degC would increase IR to say 8mOhm off the bat so it'd calc the remaining 4mOhm range vs current usage over time in seconds
High C rating would increase mOhm usage over time etc..
Catch my drift?

I really need a bench test at home so work this out myself...

Sorry if I'm repeating myself in places or it's sounds like aggro, it's not intended I can't get my head round it atm :lol:. Sometimes when something is staring you in the face you can't see it until someone points it out, obvious or not. I'd still rather see it than be in the dark guessing .

 

[dogman dan]

I just remember being told that when figuring time to stop based on absolute minimum voltage for a cell, it's the resting voltage that matters.

However, I was just posting on another thread, that you must slow down and lower your c rate as you milk out that last 5% of capacity, or it will cause the cells to get very hot. This is because the resistance increased compared to the start of the discharge. You can get more capacity out of the pack, if you desperately need it by lowering the rate, and common sense says a cool pack is not getting damaged while a very hot pack is.

But you are talking when to stop at 80% discharged. All my experience says that at 80% dod, you aren't yet at that point where full throttle will cook your cells. And again, I see no reason to think 81 or 82% discharged will be much worse than 80%.

All you need to do to go by voltage and be close enough, is figure out what your sag is under full throttle. Then set your bms to the voltage that leaves you approximately 80% dod resting, when under full load. So when it pops under load, the pack voltage is not still so high resting. Otherwise, you risk having the thing shut off when it's really only 65%. You won't see such a big voltage difference between 60% and 80%. So it gets more and more a guess till the pack is more like 90% done and the voltage spread is wider. You will be able to get more than 80% out of it, but only by riding slower and milking it out. If in doubt, give a throttle blip and see if it shuts off. Now you know.

See how this will work? If you give the pack a voltage spike when its lower, it will stop. But you can still limp along slow, and still stop before it's going over the cliff.

If you need more range than you get with your current bms setting, then carry more battery. Like a small emergency pack to get home on. Or just say, oh well, and just keep out of that last 5%. Remember, we are not talking about trying to maximize the life of a crazy expensive car battery here. You just have a bike size pack, relatively cheap to replace every three years.

 

[ccmdr]

Cheers Dogman Dan, eTrike,

This is what confused me, two different people stating different things as per normal .

Table 2 on here
vs
How To Determine Battery State Of Charge

*Edit - Found a reliable source for proof of Dogman Dan's statment: Depth of Discharge

It even gives a rough Table to take into account High Discharge rates and Colder environments

Next question: are discharge Graphs shown under load or at the resting voltage, I'm guessing now they are resting voltages because if they were loaded you would be unable to read the bounce back voltage where you measure dischargeed capacity?

 

[ccmdr]

eTrike, really thank you for taking your time to explain this.

I had been under the impression that a battery was a simple cell structure and being SLA or Lithium the same rules would apply. It seems like desulfators are much like anti-limescale for your household plumbing, but for batteries to regain a greater surface area of anode/cathode?

With regard to the M1A are they a similar chemistry to the 20Ah prismatics? I'll try a few bouts of 3.45v charged to 2.5v loaded and see what the rough capacity is. I just find it interesting that the discharge graphs are loaded SOC when most people assume it's a unloaded measurment. Given that the only information from manufacture is loaded SOC graphs is everyone going through the lengths of mapping their capacity of just going, 'yeah my unloaded voltage is 3v i'm fine to push it to 2.5v' when they've used the only source of capacity mapping the manufacture provides in the loaded discharge graph ?

I can sort the Adaptto issue by simply disabling regen, but on regen the voltage peaks over the nominal and resets everything. I'm sure it's just a setting I'm missing, time and patience and so many options with the Adaptto product.

 

[justin_le]

Say you have a 20Ah LiFePO4 pack and the manufacture recommended spec is:
So they 'claim' from the battery being fully charged to discharge it will output 20Ah at a specific Discharge rating, for this example 1C. We now want to extend the service life of our expensive batteries and have seen 80% DOD can extend Cycle Life.

Now before going any further here, can you clarify where exactly you saw that limiting the depth of discharge to 80% on a LiFePO4 chemistry cell is going to extend its service life? It might well be that your entire premise is based on someone mistakenly taking conclusions that apply to one chemistry and generalizing that iron phosphate without really checking to see that it holds. At least from what I've seen all the example tests showing LiFePO4 having exceptional cycle life were done with full discharge cycles, to 2.5 or 2.7 V/cell or whatever it may be, and didn't give any suggestion that limiting the DOD to 80% would have made it even higher.

If you or someone can point to an actual cycle life study on LiFePO4 that shows an improved longevity benefit by never even approaching the knee of the discharge curve, then that would be useful data on which you could tweak your system. But until then, it could be that you're simply putting effort into just reducing your range per charge with no tangible benefit.

 

[justin_le]

OK here's one for you, based on tests of A123 26650 cells:
https://www.researchgate.net/publication/251588109_Cycle-life_model_for_graphite-LiFePO_4_cells

Section 3.1 of this report in particular is of relevance:

When plotted in this manner capacity fade appears to be a function of DOD. However,when the same data is plotted as a function of time as shown in Fig.4, the results indicate DOD has very little effect on capacity fade. The capacity fade rate was found to be approximately the same at each DOD indicating that the effect of cycling time is more significant than DOD. After a closer examination of all the DOD data at C/2 rate, we concluded that the DOD effect was not important for the conditions investigated.

 

[ccmdr]

This is interesting, I originally just intended this the be BMS issue limiting a full true 80% DOD in real world environments, not the 23degC standard benchmark manufactures use. Are you guys going on about DOD in relation to cycle life, in the sense of:

80% DOD giving 1000 cycles
100% DOD giving 800 cycles

80% appears at first glance better, but total W discharged over that same period be exactly the same?

Justin, my asumption that 80% DOD was valid through out all lithium chemistry were based on reading a combination of DOD vs Cyclelife with the rough understanding of the term lithium ion

 

[justin_le]

This is interesting, I originally just intended this the be BMS issue limiting a full true 80% DOD in real world environments, not the 23degC standard benchmark manufactures use. Are you guys going on about DOD in relation to cycle life, in the sense of:
80% DOD giving 1000 cycles
100% DOD giving 800 cycles
80% appears at first glance better, but total W discharged over that same period be exactly the same?

That's correct. If you look at the data results from the research paper on A123 cells that I linked, you can see that cell capacity vs cycle life at different DOD's in figure 3


At a 50% DOD, you would expect to have twice the number of "cycles" as 100% DOD if the total battery output life is more of less independent of the DOD, and that's exactly what this shows. Similarly, a 20% DOD results in ~4 times the cycle life that you get with an 80% discharge, if you compare the red circles to the blue triangles. Same total energy, makes no difference to your total battery life how far you discharge it.

Justin, my asumption that 80% DOD was valid through out all lithium chemistry were based on reading a combination of DOD vs Cyclelife with the rough understanding of the term lithium ion

Well, as you can see it's a completely false assumption. The first link is freely alternating between SLA and Lithium as though those two were of the same, even sometimes stating as much with this hilariously false quote:

The above graph was constructed for a Lead acid battery, but with different scaling factors, it is typical for all cell chemistries including Lithium-ion.

With lead acid batteries, you get vastly more total power output from the cell when you keep it at or near 100% SOC, so if you just cycle the pack between 80-100% SOC you get perhaps 50 TIMES more cycles than if you cycle it between 0%-100%, ie much more than the 5x cycle increase you'd expect if it was DOD independent as is the case with LiFePO4. If you take the optimum charge discharge scenario for lead (keeping it near 100% SOC whenever possible) and used that with normal lithium cells (ie NOT LiFePO4) you'd be destroying your packs much faster than if you allowed them to discharge all the way to 0% with each cycle. Their optimal useage cases are polar opposites of each other. There's simply no rule of thumb that applies across these different chemistries.

The problem with these false generalizations is that they become propagated to the point of becoming a bogus 'conventional wisdom', causing people to waste time and effort into strategies of babying their battery when at times it's completely counter productive, or in your case serves no purpose at all.

**********************************************************************************
Summary Conclusion To anyone browsing this thread in a cursory fashion: Limiting the DOD of a LiFePO4 battery pack to 80% as the OP was seeking to do is both
a) difficult, because of the fairly flat V/SOC curve for LiFePO4, and
b) completely pointless.

 

[dogman dan]

So you are saying that 100 100% cycles is the same thing as 200 50% cycles. (for the lifepo4 graph)

Makes sense to me, one 500 wh discharge is the same energy as two 250 wh discharges.

But it would be very interesting, if you got 4x the cycles. then you'd be looking at double the wh total.

I did find though, that in general a cell stays balanced a bit better if the dod is not so close to 100%. When I'd run a ping lifepo4 down to the bms shutoff, it typically took an extra hour or so to balance charge. While stopping sooner resulted in it charging without a lot of balancing time. It did not take stopping at 80%, just not taking cells over the cliff. So 95% fine.

No idea if riding till the bms popped affected lifepan, I only know the pack sometimes had a slightly less capacity the day after a very deep discharge. Then with enough time on the charger, it would return to normal. What I'm trying to say is this, you don't have to put a hard stop on your discharge at 80% to help your battery last. You don't have to shit a brick if you stop at 82% discharged.

But it can result in a more convenient use if you carry enough battery to not need that last 20% every single trip.

Plan your commute to need only 80%, and then on the windy day you will still have plenty. Still the 80% rule, but for a different reason.

 

[ccmdr]

I've experienced the same with DOD and balancing.

This now also would indicate that whilst trip planning LiFeP04 users can 100% DOD and use 80% calc for a windy day/heavy contraflow traffic. ie, 20Ah 'rated' capacity = 16Ah usable @ 23degC.
Li-Ion and poly guys would have to 80% DOD to get a marginally longer cycle life (compared to LiFe) and then give account for 20% wastage for wind resistance, so in effect 80% of their already diminished 80% DOD wh capacity? ie, 20Ah 'rated' capacity @ 80% DOD for extended cycle life = 16Ah then 80% for headwind/hills etc.. = 12.8Ah usable @23degC...

Is my understanding correct?

Also, for a bog standard BMS both Li-ion/poly and LiFe they are all rated for 100% DOD, so unless you capacity map your cells much like DrkAngel and then x-ref that with your cells 'unloaded' voltage at 80% DOD. As each type of cell has a different capacity map wouldn't the overall varience in 'perceived' 80% vs actual 80% but much higher than originally anticipated, not your +/- 5%. For example at 80% time (not quite 1C, but still relevent), just the varience of the cells composition alone is just over 5%.

 

[ScooterMan101]

Is LiNiCoMn the same as LiFePo4 ? ( Have the same charge/discharge , sag , etc )

What are the common 18650 cells that we see on the pre made packs from Luna , Em3ev, Grin, etc. Made from ?

are they different from the LiNiCoMn cells ?

 

[dogman dan]

Let me repeat. I'm not convinced 80% DOD results in more total watt hours for other lithium chemistries either. Justin just posted data that shows for lifepo4, while cycles may increase, getting less wh per cycle means it evens out more or less. Same lifetime watt hours from the pack.

I SAID. Planning on an 80% discharges for your typical trip is a good idea for other reasons. Like having some reserve when it's a perfect storm of conditions.

You might not lose 80% because its just windy. But you will if it's windy, and cold. So plan ahead for that winter ride home from work too. And, lets face it, your battery won't put out the same capacity in two years either.

But WORRY ABOUT 80%, trying to stop at 80.1% screw that!

 

[markz]

Then basically you want it at
3.6Vmax
3.3Vnominal
2.5Vmin or 3.0Vmin - There is very little energy between the two min. levels.

 

[sketchism]

Ive always run my LiPos that i want to keep around for a while at:
4.1 Max
3.4 Min but 3.3 shutoff so it doesn't start turning off if i'm drooping under load
and i seem to go through packs a lot slower than people with similar riding style but who try eke out the extra percent above and below that.

i live in Queensland, Australia though too, our coldest days are about 10C and only a few a year and its almost always 20C-30C so our packs have it pretty easy temp wise