Battery meter logic

[bmclaurin]

How does the on-board battery meter determine estimated remaining battery capacity? I am very familiar with state-of-charge (SoC) curves (like the .csv files that are uploaded in eScribe) and methods of determining pack capacity (like eScribe's Battery Analyzer tool). I have used those tools to determine my pack's capacity, entered that into eScribe, and have uploaded the resulting .csv file to my device).

I'm just intellectually curious how the on-board battery meter uses this information to determine my pack's estimated remaining capacity percentage. Originally, I was just thinking that it simply measured the resting voltage of my pack and interpolated the capacity % from a lookup in the .csv table. But somehow I suspect it's not quite that simple in reality. If it was really that simple, then it wouldn't matter what value I entered into the pack capacity field in eScribe (since it could derive an estimate based solely on voltage directly from the .csv table).

I'm now thinking that it works something like this:

1. At the completion of each and every charging session (whether fully charged or not), the chip measures the pack's resting voltage and makes an initial determination of capacity % from a lookup/interpolation of the .csv table.

2. It multiplies that % by the total pack capacity (as entered in eScribe) to determine an estimate of how many watt-hours (Wh) are available in the pack immediately after the conclusion of the charging session.

3. From that point forward, starting as soon as the battery comes off the charger, the chip measures how much capacity has been drawn from the battery and subtracts that from the initial estimate of capacity determined in step 2.

4. The resulting remaining capacity (in Wh) determined in Step 3 is then divided by the total pack capacity (as entered in eScribe) to determine estimated % capacity remaining.

5. That % is then displayed on the chip's screen.

Now, I doubt I have this exactly right. There is probably more going on here. But that's how I have it sorted in my mind at the moment.

Can anyone comment on whether this is how it really works? It doesn't really matter at the end of the day. I'm just intellectually curious. I like to know how things work.

Thanks.

[Spector NS5 RD]

heres what i found, the battery shaped meter is directly related to the watt hour you input in escribe. i tested it by putting in 1 watt hour for the battery and vaped it to empty. it took less than 10 puffs to take the meter from full to empty. my guess its just an estimate with direct relation to what you put in for the wh's. it has nothing to do with the packs voltage at all, completely disconnected from pack voltage.  

[DannyK9]

When I first got my DNA200 it was set on 8wh which I thought was low for a 1400 3S so I done the maths and calculated it was 15.4wh so I set it to that, escribe also calculated this amount. However it doesn't seem to last any longer than when it was at 8wh, only difference being a slightly lower build. So I've no idea, as it stands I get around 12 hours vaping moderately at 30W is this good ? I suppose so.

[Margucci]

if you run the battery analyzer it will calculate how the energy extracted from the battery (the exact Wh) will change the voltage of the battery pack. essentially what it does is fully charge the pack and simulate taking pulls as a specific wattage until the battery reaches its set cutoff voltage. through doing that it knows how much energy has been used and it will adjust the battery meter accordingly. just because the battery says 1400 mAh doesnt mean that is the capacity you will experience. batteries are more efficient the lower the current draw. this is why when you run the battery analyzer its important to set the wattage to what you typically vape at (take into account that the wattage you have set isnt the wattage being used for the majority of time when under TC). also, not all packs are created equal. some are just over rated. 

[bmclaurin]

Thanks for the responses, guys. I understand how the battery analyzer works, and I've used it to determine my pack's actual capacity. And I understand that a battery's total capacity is correlated with rate of discharge. The reason I made the original post of this thread was simply to try to understand how the chip uses this information to estimate remaining capacity (i.e., state of charge) of that battery at any given point in time.

In particular, I don't understand how the on-board battery meter uses the battery's estimated total capacity (as input in eScribe), especially since the on-board meter only displays state of charge as a percentage of total capacity (as opposed to the number of watt-hours remaining). The on-board battery meter could determine state of charge (as a percentage of total capacity) simply from a measure of the battery's resting voltage and a lookup in the discharge profile .csv file.

If you handed me your battery and your discharge profile .csv file, and you never told me the total capacity of the battery, I could still give you a decent estimate of what portion (as a percentage) of its total capacity is remaining simply by measuring its open circuit voltage with a DMM and looking up that voltage in your .csv file. I don't need to know the pack's total capacity (in watt-hours) to make that estimate. And I don't need to know how much you had been vaping it prior to handing it to me.

So, I guess my question really boils down to: Why does the on-board battery meter seem to be influenced by the value that I input into eScribe for the pack's total capacity? How does it use that information?

[Margucci]

from an accuracy standpoint, once the battery analyzer is run so that the total capacity given regular usage is know, it would be best if both the battery icon and percentage were calculated based upon the capacity and not the voltage. i believe this is the case already. i have not noticed an accelerated decrease in either the battery bar or percentage when fully charged or near empty which is when you would expect to see the meter change the most if it were being calculated using voltage as most mods do. 

[bmclaurin]

Thanks, Margucci, but that's what the discharge profile curve does. It compensates for the non-linear decline of the battery's OCV (open circuit voltage) as its charge is depleted. What you're describing would be true if the .csv file contained just a straight diagonal line as opposed to a curve.

EDIT:  My apologies if I'm coming across as argumentative, as that is not my intent. I'm truly just trying to understand how this works (for curiosity's sake).

[Margucci]

perhaps im not understanding your question. from what i understand there are two scenarios:
- the battery capacity is calculated by voltage. you would have a large change in the displayed capacity at full charge and near empty. this obviously isnt the case
- the mod calculates the remaining capacity of the battery given the Wh specified in escribe. this results in a constant change in the battery meter as the battery discharges with no steep drops when fully charged or near empty. using the theoretical capacity of the battery would give a decent approximation of the capacity of the pack but for a truly accurate battery gauge the exact capacity would need to be known (and can be determined using the battery analyzer). 

are you saying something else?

[bmclaurin]

I'm saying that in your first scenario (where remaining capacity % is determined by voltage), the discharge profile curve (.csv file) compensates for the large changes in displayed capacity that would otherwise occur at full charge and near empty. Let's say your .csv file was just a straight diagonal line consisting of two points: (0%,3.1v) and (100%,4.2v). In that case, yes, you would not get a smooth change in displayed capacity % at full charge and near empty. But the .csv file contains multiple points (resulting in a curve) that account for the non-linear drop in open-circuit voltage as the battery's charge depletes.

Thanks for sticking with me through this. I like understanding how things work.

[Margucci]

i think i understand where you are going with this. yea, the discharge curve would compensate for the voltage at the extremes of capacity to give you a more accurate measure of remaining capacity. however, i dont know why this would even be necessary. it seems to me that if the total capacity in Wh is known after the test is run, and the board obviously knows how much power has been discharged each draw, why would capacity not be calculated that way. that would be by far the best method. if you notice the battery meter losing its accuracy as you near the lifespan of the battery pack you can just run the test again to calculate the new total capacity at that point in time. there may even be a way for the mod (or more likely escribe) to extrapolate the new total capacity as the pack is used. that would be a lot more complicated though. 

[bmclaurin]

I completely agree that, in principal, the chip could measure how many watt-hours had been drawn from the battery and simply subtract that from the total estimated capacity (as input in eScribe) to determine and display remaining capacity.

But, to do this, the chip has to know where it's starting from. Let's say I remove my battery from my device and give it a partial charge on an external charger (which could be a common occurrence with 18650's, for example). When I place the battery back into the device, how does the chip know where to start subtracting from? My original post for this thread was making an educated guess that this is where the .csv file comes into play. I'm guessing that the chip must use the .csv file to make a preliminary estimate of state of charge (by measuring the battery's OCV) and, from that point forward, it uses actual watt-hours drawn to estimate remaining capacity.

So, I guess I'm saying (speculating/guessing) that the chip uses a combination of these two methods (i.e., open-circuit voltage and watt-hours drawn). I guess I'm also saying that the second method alone (i.e., watt-hours drawn) is insufficient because it couldn't account for the effects of external charging, etc.

Sorry to be so tedious with this. None of this is necessary to be able to safely and satisfactorily use the DNA 200.

[Margucci]

no, my apologies. that would be 100% correct in the case of batteries which can be swapped. sorry for the misunderstanding. in this case the open voltage of the pack would certainly be needed to get a starting point. i didnt even consider 18650 batteries because until the wismec mod is released it is still a very niche case. i wonder if any changes will be needed to escribe or the firmware to account for the fact that people will be using different capacity batteries between units and/or external charging. 

[bmclaurin]

It would be handy if we could have the ability to upload multiple "battery profiles" to the device (each with their own specified Wh capacity and .csv file) and select them on the fly from the device. As it currently stands, if you want to have the ability to swap out batteries of differing capacities and maintain the integrity of the battery meter, it seems it would be necessary to manually change the capacity in eScribe each time you change batteries and upload the new settings to the device.

[Margucci]

or perhaps just something as simple as the option to base the battery meter off of battery voltage alone when using 18650 batteries. it wouldnt be nearly as accurate but it would certainly solve the issue of using batteries with different capacities.