How's everyone's [55Wh] battery health and wear looking?

Community Support
143

I have not done a full charge very recently but I will try that and see how much this value changes.

I don’t have records of the initial capacity however I remember it was below 100% (54.XX Wh). I have a record two months after purchase with a value of 98.4%, 9 days later 97.3% and 40 days after that 97.1%.

Edit: Seems I made a prior post before forgetting my OOTB value.

Let’s see if I’m wrong and in 2 months time have fallen below the 80% mark did I get a bad battery or is the charge limit behaviour causing some degradation? I’m very carefuly to not stress the battery with heat or deep cycles/fully charging it… My old iphone 3GS had 86% after 500 cycles…
IMG

145

Ok Well here is a graph from day 1 to day 375

As you can see, with my persistent checking and charging it appears as though I have none or next to no wear after a year.

In the first few months I did not monitor and even when I did on the odd occasion I had no access to the number of cycles.

Since then I set the Battery charge limit to 78% and started to use the laptop plugged in 99% of the time and the results are surprising.

Of course the readings are not necessarily ‘true’ but what else is there to go by.

What I haven’t done is to see how long I can watch videos on a full battery, that’s next.

graph~all
graph~all1337×488 74.5 KB

UPDATE: Immediately after this post I checked and it is 5.1% so that is a 2% increase during the day whilst the laptop has been of so ??? EDIT UPDATE: By 11am the following morning it has dropped to 4.0%

Almost invariably morning readings are lower, so room temp may be significant as my cabin in the woods cools a lot overnight.

For details see
http://217.155.51.23/newdevices/framework/battery/mar23.php

146

Well 4 full cycles later and now I’m down to 88.0% :man_shrugging:

147

Yes I have thought about how these measurements are taken. Clearly the info is built in somewhere as Win and Ubuntu etc. give the same info.

148

Not sure if that was directed at me but yeah the BMS I would suppose.

1 Like
149

I suppose I’d best contact support and see if they will tell me exactly how the BMS is implemented and how the info is made available.

Have just submitted a support request and clearly didn’t notice the url until is was too late. It seems I have now given my email address to some turd party.

https://framework.kustomer.help

150

I believe this is the correct URL it is what is linked on the “submit support request” button from the contact support link Framework | Support

151

Yes it is the correct one I just don’t like it when companies use third party businesses to route and monitor the requests. kustomer.help has been going since 2015

Next time I will directly email support. I’m usually more careful ??

UPDATE: I have sent a letter noting my displeasure with the third party use whilst asking for exact details on how the BMS works, I imagine it’s a part of the battery pack and further, how that info is made available for the OS

2 Likes
152

We use Kustomer as our ticketing system and have since our inception. Any frame.work domain linking to support is simply a mask as we’re utilizing Kustomer’s cloud platform. This is common practice. Knowledge Base/Support Tickets… Kustomer. Guides… Dozuki. Bugs… Jira. Welcome to 3rd Party tooling and platforms for business.

3 Likes
153

Thank you for the update.

All the best

1 Like
154

Here is a good video I found awhile back explaining the various ways that lithium ion batteries undergo physical damage and degradation. Very thorough, but I do not know for sure how many of these processes are applicable to the lithium polymer batteries sold by Framework.

3 Likes
155

Rhetoric is fine but I rarely find any data to substantiate the theory, but yes thanks for the link the video is helpful.

156

I thought I’d had enough for the last year of monitoring the battery wear but what is happening is clearly of interest.

After some 3 months of a lot of use unplugged and once I had discovered that I could ‘read’ the wear and the BIOS update that permitted the ‘reading’ of cycles I took a devoted interest.

a) I started to use the laptop plugged in >90% of the time and b) set the Battery Charge Limit (BCL) to 78%. This seemed not only to stop the recorded wear rise that had reached some 7% after four months but it started to decline.

As it started to decline I thought either the computer was reading wrong or what. I have since discovered that the battery pushes the info to the OS and the OS does very little. So I thought maybe the battery itself could do with reset.

So approx each month I reset the battery by running it down until the laptop automatically turned off and then charged it to 100%. Coincidently with that each time the recorded wear dropped. I know it’s a bir uncanny.

Now 13 months later I thought time for a change so I decided to lower the BCL to 69% This was due to the fact that at 78% the battery was till charging to nearly 4.165V which I can understand may wear the battery. So now I have the battery at 4.032 with the BCL at 69%

I was expecting a change but not the one I have. Today I did what was to the be the last of this routine checking on 31st March and what happened . . .

The wear has just recorded 2.8% 2.7% which is lower than it was upon receipt.
So either

  • The reading is wrong now and there is a lot more wear or
  • The reading is correct and the previous reading are way out.

The take is this.

If measurements are not taken over a long period with different settings and use patterns, and tabulated and graphed there is no easy was to see what is/maybe happening. Here is the graph from 4th Jan 2023 to 30th June 2023.

graph0123~0623
graph0123~06231383×511 65.6 KB

Each day I take multiple readings and the figures on the graph are the lowest for each day. To see the actual data that I take it from see the monthly data

http://217.155.51.23/newdevices/framework/battery/apr23.php . . . etc.

Graph of data since receipt in Feb 2022

graph~all
graph~all1385×513 78.3 KB

UPDATE 5th April: Since lowering the BCL to 69% the ‘wear’ has been consistently lower, between 2.6% each morning and a max of 3.0% late evening.

Screenshot of HWInfo on 9th April 2023, after 13 months ~ no joke.

wear
wear808×624 33.1 KB

UPDATE 19th April: Since I inadvertently let the charge drop a bit I used the opportunity to reset the calibration by a full discharge > a full recharge and then > discharge to my ‘new standard’ of 66%

The cycles went up 2 which I wasn’t expecting and the wear sort of doubled in that in went from 2.6 to 4.8 ???

157

I just did a deep discharge and recharge cycle, followed by a couple of more 90-100% cycles to see if it would get any more in the top end. /sys/class/power_supply/BAT1/charge* reports 3084000/3572000 = 86.3% SOH. 1.4 years in service, usually charged to 75%, not sure how many full cycle equivalents but I’m going to guess maybe 300ish.

Glad replacement batteries are available-ish in the marketplace. Will be looking forward to upgrading to the 61Wh and moving this battery to the router/wireless AP/NAS unit instead.

3 Likes
158

I received my diy framework i5-1135g7 in February 2022. As of today I’m at 196 cycles and the wear level is 11.9%.

I upgraded it to a 12th gen 1240p this past week (I spilled my drink on the mainboard/keyboard in March which necessitated an replacement).

1 Like
159

So you are still using the original battery?

What is your charging and use routine and have you tried a full discharge and charge etc.

Thanks for the info :slight_smile:

160

I think you are reading too much into this. This data is mostly within measuring tolerances and by undercharging so much you really should not have any measurable wear jet.

Since I don’t exactly know how your full charge process works I’ll just have to guess you don’t discharge to complete empty with a constant load so the zero point of the fuel-gauge in the battery just keeps drifting downwards with every time which explains your graph.

And even if you do discharge till the battery turns off if you have different wear you’ll get different 0 points because the voltage sag is different.

Would be neat if the framework ec got the battery re calibrate mode thinkpads have where you can leave the laptop plugged in and it discharges and then fully charges the battery itself giving very consistent results.

161

Thanks for the comment.

I have provided details.

For example as I said every few week I do a full discharge and then charge to 100%. Sometimes i do this twice and keep the battery at 100% for a few days.

The wear in the first few months was 7.5% and 45 cycles, this was when I was using most of time not plugged in

So sure the readings are questionable but then how else would the battery state be assessed.

As I’ve said you can see details of the changing charge rates etc.

The main issue is that I mostly use it plugged in and if lower the Charge limit effects the recorded wear in such a way as I have seen, then clearly I will just keep testing to see how long it can keep this up :slight_smile:

AS I said all the details of Charging and reported wear for 13 months can be viewed.

http://217.155.51.23/newdevices/framework/battery/apr23.php

162

Sorry in advance, this may turn into a bit of a semantic debate but in this case it’s important to define terms accurately so we are not talking about different things.

Yeah but not HOW you do a full discharge, this part is important because if you don’t go all the way to empty those watthours will be missing in the total. If you just discharge till the is shuts down that would pretty nicely explain your graph creeping downwards each time.

There is 2 was to look at it, for one there is actual wear (the one I am talking about) where the battery cells loose capacity and get higher internal resistance and self discharge and then there is the wear reported by whatever tool you are using which is the difference between the design capacity and the capacity the battery controller thinks it has (which I think is what you are talking about).

The fuel gauge in the battery basically works like this, it has a coulomb counter so it basically counts the energy that goes in and out and updates the “capacity now” value based on that. when it is charging it increments, when discharging it decrements. Of course that process isn’t perfect and so that drifts over time and might over and under report. when you fully discharge (like fully fully) it counts down till it is at zero and then keeps going until the cell voltages reach the cutoff point, which is why when you re-calibrate a really confused thinkpad battery or a battery where you swapped the cells the battery percentage stays at 0% for a long time. Once it actually emptied it starts counting from actual zero and you should get a pretty accurate capacity. Note that the actual capacity of the battery did not change just the measurement of it. And even that may not be exactly the same run to run if you don’t control the discharge rate because the battery voltage sags based on load which makes it reach the min faster on higher loads. Also some battery controllers get confused if the actual capacity is higher than the programmed design capacity XD, I once put 120wh worth of cells onto a 24wh thinkpad battery and boy was that controller a dick about it, it just capped the measured capacity at 24wh and then just stuck at 0% for hours.

Never going to 100% or fully empty does lead to accumulation of errors in the measured capacity but it does cause less actual wear.

I bet your test cycles put more wear on your battery at this point than your normal use XD

163

The discharge rate is always the same, browsing and the odd video which takes around 5 to 6 hours.

Discharge is done with the same use, until auto power on. Then I wait an hour or tow and power on and may get another half an hour use. It power off I leave it for another hour and power on and may get 15 mins. Usually it will not power on a four time.

So if the charge is coulomb counting, that is fine in measuring the actual energy.

As resistance increases as the battery gets charged, then charging becomes increasingly inefficient.

I can also measure the amount of watts used to charge externally

So the issue is how representative of the battery charge is the coulomb count ??

Regarding wear : I see it in relation to the cycles.

Framework state 1000 cycles for 20% wear, or 50cycles for 1%.

Framework don’t say how the cycles are achieved or over what time, hence my checking to see if what I do correlates.

I’m aware that as a battery wears each cycle get smaller and yes charging to 100% will be a bit stresy for the battery but then I can’t see Frameworks statement 1000/20% would not included and 100% charges.

Again thanks