LFP and Solar Charging

I recently wrote about the need to change my shed batteries from Flooded Lead Acid (FLA) and had decided to use LiFePO4 (LFP) chemistry. I'm still waiting on the AliExpress 100Ah cells (from which to assemble a 24V 100Ah battery) but they have as yet to arrive.

Important note

I'm basing this article on my system, not your and my system is designed to be 24V from the outset. I picked 24V because:

• over 12V this choice halves the amount of amps that any power demand on the inverter takes out of the battery. This has a significant impact on cables to the battery.
• a single panel could be employed (or two in parallel on different sides of the roof) and sized very closely to the battery requirements
• ideal panels were available for very good prices used

This does not apply if you have a panel system which is significantly higher than the battery requirements, for then you'll totally need a contrller ... preferably an MPPT which does DC to DC conversion.

Playing with the unknown

During the mean time I'm playing around that battery pictured in that blog post, but as I had two of them (my mate loves to waste money on Temu shit of which he knows nothing about) I ended up dismantling and then combining them.

The bundle was basically an 8S2P arrangement so with all the tape off it just unfolded into a single linear 8 cell bundle. It was then pretty easy to combine both into an 8S4P arrangement by just removing on BMS's leads and soldering heavy guage copper wire across between the two packs. The nickle stips made for an easy point to which to solder to (to you AMerryKan readers, please note we don't sodder here in Australia we solder it, try reading the letters in the word). This also made the pack more rigid (as the soldered copper held things together better). 

They are then wired back up to one BMS (leaving me with a spare ...). I don't know what the cells are rated as as there is nothing written on them, but I estimate (based on a discharge test earlier that each cell is about 5~7Ah) and so conservatively I've got about 20Ah of pack.

This is entirely sufficient for most of my lower power needs (not the shop vacuum cleaner though) and even copes with the orbital sander. I've left it "open" like this to allow me to use the thermal camera to look for any signs of thermal issues (which could lead to thermal runaway) as well as to make cell bundle measurement easier.

Charge Controllers

This is where things got interesting, as I'm using a Victron PWM controller for a little while now (even with the FLA batteries) because I got sick of the lower end MPPT contrllers shitting themselves in the summer heat. I found that the actual amps into the battery was basically the same with either system, so why fuck around when I'm not usually in a hurry to charge up?

Its been very good and compact. I noticed however that with the LFP batteries that it was taking a long time to get fully balanced and charged. I mean like 6 days to go from 3.33V per cell (which was charged with a 24V LFP compatible externals mains powered charger) to then going to 3.35Vper cell under the control of the Victron.

So I thought I'd cave in and buy a "better charge controller" which was MPPT 

had bluetooth, and app and all manner of bells and whistles and settings.

None of which really did much worthwhile and most importantly the charging not only did not go better; instead it went worse because overnight it drained more and left me with a lower starting voltage first thing in the morning.

So this got me to thinking ... you know, about charging, voltages and that my battery already had a charge controller -> the BMS... perhaps this MPPT stuff only really makes a difference when you have higher Vmax than my panel? Perhaps I was just over complicating things with a controller that assumes there is a BMS? Maybe the BMS can be the controller? Heck the Victron even says that in the case of using LFE the battery must have its own BMS to prevent over charging.

So I thought why not?

Playing with No controler?

Before just wiring it in, I started with some evaluations; like what may panel produced (or could produce).

So I know that its well within capacity to charge the cell, and won't actually take much in the way of losses to bring that Vmpp down to nearly what the full charge capacity of the battery is ... which by the way is about 3.6V per cell (8 x 3.6 = 28.8) which is pretty close to that 3.35V I mentioned earlier.

Now I know data based models aren't perfect, but they're good indicators. So I took my values above and plugged them into this online solar panel simulator.

I used 900W/square meter (becuase its not always 1000 because for a start my panel isn't optimised to the sun angle at every moment of the day) and got this:

Open circuit voltage = 33.44
Short circuit current = 8.12
Maximum power point voltage = 27.19
Maximum power point current = 7.11
Maximum power point power = 193.39

which was interesting as that Vmpp was about what I was seeing using the MPPT unit but the battery charge amps was WAY lower. In fact I've never seen more than 1 amp flow into the battery (except when using the mains powered charger.

The production of V and A looks like  this:

Even pulling the V down to lower voltages like 25V I should stille be getting near to 7 amps passed through. *(note: that 25V and 7.48amps = 187Watts, so not the "maximum power point" but batteries charge on amps and the voltage only needs to be a little higher than the batterys. Based on the simulator even if the voltage of the panel is pulled down to 24V I should still be getting plenty of amps ... yet the battery is only accepting a tiny fraction; like 100mA or less.

Given all this it seems that maybe I could just rely on the BMS to do its job and so I wired up an XT-60, unplugged the panel from the Victron and wired it directly to the battery. 

Everything went exactly as it was going before. Voltage measured across the battery was the same, volgage at each cell was the same and the amount of amps flowing was only a fraction higher (140mA).

Someting is amiss (or I just don't understand something) or maybe I just don't need the Controller at all?. 

Uncertainty amplified

So I turned to the excellent blog post, here, to try to work out what I should be seeing. If you are interested I recommend it, as its an excellent read on LFP chemistry behaviour.

However I think that there is something missing from this, and that the Constant Voltage is not actually due to the charger but is enforced by the electro-chemistry within the cell. To me this is visible in their chart showing Voltage at different C charging rates. 

Notice that they don't show amps into the battery, only the voltage and a Capacity on the X axis. This is vexxed because they actually go to the trouble of saying (in discharging) that:

If your BMS relies primarily on voltage for SOC estimation: NMC's sloped profile is more accommodating; LFP requires dedicated coulomb counting.

we also see that on the charging (upper graph) that the Ah Capacity makes almost no change for the very steep climb in Voltage. There is no time factor on this.

Myself I've found that to get these cells above 3.3V per cell requires hours (and these don't even have a large Ah capacity).

Conclusion

So perhaps I just don't need the Controller? Perhaps this "direct wiring" is good for "some circumstances" but I really don't know.

This seems to be the limits to this system, until I can get the 100Ah cells and test the system on that I just can't know more.

In case anyone is panicked here, I've got it back on the Victron now.

Shrinkflation

You may not have noticed it (but I suspect you have) that one of the ways of increasing the price of something is to give you less of it for the same price. Its been around for decades now and has acquired the (I thought well known) name of Shrinkflation. I recently bought some Tiramisu which comes in a glass "cup" again because I end up using the cups as drinking glasses. However when I picked them up in the supermarket I felt the cup was smaller ... it was, but cunningly smaller than I'd expected.

Below is a pair of pictures showing the new cup beside the old cup. The red marker shows where the smaller cup water level is and how that shapes up against the older one.

Then when I tipped that into the older cup its a lot less than half. The shape based on the sloped sides means that while the cup is just a bit smaller to look at, its less than half of what the older coup contained.

Cute.

SR500 clutch cable sizes and standards

Recently my SR500 needed a new clutch cable:

So as this bike was a stock Japan Spec 1986 model I thought it would be straightforward to order a new part. I went to Yamaha (first mistake) and they confidently identified it and sourced a part for me from a local dealer with old stock.

It arrived and was too long. So I measured it and measured the old one from end to end of the sheath (which is what actually matters). I won't actually give that measurement and you'll see why soon. Below is the cable I had on the bike from purchase (used, and a Japanese import).

(Fig 1)

My measurement proved sort of worthless because nobody (and I mean nobody) knows the lengths of the cables they sell; they only talk in short, standard or long. 

You are fucking kidding me right?

So which cable do I have as my starting cable which was perfect then the Yamaha one was clearly too long ... Hmmm ... its fucking irritating when people in engineering and technical roles speak in qualitative language not quantitative language. You know, like "how many do you want" ... a few. Gives them five ... ahh, no mate I'll need a lot more than that.

Knowing nothing I had to make some assumptions ... what sizes did I now have? So I took a punt and ordered the short from someone else.

So now with two new cables I know that Yamaha supplied me the "long" and someone else supplied the short and my original cable was the fucking standard cable.

FUCK ME DEAD

So you can see from te above tape measure that there is about 100mm between each of them. Giving us

•  +100
• standard
• -100

This means that if (refer to Fig 1 above) we measure from the end of the plain steel cap to the end of the black end (that fits into the engine) the entire length is 960mm. You can now use that as a reference number for determining what you have because that's apparently "standard".

So now you know.

Shed batteries

Well after some time I've decided that the FLA (flooded lead acid) batteries I put into the shed are past their use by date. Not bad really as they were installed new in 2022, which is about 4 years ago.

I'm presently running a pair of donated 24V LiFePO4 batteries that a mate bought (believing the Temu ad) in parallel.

you can see that the foot-print of the pair of 24V's is about half what the pair of 12V's was but they had a RC of 95 (meaning that the battery would survive for 95 minutes under a constant 25amp load before 10.5V is hit ... optimistic if you ask me) 

I've never been really satisfied with how the recharge after discarges (probably due to the internal resistance of FLA). See this post for my analysis back then. 

This is one ofthe tiddlers ... rated at 200Ah

talk about optimism ... I pulled one apart and the Cells were 7Ah and there wasn't anywhere near enough of them. Further, the BMS was clamped at 9Amps.

Well that's 400W or so, so it'll do for now ... I'll post again when the new cells arive