DIY portable, useful LiFePO4

[booster]

Very interesting finds, as always.


The voltage rebound noted is very interesting as we have heard so much about voltage control of charging and recharging being the way to go for lithium. Don't really know what to make of that, but appears that we may be back using amp hour counting to best determine recharge points?


Those are pretty extreme charge and discharge cycles on the shortened life testing, but some of the early claims for lithium weren't all that far off from them, IIRC. The high rates would really make the cutoff happen quickly once the voltage started to drop at the end, I would think. Some of the drop in manufacturers still recommend 100% of rated discharges, but in most cases we don't know how much, if any, space they leave at the bottom in particular. Many are now going quite conservative on the discharge and charge rates, though.

[Winston]

Originally Posted by markopolo

Memory effect: - idling for a sufficiently long period of time under certain conditions being used to erase the memory effect is interesting - https://www.psi.ch/en/media/our-rese...-ion-batteries

Our “old” understanding was that LiFePO4 batteries do not exhibit “memory”. To some extent, even after reading this paper, our original belief remains intact. It appears that other lithium experts also doubted that lithium has memory as these researchers noted:

“But the key was the idea of looking for it [memory] at all . . . It thus took a flash of inspiration in order to ask what might happen during partial charging in the first place.”

These are curious statements as ‘memory’ has always been a phenomenon associated with “partial discharging” (e.g. nicad). It is hard to understand why it took “inspiration” to undertake this study unless, as we surmise, it is because ‘memory’ in lithium, is a mere technical nuance of comparatively minor importance.

Another reason we’ll maintain our ‘lithium has no memory’ belief is the fact that ‘lithium memory’ can be erased:

“. . . that idling a sufficiently long period of time can be used to erase the memory effect . . .

If this 'debility' can be erased, why worry about it?

Ahh, but what is “idling”? We remain uncertain as this paper was less than clear:

“The memory effect only vanished if one waited a sufficiently long time after a cycle of partial charging followed by full discharge . . The memory effect remained however providing you waited after the partial charging and before the incomplete discharge.”

First, how long is “sufficiently long”?

And does this mean that a “full discharge” is required? Are those conservative lithium users who don’t take their ‘packs’ down to “0" SoC missing the recuperative effects of “memory erasing”?

And that last quoted passage to the effect that memory remains if you “wait(ed) after the partial charging” but before an “incomplete discharge”? Incomplete? What if you wait after a partial charging but before a Complete discharge?

Nor does this paper tell us whether our once annual full discharge regime (undertaken to test capacity) is sufficiently frequent to effect the desired ‘erasure’.

Thanks for posting this interesting article . . . for all our uncertainty, it still is a very interesting addition to our understanding.

[Winston]

Originally Posted by booster

The voltage rebound noted is very interesting . . .

Markopolo's post included some 'weighty' technical discussions that, at times, caused our layman's eyes to glaze-over. We apparently missed the discussion you referenced concerning "voltage rebound" and, as our interest in lithium charge profiles remains of high interest, we're wondering if you could point us to that discussion we missed? Thanks.

[booster]

Originally Posted by Winston

Markopolo's post included some 'weighty' technical discussions that, at times, caused our layman's eyes to glaze-over. We apparently missed the discussion you referenced concerning "voltage rebound" and, as our interest in lithium charge profiles remains of high interest, we're wondering if you could point us to that discussion we missed? Thanks.

https://www.classbforum.com/forums/f...tml#post107214

[markopolo]

Winston knows more about this stuff than me for sure.


Here is a paper - https://res.mdpi.com/d_attachment/ap...8-01825-v2.pdf - that I'm having difficult reading let alone understanding! Is cycling at low SOC's better?

[booster]

Originally Posted by markopolo

Winston knows more about this stuff than me for sure.


Here is a paper - https://res.mdpi.com/d_attachment/ap...8-01825-v2.pdf - that I'm having difficult reading let alone understanding! Is cycling at low SOC's better?

I think I had run across that article before, after video of whomever it was that talked about severe narrowing of the usable range of lithium. Still makes my head spin, but the discussion at the end is pretty clear.


Low state of charge cycling using a quite reduced range of SOC cycling greatly improved life, and it was even more of an affect than low vs high discharge rate.

[markopolo]

Did you see anything in that paper about amount of energy extracted? I searched for words like hours and energy but didn't see anything about the amount of energy obtained.

[markopolo]

A few photos of the type of prep work done prior to assembly:
prepping the copper.JPG


Two layers of the copper strip soldered together then tinned:

two layers of copper then tinned.JPG

solder tabs.JPG

Not particularly difficult but took a lot of time. I think Booster had an idea of what I was in for but I was blissfully unaware prior to committing to the project!

[markopolo]

Originally Posted by markopolo

With my latest paralleling plan, current from the LiFePO4 will be wasted as it goes through the automatic charge relay and into the starting battery and also supplies any GM related parasitic loads.

I can turn the automatic charge relay off by interrupting the small gauge ground wire using a switch but I'm stumped as to how to automate that. Automatic turn on is 13.3V and turn off is 12.8V now.

Probably just have to live with using the switch, maybe one with a led and move it to a more visible location.

An update: $20 will mostly solve the problem at least in theory. $10 charger control relay + relay + switch. The settable voltage charger control relay will trigger the other relay to make or break ground either enabling or disabling the Smart Isolator (ACR, VSR etc,). If the system voltage is less than 13.4V for example, the Smart Isolator is disengaged. The settable voltage charger control relay has a manual override triggered by a standard on/off switch so I can also force it on or off. Looking forward to trying it & checking power consumption.

[booster]

Originally Posted by markopolo

Did you see anything in that paper about amount of energy extracted? I searched for words like hours and energy but didn't see anything about the amount of energy obtained.

I did not notice anything, but didn't look for it either. The very poor scale and identification of parameters on the charting is so frustrating that it very hard to dig around for everything that should be easily seen, that I looked for the early "what we are going to do" and the ending "what we found and concluded" stuff.

[markopolo]

Originally Posted by Winston

.............First, how long is “sufficiently long”?
.........

Exactly. Time is expressed as years, days, hours, seconds etc. Sufficiently is just place holder at best until conclusions are reached.


How about the commonly seen "high state of charge"? is that 73% or 94%?


/end rant

[markopolo]

Originally Posted by booster

I did not notice anything, but didn't look for it either. The very poor scale and identification of parameters on the charting is so frustrating that it very hard to dig around for everything that should be easily seen, that I looked for the early "what we are going to do" and the ending "what we found and concluded" stuff.

I skip to the just the Conclusions often!

[JohnnyFry]

I have finished my 150 A/H LiFePO4 pack for my RoadTrek. Testing today, I took it down to 20% at C/10 rate. Recharging from the TrippLite went smoothly at 30 amp until at 100% the BMS stopped charging. Further testing with Renogy dc/dc charger went as expected. Although the battery is mounted inside in heated space, I did make provision for low wattage heating pads mounted between the cells. I also installed a 300 Watt Sine Wave inverter to accommodate electronic devices that won't play well with the MSW TrippLite.

[Winston]

Booster,

That lithium voltage ‘sags’ and ‘rebounds’ after, respectively, charging or discharging is, we think, expected. We don’t think this should preclude use of constant voltage ‘targets’ or charging.

As we’ve noted in other posts, we conduct annual full-discharge tests of our lithium pack in which, after each 10% or 20% discharge interval, we let the pack ‘rest’ (absolutely nothing connected to the cells other than the permanently mounted BMS circuit boards) for at least 10 hours. This provides an accurate (and repeatable) chart/curve of pack voltage vs SoC (although we don’t have the 1 millivolt measurement resolution that some of you have).

At this point we believe we’re on solid ground selecting a charging voltage corresponding to “the” target SoC level. It is our practice to place this voltage on the pack and leave it there indefinitely. Eventually the pack will reach the target SoC level at which point we would expect the output current from our charger (power supply) to be zero. And if we removed all loads from the system, it would be, or approach, zero.

We do not ‘float’ our batteries. Placing a charger/power supply of a predetermined voltage equal to the resting voltage of the pack (at the chosen SoC level) is not floating. When placing a charger/power supply of voltage matched to the resting voltage of a battery, no current flows into, or from, the battery. The battery doesn’t even know the charger/power supply is there. The advantage is - - when a load is applied - - the charger/power supply ‘picks-up’ the load (i.e. supplies the power to the load). The battery just sits there, neither charging nor discharging.

The only issue for us, then, is: “At what SoC do we want to maintain our lithium pack?” The jury has yet to return its verdict.

[Winston]

Originally Posted by markopolo

Here is a paper - https://res.mdpi.com/d_attachment/ap...8-01825-v2.pdf - that I'm having difficult reading let alone understanding! Is cycling at low SOC's better?

Originally Posted by booster

Low state of charge cycling using a quite reduced range of SOC cycling greatly improved life, and it was even more of an affect than low vs high discharge rate.

This paper is an eye-opener - - although it appears to be directed more to NMC/LMO chemistries than LPF. And further, its extremely low “DOD range" (defined as the percentage of total capacity that is ordinarily used before recharge) of 10% may be inapplicable to the RV market where we suspect most use more of their available DOD range. (Yet, upon reflection, there are many days when we’re on shore power, driving, or have plenty of sun, where we use less than 10% DOD). And, even for those who use more than 10% DOD, we suspect that the “keep your overall SoC low” teachings of this study still apply.

This is the first time we’ve seen the term "calendar ageing" which, as we understand it, is simply the loss of capacity which occurs over time even without using the battery, that is, while it's just sitting 'on-the-shelf'.

We have heard that it was not good to maintain a lithium battery at 100% SoC . . . and some have opined that storage at 50% SoC is preferable. But this paper suggests that there's a really big advantage to storing lithiums much closer to total discharge, i.e. at 10% SoC.

The problem for us is that our batteries are 'in service' most of the time - - we don't have long intervals of storage. Thus, if we're to gain the "calendar ageing" advantages shown in the paper, we'd have to set our target SoC to a very substantially lower level . . . and hope that we would have enough advance notice of future needs to recharge the pack to meet those needs. Of course, the unexpected unavailability of shore power as well as the unpredictable nature of sun give us 'pause' in maintaining our pack at low SoC.

So, yes, marko, it looks like we should all dial-down our SoC.

[markopolo]

Originally Posted by JohnnyFry

I have finished my 150 A/H LiFePO4 pack for my RoadTrek. Testing today, I took it down to 20% at C/10 rate. Recharging from the TrippLite went smoothly at 30 amp until at 100% the BMS stopped charging. Further testing with Renogy dc/dc charger went as expected. Although the battery is mounted inside in heated space, I did make provision for low wattage heating pads mounted between the cells. I also installed a 300 Watt Sine Wave inverter to accommodate electronic devices that won't play well with the MSW TrippLite.

That looks great Johnny. Very neatly done and everything accessible.

[markopolo]

Originally Posted by Winston

.................. As we’ve noted in other posts, we conduct annual full-discharge tests of our lithium pack in which, after each 10% or 20% discharge interval, we let the pack ‘rest’ (absolutely nothing connected to the cells other than the permanently mounted BMS circuit boards) for at least 10 hours. This provides an accurate (and repeatable) chart/curve of pack voltage vs SoC (although we don’t have the 1 millivolt measurement resolution that some of you have)..................

I just looked through the logged data. Voltage rebound reached it's peak 23.22 hours after discharged terminated.

Voltage rebound made it to 4mV of peak voltage after 7.99 hours.

[booster]

Originally Posted by markopolo

I just looked through the logged data. Voltage rebound reached it's peak 23.22 hours after discharged terminated.

Voltage rebound made it to 4mV of peak voltage after 7.99 hours.

Too lazy to look it up, how much does that much sag affect the apparent SOC?

[booster]

If the downsides of using even 90% of rated capacity turn out to be large enough to make a significant difference in battery life and the recommendations start getting to 30% usable or such, this would be a huge blow to the lithium market because of the much larger capacity needed. Both size and weight plus depth of discharge capacity are big selling points against AGMs. We would be talking about 3-4K amp hour (12v) banks so nearly 50 kwhr, for those that want AC use.


With all of this stuff, as the information trickles in, I tend to compare what we hear to existing products and applications. The primary comparisons would electric vehicles, phones, tablets, laptops, etc.


The one that sticks in my mind is that we were told that the basic spec from Apple on phone batteries was 700 cycles, and if that is correct it would certainly indicate that they have chosen to sacrifice battery life and cost for the most capacity and least space. Perhaps the RV market has already done that, also, but chosen a different balance point, but I tend to doubt that with all the varying specs we see, especially in the drop in brands.


The articles also addressed discharge rates as being a big predictor of battery life and I think that the difference in life that our personal products have shown that likely to be quite correct in existing products also. We have had multiple phones, one tablet, and four laptops. All seem to have had batteries last much longer (cycles not time) than other people we know who play display heavy games on them and discharge rapidly that way. Our tablet is used by DW for very docile games like solitaire and is probably approaching 2000 cycles now. On the laptops, the one we had that needed new batteries was also the one that had the smallest battery and discharge in much less time so probably higher rate.


Anybody's guess is probably better than mine, but it looks to me that the RV lithium market will settle in to an only slightly different set of parameters than they currently have, mainly because they would lose the market if they go toward longer battery life preservation schemes. They need to be able to provide the very common size 200ah battery bank at competitive prices and performance to AGM and adding all the stuff needed to do it in a much better way would make that impossible. (I will point out that the AGM market did this a long time ago, as the equipment used to charge them is certainly not the best, in nearly all cases, to extend battery life).


For lithium, I think it may settle for fixed voltage, full cutoff, chargers at a 95% SOC point. They would power supply type chargers that would run the coach with the battery off and no reference battery needed.


Optional heaters that are not overly complicated to use. If the battery goes into low temp shutdown, you do manual switch to turn on the heaters that would run off he power supply charger on shore power, or the alternator through a non reference needed b to b charger.


Solar and alternator chargers would also be non referenced.


If the equipment folks come up with the right stuff, it could be done reasonably, I think.

[markopolo]

Originally Posted by booster

Too lazy to look it up, how much does that much sag affect the apparent SOC?

I don't know the capacity per mV.



I was alerted to this because I needed to enter accurate voltage values for Full, 80%, 60%, 40% & 20% in the BMS software. I was allowing up to 2 hour-ish recovery periods. I also wanted to know what the 50% SOC voltage was. 50% SOC point had an overnight recovery period and the voltage was higher than what I had recorded for 60% SOC.


You see how easy it was to be 10% off on SOC using voltage.


The getting to within 4mV I mentioned might help to avoid a larger error in my example. 13.164V was peak (I think that device reports a bit lower than true voltage). Lose the 4mV and you'd still see 13.16V. Lose 5mV and that's 13.159 and a lot of meters drop the digits they can't display and you'll end up seeing 13.15V. Suddenly you're off by 14mV from actual peak voltage.