EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: Renate on January 04, 2021, 01:34:27 pm
-
Solar power can be very limited time-wise during the winter months.
There are often times that you have plenty of power but your lead/acid is saying,
"I see that you've got lots of power, but I only feel like taking a couple of amps right now."
Then the sun goes down and the battery complains, "Hey, I wasn't done yet!"
It seems that in this case a lead/acid isn't optimal.
A LiFePO seems more capacitor-like and would be happy to take whatever is being served at the time.
Yes/no/comments?
-
Only things good about lead acid is that it's cheaper than lithium (at least conventional lead acid is, sealed is often more expensive than some lithium) and prefers to stay near full charge. What would make sense is a hybrid setup that has lithium for daily cycling and lead acid for reserve.
-
Here's a reasonable documented comparison with prices and capacities:
https://unboundsolar.com/blog/lead-acid-vs-lithium-batteries
Still bloody expensive and useless, except maybe in California, but is a start. If the price of energy will get even higher in Germany, it will start to look good here as well, one must also have a quarter of million to buy a miserable house to put the panels on it (0.25mil will only get you a some kind of shack in southern Germany where it makes most sense to have solar).
-
LiFePO batteries usually have a much longer life (more cycles) than lead acid ones. So over their lifetime they cost less. But the initial investment is high.
-
Some people calculated here, that electricity is cheaper from the network, than from a Tesla powerwall.
With LiFEPO you will have to do some maintenance, balancing is often times not possible without discharging the battery protection circuits. Your inviter also probably doesnt do island mode.
I seriously don't know why would anyone want to install an extra fire hazard at home. Unless you get the batteries for free, or you make money by uploading youtube videos from your fire hazard.
-
There are often times that you have plenty of power but your lead/acid is saying,
"I see that you've got lots of power, but I only feel like taking a couple of amps right now."
Then the sun goes down and the battery complains, "Hey, I wasn't done yet!"
It seems that in this case a lead/acid isn't optimal.
A LiFePO seems more capacitor-like and would be happy to take whatever is being served at the time.
Both lead-acid and LiFePO batteries have the same limitation; since both use voltage for charge termination, current drops when as full charge is approached. Flooded Lead-acid batteries are actually better about this because they rely on controlled overcharging for cell equalization and stirring the electrolyte so they are charged to voltage slightly higher than their float voltage.
But in practice the solution is to use a larger battery bank so that it can absorb all of the power available at the end of the day.
-
Well, to put this in perspective...
My case is mobile, 600 W solar, 40 A controller, 224 AH 12 V (2 x GC2 AGM).
To minimize wasted sun I've started to let the battery take the full power in the morning until I can turn on a load without affecting charging.
Yes, that does waste the sun because I have to wait until there's 125 W wasted before I put a load on.
But that does leave me with the biggest time span where the battery is getting something.
-
Is your solar charge controller of the MPPT type? If not, you're indeed missing out.
-
Is your solar charge controller of the MPPT type?
Yes, of course it is.
Here's a graph to show you kind of what's going on.
It shows an actual day, yesterday.
The orange dots (that I hand drew, but I'm writing something to do that) show the actual available power.
We have clear skies here and at the current sun angle/efficiency/whatever it peaks at 300 W at solar noon ~12:30.
You can see at the start on the left the battery sucks up all the power from the solar panels.
At 9:30 you can see it starts saying, "Oh, I'm so full", but keeps on minimally charging for the next 7 hours.
The middle and right rectangular peaks are 125 W loads.
-
I'm not sure I see a problem with "wasted" solar electricity, it's not like it costs anything. Surely the goal is to have sufficient energy stored in the battery at the end of the day to last until it's next sunny. How far from full is the battery when the sun goes down and how does the total stored energy at that point compare with your needs?
In your graph the red line is power, what's the green - battery voltage? A scale would be useful. More useful still would be a state of charge plot.
-
I'm not sure I see a problem with "wasted" solar electricity...
Well, yesterday I was running a minimal load.
If I start running a load before the sun comes up, I discharge more and it takes later in the day before the battery even starts charging.
Basically, that would reduce the time during the day when the battery is getting any charge by 4 hours.
The green is the array voltage, I didn't want to complicate things. It runs ~70 V.
Yeah, state of charge would be good, but the best would be an acid test. (But not on a sealed battery.)
-
If I start running a load before the sun comes up, I discharge more and it takes later in the day before the battery even starts charging.
I am not sure I follow, you run a load before the sun comes up and the battery is emptier when the sun rises. The battery will then charge from sunrise until later in the day than if you didn't run the load, as it takes longer to get to the "nearly full" point where it slows down.
Basically, that would reduce the time during the day when the battery is getting any charge by 4 hours.
I don't see how this is true unless your load is applied for four hours during daylight, and exceeds generation that whole time.
Yeah, state of charge would be good, but the best would be an acid test. (But not on a sealed battery.)
Do you at least have the ability to plot battery voltage and current?
What I am really trying to get at is that your battery charging slows down at about 9:20 am in your graph, tapering off to 30 W by midday. This is a charge current of around 2 A on a 224 Ah battery, around 0.01 C. Unless your battery is heavily sulphated I would expect that to mean it is very much full, 90-95% by that time. In that case, what's the problem? You've stored (almost) as much energy as the battery is able to, either it is enough for your needs or you need a bigger battery. Are you really worried about the last few percent? Or do you have reason to think the battery is much less charged than I am suggesting?
Don't be mislead by the fact that the battery is still drawing 2 A at sunset, lead-acid charging current does not taper down to zero when it is full (unlike lithium), instead it tapers down to a near-constant current unless you lower the charging voltage right down. Depending on your charging voltage a battery that size may draw an amp or so forever without storing any additional energy, it just gets warm.
-
If I run a 125 W load than the power curve obviously has to be greater to leave anything for the battery.
In the current circumstances, that lops off about 4 hours of the day.
Yes, I have curves for all that.
The thing that I'm trying to reconcile is that the performance doesn't jive with the simple calculations.
224 AH * 12 V = 2.5 kWH
Of course we can't take all of it, how about half?
1.25 kWH / 200 W is 6 hours.
So why I am going from 12.7 (OC) to 12.0 (16.5 A load) in a space of 2 hours?
Siging off, I don't go below 12.
-
He's not the most technically educated person but Will Prowse does some good enough research and testing of consumerish DIY solar, particularly for mobile applications (like yours?) https://www.mobile-solarpower.com/ (https://www.mobile-solarpower.com/) https://www.youtube.com/c/WillProwse/featured (https://www.youtube.com/c/WillProwse/featured)
I can't speak too much about specifics of running lead acid systems but LFP (LiFePO4) does have better lifetime cost compared to lead acid. I'm not too sure about how maximum charge rates and charging stages used in lead acid work (bulk, absorption and float) so I can't really compare but looking at LFP alone, maximum charge rates of at least 0.25C are pretty common for cells (and packs) designed for bulk solar storage. So in theory if enough power is available then they can be charged to full from flat within ~6hrs even including the "top-off" constant voltage charging stage, many batteries are available which are rated for double the charge rate. Doing deep cycling with any lead acid will kill it off relatively quick compared to lithium ion so if you're using up the full capacity daily then you really don't want to use lead acid. Especially since you're mobile, you'll be able to benefit from the lower size and weight of LFP; typically less than half the weight and size for LFP compared to deep cycle lead acid.
Failure modes in lithium-ion, even for LFP are more severe than lead acid and you can't do overvoltage equalisation as you can in lead acid. These problems are solved by a BMS system which comes integrated in most LFP battery packs designed as drop-in replacements for lead acid. A cheap way I've seen cell balancing done in some systems is a device across each cell which acts like a zener diode limiting maximum voltage and resulting in similar behaviour to overvolting lead acids for balancing but much lower leakage in lithium-ion means balance seems to be much less necessary (again not too familiar with operating lead acid systems). Most just switch in a resistor for <0.1A when imbalance over a threshold is detected and that's normally more than enough.
Cold affects both chemistries but again with failure modes being more severe in lithium-ion, cold weather (<0°C <32°F) is often disabled by competent battery management systems but is still possible to do with reduced current (like lead acid). So heaters or battery packs with integrated heaters are something you might want to consider with either chemistry if you're operating in such conditions.
LFP completely beats out lead acid in terms of overall cost and performance apart from the upfront cost for name-plate capacity but in the end you end up paying more for usable lifetime throughput if you use lead acid. The only time I see it still used is legacy systems, some standby power/UPS and otherwise poorly optimised systems.
-
So why I am going from 12.7 (OC) to 12.0 (16.5 A load) in a space of 2 hours?
Siging off, I don't go below 12.
Depending on exact chemistry and ambient temperature you might expect 12.7 V to be 100% SOC (if the battery has been sitting neither loaded nor charged). Under the same condition 12.0 V would correspond to around 50% SOC, but if measured on-load you might reach 12.0 V at 60-65 % at moderate load.
So you are saying you drop from 12.7 to 12.0 V with 200 W for 2 hours? That's around 32 Ah removed, only about 14%, and I agree it should not drop nearly so far with such a load. This could correspond to a damaged/sulphated battery (how old is it? How often has it been run to zero?). Poor charging seems less likely as you are starting at 12.7 V, but what voltage is the battery at when input power starts to drop, and what voltage is it at say 3pm when you've got plenty of incoming power but the battery seems to be as full as it gets?
What's your battery temperature?
-
The batteries are less than a year old.
The temperature here goes from 40°F - 70°F.
The batteries have never seen a terminal voltage less than 12.0 V
A normal charge sees them at 14.4 V for two hours, then down to 13.8 V.
I used to use cheap (generic big box store) deep discharge 80 AH batteries.
I used to normally use them down to 11.5 V.
They didn't do so hot.
So when I went to my shiny new (expensive) AGMs I never let them drop below 12.0 V
When they were new I could run 130 W load in the dark for 5-6 hours with impunity.
I do acknowledge that somehow the capacity of the batteries has degraded although I have tried to treat them right and live within my energy budget.
The previous graph was from this morning, in complete darkness.
The next graph is from this morning charging.
Ignoring the issue of the condition of the batteries we are still left with the fact that at the far right we are ignoring 150 W (~270 W extrapolated - ~120 W) that another chemistry could be using.
On the far left we have ~150 W constant with ~50 W cycling (the square wave).
A bit later we see just the ~50 W cycling.
You can see that the ~50 W modulation has more effect further right than far left so we can tell that we are not just looking at IR loss.
-
Doesn't that just say you could use more battery capacity?
-
I'd support the direction of what sandalcandal says above although I'd go a little further.
I've been running a leadacid setup in a van for the past 4 years and recently became aware of LiFEPO4 as an alternative. After spending a little time looking into it, I figured it was worth a punt. Now, after a few weeks hands-on experimenting, I'm convinced it's a sea-change technology.
Firstly, energy density: this is not 'a couple of times' better, It's more like 4 times. I have 3 x 120ah AGM, so in theory, 50% of rated leadacid capacity being usable = 180ah.
The LiFePO4 cells I just got, are rated 280ah, although I've measured them to hold 300ah using recommended BMS settings. Setting BMS parameters ultra conservatively, I've still got 290ah.
These cells together are smaller than just one of my leadacid units and yet deliver 1.6 x the total capacity. I make that, 4.8 x energy density in practical terms. (I know it's less in theoretical chemical terms)
Secondly Cost: I wanted to get the cells I have quickly so bought from a source warehoused in Poland (hence already inside the EU) and they arrived after 8 days. They cost a little over £600. The BMS was another £120. Compare that to the £300 I paid for the leadacid units, multiplied by the capacity increase £300 x 1.6 = £480. OK, so it looks like an initial cost of 1.5 x lead acid. So we'll be talking about the life cycle benefits to justify it however... if I were to buy the same cells from China (both sources are Chinese, the first warehouses stock in Poland), I'll have to wait 6-8 weeks for delivery but the cost saving is huge: eg These are just £364 (https://s.click.aliexpress.com/e/_9RhXBN) (I have an order from there in transit at the moment). So with the same BMS, that's a total of £484, or simply, the same cost as leadacid.
Thirdly performance: the general guide is that leadacid should be charged at about 0.1C while LiFEPO4 can be charged at 1C. Similarly, this is the max discharge rate at which they remain efficient.
So pulling 16A from a 100AH leadacid is already beyond its comfort zone (you'll be losing more than you're using), while pulling 5 x that from LiFePO4 is still in it's happy place. As I say, the capacity of my cells is rated 280ah. I'm charging them at >100A and only seeing very slight heating of the cells. Gotta make sure those busbars are screwed down tight though!!
I can draw 100A from the LifePO4 for an hour and use 100AH from the capacity.
I wouldn't be able to draw 100A from my leadacid set for more than about 20-30 mins before it needed 'a rest'. I'd be surprised if I got as much as 45 mins out of it in total at that discharge rate.
Fourthly, Life Cycle: Typical life cycle...
leadacid = 4-700 cycles
LiFePO4 = 2-4000 cycles
'nuff said.
Finally safety v convenience: Lithium batteries have a rep for bursting into flames, exploding etc. The number of times, I've mentioned my 'experiments' to people and had a response like, "better invest in a decent fire extinguisher". But 'those' lithium batteries are not LiFePO4. Some of them have higher energy density, yes, but are less stable chemistries and hence come with associated risks. LiFePO4 has the perfect balance to be the game-changer (IMHO) because of it's massively increased energy density while being no more volatile than Leadacid. The limits are more strict though so while more factors tend to diminish the performance and life cycle of leadacid, exceeding the ultimate boundaries of LiFePO4 is more likely to kill the cell entirely (hence using a BMS to look after it).
In summary: the balance has shifted - LiFePO4 is just so much better - I'm done with leadacid :-+
-
Doesn't that just say you could use more battery capacity?
Yes, I need more capacity, but that boils down to two issues:
1) What is the chemistry that can "make hay while the sun shines" and "leave nothing on the table"?
2) Why did my shiny new batteries lose 2/3 of their capacity in a year when I didn't abuse them?
I'm done with leadacid
Sounds good to me.
I had to have a tray welded in to take two GC2s instead of a BCI 27 series battery.
I was thinking of one of those enormous 12 V 200 AH beasts that have two handles.
Still, I'd have to grind out the old tray, bend and weld in a new tray.
I'd have no problem doing that but I don't have a plasma cutter, a brake or a welder so I'd have to pay people. :(
-
Doesn't that just say you could use more battery capacity?
Yes, I need more capacity, but that boils down to two issues:
1) What is the chemistry that can "make hay while the sun shines" and "leave nothing on the table"?
If I understand the question correctly you want a battery chemistry that allows you to maintain the full charge current all the way to termination? NiCd or NiMh might come close but reliably detecting end of charge state is not at all easy.
The obvious answer is to have excess storage capacity i.e. enough capacity to prevent the batteries reaching the final slow part of the charging cycle most of the time. Lithium Ion could be a good candidate for this as you can significantly extend cycle life by stopping before you reach the CV stage (at the expense of some capacity).
-
From my experience on large UPS systems (1000+ kVA), I can tell you that AGM is the worst possible battery technology for cycling applications. Sure, they're 'maintenance free' but that's because you can't even look at the electrolyte, much less do anything about it. In this case 'maintenance free' means 'maintenance not possible'.
For 1000 years (give or take a couple) utilities have been using lead acid batteries. Telephone systems, electrical switching, backup at nuclear power plants - all of the important applications use lead acid - most in clear glass cases. You can see the sulfation, you can see the liquid level and you can measure the intercell resistance. The downside is the vertical clearance needed for hydrometer testing. LA battery systems take up a lot of real estate. And a verifiable maintenance program...
If anybody ever proposes many batteries in series to get to a terminal voltage (ok) and many strings in parallel to get the current (not ok), run away as fast as you can! One dead cell in one battery in just one string will leave the remainder being overloaded. Maybe you can lose one string but you probably can't lose two. There were 400+ batteries in the system, IIRC, and it was statistically unlikely that all of the batteries would be fully capable. The UPS had less than stellar performance!