After more thorough testing of the used LiFePO4 battery that I described in my previous post, I’ve reached the sad conclusion that it’s essentially a worthless piece of junk. Encouraged by stories from mobile-solarpower.com and other sources, I’d hoped this used battery might still have about 80% the capacity of new. Even the battery’s own datasheet said an ‘end of life’ battery still retains 60% of the original capacity. But I measured the usable capacity at just 27-29% of original, and with a collection of other battery problems too. It’s basically unusable. What a disappointment. It’s an expensive lesson in the risks of buying used equipment.

The trouble began immediately after I unpacked the battery and measured 0 volts at the terminals. It was completely discharged, to the point where the low-voltage cutoff of the built-in Battery Management System had disconnected the cells. The seller said it was at 12.8V prior to shipping a week earlier, and it definitely shouldn’t have self-discharged so much over just one week.

I was able to recharge the battery with an Accucel 6 charger in LiFe mode, waking up the BMS and charging the battery to about 95-100%. But it reached a “full” charging voltage much too soon, at about 14.2V after putting in only 13 Ah. That’s roughly 29% of the battery’s nominal 45 Ah capacity. Not a good sign. Also worrying: the built-in BMS appeared to kick in at a charging voltage about 14.2-14.3V and disconnect the cells to prevent overcharging, which caused my charger to report an error. This contradicts the battery’s datasheet which says 14.6V is normal and the BMS won’t kick in until 15.6V.

After disconnecting the charger, I measured an open circuit voltage of 13.664V. Over the next day the voltage continuously dropped, though the rate of decay eventually slowed. At 20 hours post-charging, the voltage had dropped to 13.349V, or 3.337V per cell. According to this chart, this indicates a state of charge greater than 90%.

To perform a capacity test, I connected a 12V LED light with a measured current draw of 0.94 amps, and timed how long the battery survived. The setup is shown in the title photo above. During the test I also periodically measured the battery voltage. The battery lasted 13.5 hours, for a total measured capacity of 12.7 Ah or 29% of nominal capacity. That’s a very disappointing result, but at least it was consistent with the charging results. Here’s the discharge curve:

Not only did I get just 12.7 Ah instead of closer to 45 Ah, but there was another problem too. Five minutes after I measured the last data point (806 minutes, 12.259V) the battery completely shut off. This must have been the low-voltage cutoff of the BMS again, except it happened much too soon. The datasheet says a normal discharge continues down to 10.0V, and the BMS doesn’t engage until 8.8V! The discharge curve is very steep at this point, so I probably only missed out on about 5% of extra capacity, but it’s another indicator that things weren’t working correctly.

As a final test, the next day I recharged the battery a second time at just 1 amp charging current – C/45. That’s a very conservative rate, because I wanted to be sure I hadn’t somehow failed to actually charge the battery fully on the first time around. The results were even worse than before, and I only managed to put 12.1 Ah into the battery before it reached full, 27% of nominal capacity. After disconnecting the charger, I measured an open circuit voltage of 13.6V, which indicates a fully charged battery. So there’s no mistake… it’s just a dud.

So now I’m left with a $167.50 used battery with vastly less than the expected capacity, apparent BMS problems, and a possible self-discharge problem too. This seller didn’t make any specific performance claims other than that the battery came from “well-maintained equipment”, but I don’t think they intended to sell a battery in this poor state. If they were aware of the battery’s problems, I think an honest seller would have thrown it in the trash, or perhaps sold it with a warning like “for parts only – battery does not hold a full charge”.

I’ve contacted the seller to request a return or a refund. No reply yet. My strategy in such things is to explain nicely what I’m seeing, and ask politely for help, always working from the assumption that it’s a good faith misunderstanding. In my experience this works much better than angry rants and accusations.

Have you ever bought a used battery? How did it work out? What information might you look for when choosing a used battery to avoid problems like this? Leave a comment and share your thoughts.

My used LiFePO4 battery (lithium iron phosphate) arrived yesterday. Who buys used batteries? I do, apparently. I’ve been putting the battery through its paces to see how much of the original capacity of 45 Ah / 675 Wh still remains. Unfortunately it looks like my battery may be a dud.

 
Where are the Battery Terminals?

My first challenge was simply connecting anything to the battery terminals. This battery is supposed to use female threaded button terminals, which are basically just M6 x 10mm steel bolts that screw into the battery, but they weren’t included. The only thing I have to work with are 6mm female threaded openings. For taking measurements with a multimeter, I can hold the probes to the terminals, but there’s no way to make a permanent connection for charging or discharging.

With all the local stores closed and everyone in California ordered to stay home, I can’t just run out to Home Depot for M6 bolts. All of my attempts at jury-rigged connections failed. I finally had to resort to sticking a loose piece of wire into the terminal opening, but it’s a poor connection.

 
Battery Charging

Initially I was unable to measure any voltage at all between the battery terminals. 0 volts! I contacted the seller, who reported that the battery was at 12.8V when he shipped it a week ago. So what happened?

Knowing that the battery has an integrated BMS (battery management system), I suspected that the BMS might have disconnected the internal battery cells to protect them due to over-discharging or some other fault. From the back of a closet I dug out an old Accucel 6 battery charger from my RC aircraft days, and used it to feed in constant current at 0.5A. The battery behaved like it was charging, starting at a voltage around 10V and quickly ramping up to 11-12V. But whenever I disconnected the charger and tried to measure the open circuit battery voltage, I saw something around 6V that quickly died away to zero.

I decided that maybe I just hadn’t charged the battery enough to convince the BMS to reconnect, so I kept going. After putting about 500 mAh into the battery (about 1% of the battery’s theoretical capacity), the BMS “woke up” and I could measure a stable open circuit voltage at the terminals. I was also able to power some 12V LED light strips for about 15 minutes, until I grew bored of the test.

Something is puzzling here, though. LiFePO4 batteries have a very low self-discharge rate. They can last for months on the shelf. If my battery was at 12.8V last week, it shouldn’t have dropped so far in a week to engage the BMS low-voltage cutoff of 10V or less. Hmm.

I continued charging the battery with my Accucel 6, now at a current of 3.8 amps. All went well for a few hours, and I put about 13 Ah into the battery, which should be about 25-30% of its capacity. But then the voltage quickly shot up to around 14.2V and the Accucel aborted with a “connection break” error. I checked all the connections, and tried to resume charging several times, but kept getting “connection break” or “over voltage” errors. The battery wasn’t actually over voltage, and the spec sheet says to use constant current until the voltage reaches 14.6V, then hold that voltage until current decreases to 1 amp.

No matter what I tried, I couldn’t get the battery to charge any further. Connection problem? Charger problem? Battery problem? Starting from a dead battery, it can’t have been full already after putting in only 13 Ah. Perhaps my poor terminal connection is the cause. Or maybe it’s the Accucel 6, which at 3.8 amps and 14.2 volts is operating beyond its 50 watt rating. Or maybe it’s the 15 volt power supply for the Accucel 6. I’m not sure whether it uses DC-to-DC conversion to enable charging batteries at a voltage higher than the supply voltage.

Without a way to reliably charge the battery to 100%, I can’t do a discharge test, so I can’t measure the capacity of the battery.

 
Self-Discharge

Defeated for the moment, I gave up on charging and disconnected the charger. I measured the open circuit voltage at 13.664V, or 3.416V/cell. According to the state of charge table for LiFePO4 cells, that indicates it was more than 90% charged, close to 100%. But how could that be, after starting from zero and only putting in 25-30%? A few minutes later I measured the voltage again, and it had dropped a tenth of a volt. Even with nothing connected to the battery, the voltage kept dropping over the next few hours:

Divide these numbers by 4 to get the cell voltage, and compare them to this LiFePO4 SOC data. Is this normal? Maybe I need to wait a few days to see how much further the voltage drops, before I can be sure.

I’m guessing this battery may have an internal short circuit somewhere that’s causing it to discharge even when it’s sitting unused and disconnected. This would explain why my open circuit voltage measurements are falling so quickly, and why the battery dropped from 12.8V to below the low-voltage cutoff during a week of shipping time. If I’m right, unfortunately that makes the battery a useless $167.50 lump of lithium and plastic.

Is there a better explanation here? Maybe I’ve overlooked something? If you have experience with deep cycle batteries, especially lithium iron phosphate batteries, please leave a comment and share your thoughts.

What’s the most expensive component in a small off-grid solar setup? You might assume it would be the solar panels, but the cost of panels can be surprisingly low. For many people, the batteries are the single most expensive element in the system. So what batteries offer the best value?

Lead acid batteries and lithium batteries are both popular options for solar energy storage. When discussing solar, it’s important to note that “lithium” usually means lithium iron phosphate – LiFePO4. This particular lithium battery chemistry is different from the “lithium ion” batteries commonly used in laptops and mobile phones, which is typically something like lithium cobalt or lithium manganese. It’s also different from the LiPo (lithium polymer) batteries used in applications like RC cars and airplanes. While these can both be used with solar, LiFePO4 works at a different voltage level that’s a better match for common 12V equipment. It’s also safer and less explosion-prone.

From a technical standpoint, LiFePO4 batteries are superior to lead acid batteries in nearly every respect.

Usable Capacity – A deep cycle lead acid battery can be regularly discharged to a depth of about 50%. Frequent discharges below that level will harm the battery and shorten its lifetime. If I’ve got a 100 Ah lead acid battery, only about 50 Ah is usable in practice. LiFePO4 batteries can be regularly discharged to a depth of 80-100%.

Lifetime Charge Cycles – When cared for properly, a sealed lead acid battery should last about 600 charge/discharge cycles. If charged and discharged every day with a solar setup, that’s a lifetime of less than two years. A LiFePO4 battery should last about 4000 charge/discharge cycles, for a lifetime of roughly 11 years with daily use.

Charging Efficiency – A charge/discharge cycle with a lead acid battery is about 85-90% efficient. I’ll get out about 85-90% of the power I put in. A LiFePO4 battery is 99+% efficient.

High Current Capacity Loss – A typical “35 Ah” lead acid battery is usually measured at a C/20 discharge rate: depleting the battery in 20 hours at a rate of 1.75 amps. Discharging the battery at a faster rate than C/20 will reduce the battery’s effective capacity. At a C/5 discharge rate of 7 amps, the effective capacity is reduced to only 28 Ah. A LiFePO4 battery is much less sensitive to the discharge rate, and will provide close to the rated capacity even at C/2 or faster.

Weight – A lead acid battery weighs more than twice as much as a LiFePO4 battery with the same Ah rating.

 
Hunting for Battery Bargains

If somebody’s giving me a free battery, then LiFePO4 is the obvious choice. But is it worth the extra cost? Hunting Amazon for inexpensive choices, I found this Weize 35 Ah 12 volt lead acid battery that’s currently $63. I couldn’t find a 35 Ah LiFePO4 battery, but I could buy three of these TalentCell 12 Ah 12 volt LiFePO4 batteries for a total of $249. That’s about 4x more expensive. Ouch.

To compare the options, instead of measuring the battery capacities in amp hours, it’s helpful to measure capacity in watt hours (volts times amp hours) instead. This is because lead acid and LiFePO4 batteries operate at slightly different voltages. Then I can compare the cost per watt hour of nameplate battery capacity.

lead acid 35 Ah x 12 volts is 420 Wh, $0.15/Wh
LifePO4 12 Ah x 12.8 volts is 153.6 Wh, $0.54/Wh

The simple difference in upfront cost per Wh is 3.6x more for LiFePO4, then. But what about the differences in usable capacity, lifetime charge cycles, and charging efficiency? Let’s calculate the cost per actual usable Wh over the lifetime of the battery.

lead acid 420 Wh x 600 lifetime cycles x 50% per cycle x 90% efficiency = 113.4 kWh = $0.55/kWh lifetime
LiFePO4 153.6 Wh x 4000 lifetime cycles x 80% per cycle x 99% efficiency = 486.6 kWh = $0.17/kWh lifetime

Over the battery’s lifetime, the LiFePO4 battery represents a much better value. My cost to charge and discharge a kWh is less than one third the cost with a lead acid battery.

Still, there’s that painful upfront cost. Can I do even better?

 
Used LiFePO4 Batteries

For the home DIY enthusiast and backyard tinkerer, used LiFePO4 batteries are an interesting prospect. If you’re like I was, the concept of buying used batteries may seem about as desirable as buying used toothbrushes. However, with the very long lifetimes of LiFePO4 batteries, pre-owned batteries can make good economic sense.

Many LiFePO4 batteries are used in medical and industrial applications where they’re replaced on a fixed schedule every 2-3 years. These batteries are probably not used every day, nor routinely discharged fully. A gently used three year old LiFePO4 battery might only have 500 cycles on it, and still retain 80% or more of its originally rated capacity.

After scouring eBay for a few days, I purchased this used 45 Ah / 576 Wh LiFePO4 battery for $167.50 shipped. It advertised “scheduled removal from well maintained equipment”. I’m taking a chance of course, but if this battery retains 80% of its original capacity, then $167.50 for what’s effectively a high-quality 36 Ah / 461 Wh LiFePO4 is a bargain. At $0.36/Wh it’s about two-thirds the cost of el-cheapo brand LiFePO4 batteries available new.

The only puzzling thing about this battery is the maximum charge and discharge rate of 20 amps. While it’s more than I ever anticipate needing, it seems curiously low for a 45 Ah LiFePO4 battery, representing a charge/discharge rate of 0.44C. From the reading I’ve done, a LiFePO4 battery should easily support 1C discharge and probably even higher.

I’ll report more about this used battery once I get my hands on it.

While sheltering in place this week, I’ve been tinkering with a small-scale setup for solar power generation. I’ve got a 100 watt solar panel, and access to the sun. What fun things can I do with this? Is it actually useful? Let’s find out.

My first thought was to power some equipment directly from the panel, but that’s not practical for most situations. Even if I could tolerate only having power during daylight hours, the output voltage and available power from a solar panel fluctuates too much from moment to moment. I’d either need a DC-to-DC voltage converter with a wide input range, or some equipment like a pump that can tolerate a wide voltage range and doesn’t mind frequent stopping and starting.

For most purposes it’s better to charge a battery from a solar panel, and then use the battery to power other equipment. I already have a solar generator (a large battery with integrated charger, inverter, and other conveniences) that was ideal for this experiment. I only needed to connect the panel’s MC4 output to the solar generator’s MC4 input adapter cable, stick it in sunlight, and wait.

 
100 Watts? Not So Much

With a 100 watt panel and something close to 12 hours of daily sunlight, I expected to get something close to 1200 watt-hours of electric production daily. My solar generator has a 150 Wh battery, so it should only take about 1.5 or 2 hours to charge. So I confidently set up the equipment, and after an entire day in the sun I only managed to increase the battery level by about 40%. What?

Maybe 1200 Wh was a little unrealistic. Or a LOT unrealistic. After some reading, I concluded the panel would probably never output 100W unless it was noon on a bright sunny day somewhere near the equator. But I might hope to get about 70W at noon at my latitude, with lower power output during the morning and late afternoon. Factoring in shadows from trees and other buildings, I decided I might expect to get about 400 Wh of average total daily production, with more in summer and less in winter.

OK then, 400 Wh should still be enough to charge my solar generator’s battery almost three times during the course of a day. So why wasn’t I getting that result?

 
Measuring Solar Panel Output

It’s not so easy to measure the power generated by a solar panel. With a multimeter I could measure the open circuit voltage, and the short circuit current, but multiplying the two figures wouldn’t tell me the power. I need a load to get a useful measurement for power output. But a fixed resistive load won’t work, not even a 100W-rated resistor, because it likely won’t bring the solar panel to the correct voltage for optimum operation. That optimum voltage varies from moment to moment, based on the sunlight hitting the panel. To do this right, I needed a solar charger like the one integrated into my solar generator. Then I needed to measure the current and the voltage simultaneously. I could have built some wiring adapters and used two meters for the measurements, but instead I bought a cheap inline power meter and soldered MC4 connectors to it.

I measured 19.7W in full sun at noon. Huh?! No wonder the solar generator’s 150 Wh battery takes forever to charge. Is there something wrong with my panel? After several days of tinkering with the setup under different lighting conditions, I never saw a continuous output higher than 23W. Most of the time it hovered right around 20W. Hmm.

I began to suspect the solar generator was at fault. Sure enough, buried in the manual were the specs for the solar input: 13V-22V / 2A max. With my solar panel, that means I’m theoretically limited to about 40W max (2A at almost 20V). I’m not sure why I rarely saw more than 20W though, and never saw more than about 1.3A of current. Maybe the integrated solar charger is even more limited than the manual suggests? Maybe I have bad wiring, or another problem?

As an engineer, 20W from this panel is insulting! Even if I have no practical need for this solar panel, losing 80% of its output is unacceptable to me. To save my pride I’ve begun to research plan B, which will involve a stand-alone solar charge controller featuring a much higher maximum charging rate, and a separate battery. More about that soon. Maybe I’ll put together a solar-powered Mac Plus.

 
Finding the Parts

Some links below may be affiliate links. BMOW may get paid if you buy something or take an action after clicking one of these. As an Amazon Associate BMOW earns from qualifying purchases.

Here’s the equipment I used.

Renogy 100W monocrystalline solar panel. You can find slightly cheaper panels, but the Renogy has an extra-sturdy aluminum frame and a strong reputation for quality. This particular panel is also more space-efficient than most other 100W panels, if minimizing area and weight are important to you.

Suaoki 150Wh portable power station. Despite its annoying 2A charge rate limitation, I love this thing and use it all the time. You can charge it from solar, from a car, or from a wall plug. It has a built-in 100W inverter for running small appliances, USB ports (including a Quickcharge 3.0 port) for charging phones and tablets, 12V ports for DC lights and other appliances, and an integrated high-brightness emergency lamp.

200A Inline Watt Meter. I’m not confident it will actually handle 200A, but it works nicely for lower currents involved in small-scale solar. It shows volts, amps, watts, accumulated amp-hours, watt-hours, max watts, and min volts.

MC4 Male/Female Solar Panel Cable Connectors. Solder these to the watt meter.

DROK 12V Battery Meter with adjustable limits. Other cheap battery meters typically have fixed voltages for the 100% and 0% charge state. This one makes it possible to set custom values for the upper and lower bounds. Measured from one of the Suaoki’s 12V ports, I measured 12.33V fully charged and 8.96V just before the low-voltage cutoff disconnected the battery. The discharge curve isn’t linear, so the meter won’t go smoothly from 100% to 0%, but this is still vastly better than the simple built-in 5-bar power gauge on the Suaoki.

Recently I’ve been looking at battery datasheets, in preparation for an off-grid solar project. I’ve noticed something strange about the “constant power discharge” numbers in the datasheets of several 12V lead acid batteries. Here’s an example from a 20 Ah Euroglobe sealed lead acid battery.

In the Constant Current Discharge table, if you discharge to a final voltage of 1.80V/cell (10.8V total voltage), the entry circled in yellow shows that you can get a current of 1.00 amps over 20 hours. The voltage will drop from around 13V down to 10.8V during that time. Let’s call it an average of 12V times 1A – that means you can average about 12 watts for 20 hours.

But wait. In the Constant Power Discharge table, if you discharge to a final voltage of 1.80V/cell over 20 hours, the entry circled in yellow shows a power of just 1.98 watts. That’s far less than 12 watts. Why?

Other table entries show something similar. It’s 11.3 amps constant current for 1 hour – that should be an average rate of about 136 watts, but the Constant Power Discharge table shows a measly 21.6 watts. It’s not just this particular battery either. Here’s a 35 Ah lead acid Mighty Max battery that shows the same curious pattern in the Constant Power Discharge table.

So what’s going on here? Am I misunderstanding what these tables mean? Or is there some other factor that limits the power to a much lower number than is suggested by the constant current data? I’ll keep digging for answers.

My family and I have been ordered to shelter at home until at least April 7, to help stop transmission of COVID-19 in the San Francisco area. Many of you may already be living under similar orders, or will be shortly. Travel is restricted to only the “most essential needs” – basically food and health care.

As you can imagine, this will severely impact business shipments. BMOW will still be accepting new orders during this time, but it will likely be impossible to ship anything until April 7 or later. Please be patient, and if you’re not prepared to wait at least 3-4 weeks for delivery, then please hold off until mid-April to place your order.

Stay safe everyone. Remember to wash your hands.