LiFePO4 - The future for off-grid battery banks?

[CrazyJerry]

Thanks Simon. Recently I stumbled across this forum and then discovered that there wasn't a lot of historical data with lipo in the off grid systems - so maybe some folks can benefit from my experiences.

Upfront, my standard disclaimer is:

1) I'm not selling anything. Not a business owner. Most people I deal with have very, very, limited funds and are off-grid. This makes us very accountable for the energy we use.

2) When referencing items I would only use a name brand as a specific example so there's some context. “Generically speaking” works in certain situations but actual names or part numbers can be very helpful when researching or validating claims. Moderators please feel free to edit out if necessary.

3) Based on the ongoing 13 year experiment here at 9 Main, I call'em like I see'em. If something works I share that, if something is sub-par that also is shared. If I've destroyed something due to ignorance or any other reason, I'm happy to include that also to provide context.

4) Anyone who reads what I've written should also note that I have zero responsibility if you try any of this. My experience is that some people "skim" or "thumb" over information and then find themselves in a pickle because one or more important "details" were missing.. Aside from the safety aspect, operating in this manner can deplete a piggy bank fairly quick so understand the risks and benefits, then proceed with caution. Onward ho!

GBS 20ah Lithium batteries: Part 1.

The batteries referenced a couple of posts back were purchased in March 2011 for a transportation experiment and actually put into service on April 4th, 2011. Listed as LiFeMnPO4 chemistry, they were originally purchased online (from GBS) as four pre-assembled 12 volt 20 ah packs that were then assembled in series totaling a larger 20 amp hour main pack at 48 volts. This was for a Go-One3 velomobile that I rebuilt and then "electrified". These batteries were used hard for the first 6 months (all summer) every day in that trike. The batteries were mounted in enclosed saddlebags on each side of the trike - so it got a little warm in there at times – partly through use and party from baking in the sun.

The charging and discharging regime for those first 6 months were dictated by the spec sheet on the batteries / charger /balance boards - EX: charge them up to the point at which point the balance boards would come on: 3.7 volts per cell.

In doing so, I did notice the warm feeling in the saddlebags where the batteries were. To try and get more data as a baseline, I used an infrared heat gun at the terminal posts, the temperature on all 16 cells in the 48 volt series pack would climb from about 75 degrees starting temp - to 96 or so at the end of the complete charge cycle. NOTE: This would be after using the trike with the pack voltage depleted and then resting at 48-48.5 volts (roughly 3 volts per cell). The recharge time would take a few hours with the supplied 6 amp output charger - significantly less time (about 1 hour and 15 mins using 600 watts of solar - more on that in a bit..)

As a side note, just about a month or so into the project (still 2011), I ended up purchasing a second identical pack of these batteries (so I could rotate them), and noticed the stark difference at the point the new balance boards came on - . Seemed the early balance boards were now reconfigured and the kick on point for the balance function was indeed lowered to just 3.55 volts per cell. I still have the 2011 spec sheets as noted above whereby the balancing function is listed to begin at 3.7 volts per cell... There must've been a reason for the big change..

Via the GBS website, the spec on the batteries changed as well. No longer was the recommendation to keep the voltage on these batteries between 2.80-3.80 volts per cell. As of this writing, GBS lists their 20ah cells operating voltage range at: 2.8-3.6 volts. The change is only to the upper end of the voltage range.
At that point in the experiment, I had two identical packs consisting of 16 cells each (48volts total / 20 amp hours) - but the balance boards engage at different set points - what to do?

Using the infrared heat gun on the new pack with the lower charge voltage / balance boards, I notice a lot less heat at the terminals. In most cases it wasn’t exceeding 82 degrees - so assuming the second set of batteries were manufactured exactly the same as pack #1, I now have a measurable difference between them based on a voltage charge recommendation. I opted to go with the lower charging voltage simply for the heat reduction.

Discharge on the trike was within GBS 20ah cell recommendations - that is until such number was also reduced. The 2011 spec on these cells for max continuous discharge was 3C (60 amps) and pulse at 10C. Today, their website lists the same battery max continuous discharge current at 2C (40amps) and 10C impulse. I largely ignored this new spec and set the trike's controller to deliver a “compromise” at 50amps max to the rear hub, and that happened quite frequently - mostly because at only 110 lbs – the trike would zip right along! It was a blast. Many times the ending voltage after a 50 or so mile trip would be around 50-51 volts. There were a couple of 100 milers where the ending voltage was 46-48 volts resting.

Early on, charging these cells was also an experiment. I could (and did) do this from a variety of ways. The ac 110 volt input (GBS supplied) charger that came with the batteries was the easiest - but slow at only 6 amps output. Piping 600 watts of solar into an Outback MX-60 MPPT controller was considerably faster and to be the main "charging station" for the Go-One3. (These controllers are discontinued and Outback now uses the FM60 as the new model). Programming the MX-60 to mimic the ac charger for the batteries was pretty easy. I set bulk to 59 volts, the absorb setting for one hour, and end amp current set to 2 amps. So when the sun was shining, 600 watts of solar went into the MX-60 controller and then would raise battery pack voltage to 59 volts. The controller would then display the absorb mode and hold the voltage. During the absorb 1 hour timeframe, the charging amps would taper down. When they reached just 2 amps the controller would end the absorb stage and then go to float (float was set to 13.2 - basically terminating any further charging). This would generally happen before the 1 hour absorb timer ended. It was a pretty slick setup - the off-grid home system was now providing the fuel for the 30 mile round trip work commute and plenty of other trips too - very cool indeed. With the ability to quickly swap out a depleted pack for a fresh one, the Go-One3 was already ready to go!

Another charge source was introduced for emergencies. I assembled my version of a regenerative brake system into the Go-One3. Using the 3 phase wires (from the Crystalyte rear hub) through a large alternator rectifier, gave me direct current to play with - but the voltage wasn't high enough to regen into the batteries unless I was going down an olympic ski jump at 60 miles an hour. I needed a way to step the voltage up and I found a unit that did exactly that. Incorporating that into the system allowed for a manually engaged brake regen unit at much lower speeds - very useful. For emergencies/roadside charging, piping any DC source of 7 volts - 48 volts (that has some ooomph behind it) into the regen auxiliary plug, would be stepped up to, and not beyond, 56 volts for the batteries. So, in a pinch I could easily charge from any 12 volt car battery. The 20ah cells were now being pelted from a variety of charging sources and sometimes during extreme scenarios as in the case of the regen unit in the hills. I seldom had to use the regular cable operated brakes – the drag from regen worked excellent as a brake..

This was just the beginning of this lithium journey. These batteries showed promise but the need to expand quickly set in. They needed to do more....

To be continued in Part II!

~CrazyJerry
The Aurora project (velomobile) webpage for summer of 2011 is available at the following weblink:

http://diesel-bike.com/Aurora/aurora.html

[Willy T]

We all bemoan the anecdotal bench tests that are out there as not anything you can hang your hat on, it's just as bad trying to decipher the data from regular users. I reached my one year anniversary this week and it's milestone if the fact they made it a year makes a difference. I consider all my data worthless for a couple reasons, one the temperature exposure during the summer was above 90° and I have now made the decision to put them into a storage mode this winter.

The basic internet contention is to drop the SOC to 50-60% and leave them alone to self discharge. It then becomes that 5 months could present some problems in a remote location and un-monitored. The choices are, remove everything and let the cells sit naked, maintain a low float voltage or use a VCS and have a low set point and raise the voltage to a high point within a given range.

I see any meaningful data as needing lot's of footnotes, calendar years and cycles are nice to look at, but having a Capacity tests along the way will be the most important , assuming that there is not a rash of cell failures.

[karrak]

Throwing in a current calendar / cycle example here: 1650 + days/cycles on bank "A" (LFP) and still performing quite nicely. These are the smaller versions - but many - that you are using PNjunction. It's never boring being part of a living experiment - although there's always a few eyebrow raising moments to keep complacency from taking over!

Congratulations, you are one of the very early adopters of LFP batteries that is still commenting online. I don't know of any older systems.

I would be interested to know more information on your system, what sort of battery management/checking you use, what voltage or range of SOC you keep you battery between, what your experience with battery balance has been, whether you have noticed any change in the the performance of the battery over the four and a half years you have had it, and of course what your eyebrow raising moments have been. Not asking for much am I?

I think complacency is probably a major killer of batteries of all types.

Simon

[CrazyJerry]

I sincerely wish him the best, but it is wise to remember that NONE OF US has demonstrated a real-world calendrical lifespan reaching the mythical 2000+ cycles (ie daily cycle, not hammering them repeatedly). Ie, nobody I know that uses LFP has ever said something like "yeah, I'm on my second bank now - my first one lasted 8 years and about 2000 cycles, so I just replaced it with the same". If a cell / bank that is sitting around between 40-80% SOC on a float-standby basis, that doesn't count for us daily-cyclers.

Thing is, instead of just talking about it, get some cells from Dax, put it together and join the fun for a small experimental bank. Or any of the prismatics. 12v even with a simple 4S setup!

I tell ya' lurkers, actually doing it physically beats the pants off talking about it.

The problem is that there will never be a one-size-fits-all solution with LFP, and thus a never ending thread.

Very well said. Throwing in a current calendar / cycle example here: 1650 + days/cycles on bank "A" (LFP) and still performing quite nicely. These are the smaller versions - but many - that you are using PNjunction. It's never boring being part of a living experiment - although there's always a few eyebrow raising moments to keep complacency from taking over!
~CrazyJerry

Edit: Actual number or days / cycles on these batteries is 1,670, so, in 330 days from today they will hit the 2,000 cycle (and days) mark. There are 32 of the 20ah cells configured to 12 volt and 160ah. They are predominantly charged from solar.

[SunEagle]

Wut? I'm going with Hoffa.

Nah. He is in the NJ Medowlands under Giants stadium.

[PNjunction]

I asked them what type of LFP - "CALB LFP cells". Hard to tell how reputable they are but to my eyes it looks like a well built, well thought out product. Still, trick or treat, who knows.

Exactly. Are they the new-old-stock original formula SE "blue" CALB cells to save costs? The question here is exactly what models are inside, and who is their supplier? What is their manufacturing relationship with Calb?

The blue cells, while ok for our purposes, are getting long in the tooth if they are using those.

[vudu]

Huh? When I taught urchins in a community college, I liked pointing out the morons so they would leave and quit wasting my and other students time and money with one simple question they cannot answer correctly. their eyes just glaze over and roll into the back of their heads. Only to find out later after being laughed at for being stupid.

Who is buried in Grant's Tomb?

Wut? I'm going with Hoffa.

[DanKegel]

When I taught urchins in a community college, I liked pointing out the morons so they would leave

I see old habits die hard.

[Sunking]

https://youtu.be/O1jFUHcYC3U?t=5m41s

I asked them what type of LFP - "CALB LFP cells".

Huh? When I taught urchins in a community college, I liked pointing out the morons so they would leave and quit wasting my and other students time and money with one simple question they cannot answer correctly. their eyes just glaze over and roll into the back of their heads. Only to find out later after being laughed at for being stupid.

Who is buried in Grant's Tomb?

[vudu]

CALB LFP cells

https://youtu.be/O1jFUHcYC3U?t=5m41s

I asked them what type of LFP - "CALB LFP cells". Hard to tell how reputable they are but to my eyes it looks like a well built, well thought out product. Still, trick or treat, who knows.

[PNjunction]

I don't know what that video has to do with dax, if you go to the company website mentioned in the video you will find they are based in the US.

It just shows that I'm not picking on him personally, I'm just pointing out that the common element in both cases is buying things sight unseen. It has nothing to do with location, or differences in quality.

I understand from a business standpoint that one wants to have the flexibility of changing your supplier without having to deliver those details to the customer. For instance, some powersports LFP battery suppliers used A123, and people thought they were buying those internally based on previous sales. But lo and behold, a year or two later when they built up a name, they are no longer using A123, but something else, and you have to cut one open to find out who the replacement is. For some, that brand-switch is unacceptable when it is not mentioned in detail any longer.

The extreme analogy here is buying what was supposed to be a Rolls-Surrette, but a few years later, the Costco battery shows up at your door instead. Since we are discussing a *proposed* retail business and not the current private contractor biz, analogies are all we have to go by.

[lkruper]

I don't know what that video has to do with dax, if you go to the company website mentioned in the video you will find they are based in the US. They must be doing something right if their customers include NASA, all branches of the US military, Lockheed Martin etc.

Simon

I guess you missed my post regarding some items paid for by the US military like $600+ toilet seats. Just because the Military buys something does not mean that it was cost effective, intended for or reliable enough for national security.

[Living Large]

I don't know what that video has to do with dax, if you go to the company website mentioned in the video you will find they are based in the US. They must be doing something right if their customers include NASA, all branches of the US military, Lockheed Martin etc.

Simon

At a glance they may appear professional and even reputable, but then you see a 5th grade grammar error such as: "If your looking for the latest portable solar panels." Fail.

[karrak]

Still, this is what I'm talking about. And to be fair to Dax, here goes:

We don't know what the cells actually are, nor do we know what the "mppt technology" controller really is, nor do we know what sort of dc/ac inverter they are using. All of it is sight-unseen except for some pretty portable cases. I want to know what's inside before buying.

In addition, pics of some of the panels just simply ground mounted with clumps of grass across the bottom throwing shadows shows that they (or maybe just the photographers) don't know that you need to keep shadows off mono/polycrystalline panels. They really should have cleaned up in front of some of those shots. Unwary users will just plunk them down on the lawn like in the photos, and then wonder why nothing seems to charge well.

Uh, yeeeaaah, one could do much better by constructing a kit of their own.

I don't know what that video has to do with dax, if you go to the company website mentioned in the video you will find they are based in the US. They must be doing something right if their customers include NASA, all branches of the US military, Lockheed Martin etc.

Simon

[vudu]

That .be domain-hack kind of freaks me out.

It's the link given by youtube when you share the video at a certain spot rather than copying the url. You save a whole character or something.


I'm not even entertaining buying it. The details are quite opaque.

The "large format" is what caught my attention in regard to this thread.

Actually, there is a lot to learn from a "learner project" like this, so if you take notes, it isn't money wasted. Perhaps you can still go lead-acid with a redesign.

Probably so but as a learner project, I'd just assume start with LFP. I've learned :P Ha, we'll see, thanks again for your insights. Paul