Lifepo4 replacement quandary

mcgivor · July 2018

Having spent much time reading posts on LiPo4 batteries, I've come to the conclusion that's the way I want to progress away from LA technology. Reading between the lines it would seem prysmatic cells are the way to go, the search for cells is overwhelming, I've been in contact with suppliers many of whom are willing to sell the cells but have no BMS system designed and or supplied by them, which means an after market BMS, not that that's a problem, it just when something goes wrong each will blame the other for the failure, should it happen. There is one supplier who has an integrated battery with BMS claiming up to 4 series 4 parrallel with their 12V 180 Ah batteries each of which has its own BMS, not sure how that works, here is the link to the supplier https://www.alibaba.com/product-detail/Polinovel-Latest-HD-12v-24v-180ah_60777826187.html?spm=a2700.galleryofferlist.normalList.11.1cdc65caXXOdg0&s=p Which way to go is the question. Oh yes China supplier, but when in Asia shop in Asia.

westbranch · July 2018

they even have a 10 Kwh Power Wall just for you... https://polinovel.en.alibaba.com/product/60807112821-806855448/Power_wall_solar_system_battery_51_2V_140Ah_oem_lithium_ion_battery_10kwh.html?spm=a2700.icbuShop.82.16.28b81a5fwbK2QT

westbranch · July 2018

looks like they also had an earlier model 9kW http://www.plgenergy.com/newsitem/277057753

ps their most recent news postings are from 2015...

karrak · July 2018

I wouldn't go for an LFP battery with integrated BMS unless it was from a supplier with a well known track record. I always make the assumption when designing a system that each part of the system is likely to fail during its lifetime and ask myself what would happen if each part of the system failed. Failure of the BMS could at best make the battery inoperable which could be expensive if the supplier doesn't back their products up with a decent warranty. At worst a BMS fault could take the battery out of its safe operating region which would be a safety issue.

IMO if you are building a DIY system I think you should be very confident that the if the battery and/or BMS fails that the whole power system will fail safely.

With this in mind I have just designed and installed my third LFP system for a friend which uses individual prismatic cells with one of these BMS's. I chose this as

it is a fairly simple design based around a well know Li-ion battery monitoring chip made by Texas Instruments
Can be monitored and controlled via bluetooth or USB.
Is easy to reverse engineer
Has been working well on our E-bikes for over a year.
There is lots of user feedback, especially on the Endless Spheres Forum

I would stick to LFP prismatic cells that have a good track record. CALB and Winston prismatics being good examples.

Simon

mcgivor · July 2018

@karrak Thanks for the opinions, the integration of the BMS is a concern for the exact reasons you point out, the advantage of prysmatic cells is the ability to actually check the individual cell voltage by simply removing it from the bank, not easy with a manufactured all in one unit. The question of cell quality is always there, will look into the two mentioned, but it would seem there is a lot of rebranding or outsourcing of cells, price is another factor the quotes I've received so far are almost equal Ah to Ah equivalent ,with LA here in Thailand but given the deeper discharge capabilities cycle expectancy and no partial discharge issues makes them a bargain. However the primary reason is temperature, with ambient temperatures mostly 30°C or above batteries usually above that, LiFePO4 are supposed to tolerate higher temperatures better and cooling them would be easier given the sheer mass of LA. Lots to consider.

mcgivor · July 2018

So prysmatic cells it is, ordered them today 32 100Ah cells, will be a 4P 8S arrangement for 24V nominal, most vendors I've communicated with want to build a battery, usually using small cells in a sealed package with the BMS built-in. Did look into Winston and Calb, sent a request for a quote but never got a reply. The company I purchased from Hunan AUK has positive feedback generally from Australia, the price $2000 for 400Ah plus shipping $320. Looking into BMS systems now, it's going to be an adventure, but if it can be done, I will learn.

westbranch · July 2018

I have to assume that that is $2000 Auz??

Unable to find 100Ah cells on Website, LINK?

Dave · July 2018

I've been looking into the same thing. I called around but I found this place to be extremely helpful (for the totally ignorant me) and have a wide range of systems to choose from:

https://www.electriccarpartscompany.com/

I was speaking to Raj.

BTW, I've seen a number of people post here about going to Lifepo4 but generally they seem to be talking about using a schneider controller.

I have a Morningstar Tristar 60 MPPT that is fully programmable and I am wondering what thoughts are here about using that controller with lithium batteries? Don't mean to hijack so I can start a new thread if that would be better?

Estragon · July 2018

Lifepo4 are a lithium (hence the "Li"). They tend to be safer, and in our application, where power to weight & size ratios aren't critical, probably the best choice among the various current lithium chemistries. As for using the Tristar, as long as it can be programmed properly for the particular battery (and bms) under consideration, it should do the job IMHO.

mcgivor · August 2018

westbranch said:

I have to assume that that is $2000 Auz??
Unable to find 100Ah cells on Website, LINK?

$2000 US, wish it was Auz $, website search for prysmatic cells gets a message data update, here's link to where I found them.

https://www.alibaba.com/product-detail/Long-cycle-life-Rechargeable-100Ah-lifepo4_60725516304.html?spm=a2700.7724838.2017115.35.4f7a4665cgN8cb

Raj174 · August 2018

@mcgivor
That is an unbelievably good price. In the states the equivalent would be over $1000 more. And I know you're going to be amazed when you start using then. Did you get a charger to balance them yet? The initial balance is the most critical aspect of using a LFP bank IMO. Mine has been in service a little over 2 years now. Let me know if I can assist in any way with the installation.

Rick

mcgivor · August 2018

@Raj174
No not yet, wlii be speaking to the supplier about a BMS, will ask him, you mean a charger to charge each cell individually I assume.

Raj174 · August 2018

Yes, each cell and then the entire bank is top balanced in parallel at 3.6 to 3.65 volts. It does take some time.

mcgivor · August 2018

This is a future replacement, I'll take all the time nesesary to learn before switching over, nothing like finding out you got it wrong when you depend on it.

westbranch · August 2018

Thanks for the link. I suspected Alibaba after going to their website and going in circles for a while trying to get past bicycle battery packs etc

Raj174 · August 2018

If the cells are good, I think all will go well. You just take it one step at a time. When I setup 16 (+1 spare) 195AH cells I balanced them in parallel and then made a 16S bank. Started cycling them and checking for cell voltage deviation. I figured, I would get a BMS if increasing imbalance made it necessary. Well after about 6 months with only 10 millivolts deviation I was satisfied that imbalance was not an issue. I did get a BMS, but just to monitor individual cell voltage with high and low limits. It also monitors bank voltage with upper and lower volt limits. It's like most things that seem like rocket science but it's not. In fact they are very much like any other battery, except for the 80% DOD and the almost total lack of necessary maintenance.

Rick

karrak · August 2018

The primary job of a BMS is to make sure that no individual cell in the battery goes outside its safe operating temperature and voltage range.

There are three reasons that I can see that could cause a cell go out of its safe operating voltage range.

The battery is not balanced or the cells have differing capacities. With an LFP battery these conditions will only show up when the battery is nearly full charged or fully discharged.
The charge controller/inverter is incorrectly programmed or a fault in either means that they don't stop charging/discharging the battery when it is full or empty.
A fault in the battery or battery wiring.

A BMS can stop this occurring in a number ways

Instruct the charge controller or inverter via a hardware/software interface to stop charging/discharging when it detects that any cell has gone outside it safe operating conditions.
Raise an alarm which requires a person to intervene and shutdown the inverter or charge controller.
In the event of any cell voltage going too high stop the charge controller charging the battery by preferably disconnecting the solar panels from the controller or disconnecting the battery from the charge controller or in the event of any cell voltage going too low disconnect the inverter from the charge controller.

IMO it is best to have a BMS that uses method 1 with as a backup if the charge controller or inverter does not shutdown when instructed to resort to method 3. I think ta BMS should also raise an alarm that a fault has occurred even if the fault is resolved by the BMS.

Two other important functions that either have to be done manually or can be done automatically are monitoring the individual cell voltages and balancing the battery if it goes out of balance.

Simon

karrak · August 2018

mcgivor said:

The company I purchased from Hunan AUK has positive feedback generally from Australia, the price $2000 for 400Ah plus shipping $320. Looking into BMS systems now, it's going to be an adventure, but if it can be done, I will learn.

That is a good price, the latest batch of 100Ah CALBs i got were ~US$100ea.

Simon

karrak · August 2018

Dave said:
I have a Morningstar Tristar 60 MPPT that is fully programmable and I am wondering what thoughts are here about using that controller with lithium batteries? Don't mean to hijack so I can start a new thread if that would be better?

I had a look at the manual for the Tristar and as far as I can see it can only be stopped by sending commands to it via Ethernet or RS232. This is not necessarily a showstopper but would probably either need a BMS that is compatible with the Tristar or some micro controller like the Raspberry Pi that I use to monitor the BMS and send the appropriate commands to the Tristar. The other option would be a BMS that will preferably disconnect the solar panels or the battery from the Tristar.

Simon

karrak · August 2018

Raj174 said:

Yes, each cell and then the entire bank is top balanced in parallel at 3.6 to 3.65 volts. It does take some time.

In the past I would have agreed with this but it does need special equipment and is time consuming.

For the last two off-grid systems I have set up, rather than doing this I used a BMS that would shut down the charge controller if any of the voltages on the individual cells went too high and slowly ramped the charge voltage up starting at 51.0V for a 15S battery (3.40v/cell) and adjusted the balance of any individual cells whose voltage was higher than the average. This balancing could be done with automatic cell balancing or manually.

Do not try this unless you have a BMS that will shut down the charging automatically and do it under supervision.

Simon

mcgivor · August 2018

For a long time I've been a fly on the wall, reading much of what Simon @karrak has to say, obviously there is experience behind the information, which qualifies as knowledge. Rick @Raj174 I don't mean to take away from your input and totally understand your rationale, it makes sense, but requires a human, preferably the operator to keep watch, as the BMS. Unfortunately there's times when I'm away, for that reason a BMS that looks after my interests is of importance. There are so many to choose from, many appear to be a PC board without much else, I did find one which, although not the cheapest by a long stretch, seems to offer the features I like. One of them is drive outputs for mechanical relays/contactors with a NO contact rather than electronic switching within the BMS itself taking current readings from an external shunt, Omron has some excellent components for the purpose see pdf. Being involved with life safty equipment in the past, software is generally not an acceptable means of communication from safty standpoint to shutdown equipment, old habits die hard, not that this is life safety, unless there is a fire of course. Here is the BMS I'm considering https://www.alibaba.com/product-detail/2S-16S-100A-continuous-300A-peak_60734485241.html?spm=a2700.details.maylikehoz.13.ba31347bDCnDhF Really appreciate the input, let me know opinions on the BMS proposed.

Ivor

karrak · August 2018

mcgivor said:

Here is the BMS I'm considering https://www.alibaba.com/product-detail/2S-16S-100A-continuous-300A-peak_60734485241.html?spm=a2700.details.maylikehoz.13.ba31347bDCnDhF Really appreciate the input, let me know opinions on the BMS proposed.

This is actually a Chargery BMS16and has been around for a while.I think Chargery has a pretty good name. I would look in the Endless Spheres Forum for more information.

I agree that linked up to the Omron relays that it would do the job.

The one drawback with this BMS is that as far as I can see it doesn't output any data and can't be interrogated from an external source.

In this day and age I like to have gear that can talk to external equipment. I have got very used to being able to check up on the three systems I have installed and the one I am currently building remotely via the Internet.

This BMS doesn't do cell balancing, were you going to do this manually?

Regarding having to rely on software, you are relying on the software in the BMS working correctly. Hardware that doesn't rely on software can also fail to work when it is needed. I am a fan of having pieces of equipment being able to work autonomously but also being linked to other equipment that can check if they are working.

Simon

Dave Angelini · August 2018

Sometimes a picture is best! Hi Simon! You said it perfectly! Hi Reliability is critical if you expect the life out of these systems.
The screen is from one of the best in the world at this, in so many applications!

Dave · August 2018

karrak said:

Dave said:
I have a Morningstar Tristar 60 MPPT that is fully programmable and I am wondering what thoughts are here about using that controller with lithium batteries? Don't mean to hijack so I can start a new thread if that would be better?

I had a look at the manual for the Tristar and as far as I can see it can only be stopped by sending commands to it via Ethernet or RS232. This is not necessarily a showstopper but would probably either need a BMS that is compatible with the Tristar or some micro controller like the Raspberry Pi that I use to monitor the BMS and send the appropriate commands to the Tristar. The other option would be a BMS that will preferably disconnect the solar panels or the battery from the Tristar.

Simon

Thanks Simon, I'd definitely be using a BMS but that is more for balancing among the batteries. If I understood your comment, your issue was that the Tristar wouldn't stop the charge without an ethernet or RS232. That controller is quite programmable. I think you'd just program it to bulk up to the spec voltage (provided by manufacturer) and then set it to absorb at that voltage to complete the saturation and then drop down to whatever maintenance voltage is required or cut off completely whenever the absorption/saturation was complete. If I have this wrong I'd like to hear about it because this is one of the main issues I am trying to understand now. What sort of compatibility issues between the BMS and the Tristar would I need to worry about?

The prices McGivor shows look incredibly good. I was looking at a 48V 200Ah system with BMS and miscellaneous cabling etc and I think with shipping it was pushing $7k, which I didn't think was all that bad compared to a LA system where's I'd need around 500 Ah to be comparable and would likely be $3-3.5K.

Raj174 · August 2018

mcgivor said:

For a long time I've been a fly on the wall, reading much of what Simon @karrak has to say, obviously there is experience behind the information, which qualifies as knowledge. Rick @Raj174 I don't mean to take away from your input and totally understand your rationale, it makes sense, but requires a human, preferably the operator to keep watch, as the BMS. Unfortunately there's times when I'm away, for that reason a BMS that looks after my interests is of importance. There are so many to choose from, many appear to be a PC board without much else, I did find one which, although not the cheapest by a long stretch, seems to offer the features I like. One of them is drive outputs for mechanical relays/contactors with a NO contact rather than electronic switching within the BMS itself taking current readings from an external shunt, Omron has some excellent components for the purpose see pdf. Being involved with life safty equipment in the past, software is generally not an acceptable means of communication from safty standpoint to shutdown equipment, old habits die hard, not that this is life safety, unless there is a fire of course. Here is the BMS I'm considering https://www.alibaba.com/product-detail/2S-16S-100A-continuous-300A-peak_60734485241.html?spm=a2700.details.maylikehoz.13.ba31347bDCnDhF Really appreciate the input, let me know opinions on the BMS proposed.

Ivor

No worry, only you know what you're comfortable with. By the way, the Chargery BMS16 is the one I have and as Simon indicated, may not be what one you're looking for since it doesn't do active or inactive balancing nor even have a communication feature.

Rick

mcgivor · August 2018

So Rick other than the Chargery BMS 16 not being able to ballance or communicate with external devices how is it from a quality performance standpoint. My understanding is balancing is not a regular task, but rather a one time affair done before assembly of the bank to ensure all cells are at, or as close as possible, (please correct me if I'm wrong ) to equal SOC, so doing it manually with single cell chargers is possible, albeit time consuming. The remote communication or a access via a PC is not something that interests me to be honest, why I like the fact it has its own display.

Simon, the reference to software control was to do with the use of data to comand a function, specifically with regards to fire alarm systems, where a physical relay contact is required as a fail safe, to shut down ventilation fans for example. Direct digital control , DDC, engineers always wanted to use digital input to their systems to shut down equipment over a LAN network, not fail safe. Also the transmission of alarm signals couldn't be transmitted via a VOIP telco system, it had to be physical land line, that is changing however with the use of multiple communication means to monitor each other, VOIP and cellular as examples, times are changing and physical land lines are fast becoming a thing of the past. The use of mechanical relays to disconnect charge and discharge, as apposed to internal mosfets, or other means, would make it reletivaly easy to include a second BMS simply as a monitor of the primary system, interlocked via a second contact in the relay coil circuit. Perhaps I'm over thinking this a little, but redundancy is not nesesarally a bad thing.

karrak · August 2018

Dave said:

Thanks Simon, I'd definitely be using a BMS but that is more for balancing among the batteries. If I understood your comment, your issue was that the Tristar wouldn't stop the charge without an ethernet or RS232. That controller is quite programmable. I think you'd just program it to bulk up to the spec voltage (provided by manufacturer) and then set it to absorb at that voltage to complete the saturation and then drop down to whatever maintenance voltage is required or cut off completely whenever the absorption/saturation was complete. If I have this wrong I'd like to hear about it because this is one of the main issues I am trying to understand now. What sort of compatibility issues between the BMS and the Tristar would I need to worry about?

For everyday operation if the battery is balanced and there are no faults or stuffups you are right that just setting the bulk,absorb and float voltages is all you need to do. The problem is that if something goes wrong and one individual cell voltage goes out of its safe operating range the BMS needs some way to shut down the charge controller or the inverter.

I think the bare minimum is to have the BMS generate an alarm which a person would act on. Rick@Raj174 and I use this approach with our systems because we have a good understanding of the batteries and keep then under surveillance. For the systems I have installed for friends I have taken a more conservative approach. The system I installed five years ago will disconnect the solar panels from the solar controller if a fault occurs. The two new systems I have just built go one stage further and use a Raspberry Pi micro controller that I have programmed to monitor the BMS and will send commands to the PIP5048 Inverter/charge controller to shut it down. As a further level of protection if this doesn't work the BMS will autonomously disconnect the solar panels from the controller or disconnect the battery from the Inverter. To me this is the gold standard which would require a number of separate failures before there is any safety issue.

I would be interested to know what the Electric Car Parts Company has to say about BMS systems and battery protection

Simon

karrak · August 2018

mcgivor said:
The use of mechanical relays to disconnect charge and discharge, as apposed to internal mosfets, or other means, would make it reletivaly easy to include a second BMS simply as a monitor of the primary system, interlocked via a second contact in the relay coil circuit. Perhaps I'm over thinking this a little, but redundancy is not nesesarally a bad thing.

The one problem with using mechanical relays is that they use quite allot of power. For your system if you are using the Omron relays I would think you would have to use a G9EC-1 as the main inverter relay and G9EA-1 for the charge controller. Combined these draw ~16W which equates to energy use of ~0.4kWh/day. Maybe a better arrangement would be to use MOSFET switches or can you shut down your inverter and charge controller via external inputs or switches?

The BMS I mentioned in in an earlier post uses onboard MOSFET switches to disconnect the battery, will monitor the cell voltages and do limited balancing and uses very little power, in the order of a few mA if not communicating with the rest of the world.

The attachment to your post describes the way that I now do my initial balance. I use a bench power supply to boost and a 1 Ohm resistor to drop individual cell voltages. I think the CALB cells that I am in the process of commissioning are all within ~1%SOC balance and am trying to balance them using the onboard balancing circuitry of the BMS even though the balancing current is only ~50mA. It is taking a while not because of the small balance current but because of the lack of sun. For a change we are having a wet and cloudy winter.

Simon

Dave Angelini · August 2018

I use a Raspberry in addition to the BMS monitor that I worried would shut down power when I was not home. They both get the can bus data but it takes both to shut down power. I also tested the system with what is called a failed power bridge test. It is basically a short on DC in to the inverter. The first short test tripped the battery breaker. The second short blew the fuse and tripped the breaker. I like both and feel pretty safe. This was all 2 years ago and all is well.

There should only be 2 things that shut down power offgrid, it is a new wrinkle that the battery is a 3rd. My solution takes it back to a dead battery or overcurrent protection causing loss of power. I definitely would not use relays. Just protect your network cables or double them.

Take a look at JEHUGARCIA on Utube for some ideas. I would not do any of this but for hobbyist he is pretty popular. Typical Southern California mindset and he is kind of funny.

https://www.youtube.com/watch?v=e_kTMDToieE&t=3s

Currently testing a second battery / inverter and it's on the bypass of the power distribution panel. This gets us to a completely redundant power system which is nice. Since it is new and has low hours on it I have to watch it alot.

It sounds like you (mcgivor) are still thinking this out? Maybe just wait another year and see? I would not jump into this if you question all the work. As I have said, this (homebrew) is not for my clients and I just do not have the time for it. Good Luck

Raj174 · August 2018

mcgivor said:

So Rick other than the Chargery BMS 16 not being able to ballance or communicate with external devices how is it from a quality performance standpoint. My understanding is balancing is not a regular task, but rather a one time affair done before assembly of the bank to ensure all cells are at, or as close as possible, (please correct me if I'm wrong ) to equal SOC, so doing it manually with single cell chargers is possible, albeit time consuming. The remote communication or a access via a PC is not something that interests me to be honest, why I like the fact it has its own display.

Simon, the reference to software control was to do with the use of data to comand a function, specifically with regards to fire alarm systems, where a physical relay contact is required as a fail safe, to shut down ventilation fans for example. Direct digital control , DDC, engineers always wanted to use digital input to their systems to shut down equipment over a LAN network, not fail safe. Also the transmission of alarm signals couldn't be transmitted via a VOIP telco system, it had to be physical land line, that is changing however with the use of multiple communication means to monitor each other, VOIP and cellular as examples, times are changing and physical land lines are fast becoming a thing of the past. The use of mechanical relays to disconnect charge and discharge, as apposed to internal mosfets, or other means, would make it reletivaly easy to include a second BMS simply as a monitor of the primary system, interlocked via a second contact in the relay coil circuit. Perhaps I'm over thinking this a little, but redundancy is not nesesarally a bad thing.

I would say that the Chargery BMS 16 is of decent quality. It has a good display and the claimed accuracy of 5 millivolts appears to be correct. This BMS was obsoleted when I added a 17th cell to my bank about 6 months ago, but I liked it well enough to go with the Chargery BMS 24T, which I recently found and will be ordering soon.

I don't use the balancing method describe in the attachment, I use a single cell charger to get all the cells up to 3.38 to 3.4 volts then put them all in parallel with busbars. Or, if they are already close in balance, charge them in series up to the same voltage then put them in parallel. Once in parallel, they are charged with the 3.65 volt charger until they reach 3.6 volts. The charger I use for this is 10 amp.
IMO this initial balance is good for 2 or 3 years, at least that is what I found with my HiPower cells. A rebalance could be done when necessary, it just involves paralleling the bank as described before. I have tried using the rebalancing method Simon mentioned, manually draining individual cells with a resistor to bring them into balance, but it is a bit too tedious for my liking. So I stick to what works for me.

Rick

Dave · August 2018

what a fine resource this blog is! i had
not thought about the need for an auto disconnect but I'll contact electric car parts for their perspective and update here.
before i move forward, I'm hoping to find a buyer for my batteries since they should have quite a bit of life left and i can't see just wasting them.

Lifepo4 replacement quandary

Comments

Categories