DIY portable, useful LiFePO4

[markopolo]

I forgot to add that I paralleled the LiFePO4 pack yesterday with 200Ah of lead acid batteries on float charge. That system is a battery backup for a sump pump.

balanced cell groups.jpg

The LiFePO4 cell groups are well balanced this morning. (note: Balancing was set at 3.5VPC (14V) yesterday and therefore is effectively Off). The voltage is right where it could be OK or could cause some capacity loss over a long period of time. It depends on the cells.



Read more 1/3 down the page here: https://marinehowto.com/lifepo4-batteries-on-boats/

Can you hold a standby voltage at 3.400VPC or 3.35VPC or lower? Absolutely, but we don’t really know the long term affects other than to say it is it may shorten the life of some cells and may cause little to no harm to others. The premium cylindrical cells we tested at 3.400VPC (using a very expensive very linear power supply), lost no quantifiable capacity but some of the cheap cells lost as much as 16% in the same time frame using the identical charge source. Do you or will you know the quality of the cells inside your own battery and how they actually handle a “standby voltage“?

The cells that I have rest at 3.4234 volts or higher after sitting for a week. That's the lowest of the group of 40 I've prepped for pack #2. The four lowest will be excluded from the group as only 36 are needed for pack #2. The 3.406 VPC (13.624V) this morning is less than 100% SOC but is it low enough for long term standby?

I'll be more comfortable using 3.350 VPC (13.40V) for long term standby.

Fortunately, I have specific instructions on how to lower the float voltage on the Samlex SEC-1215UL and have the required tools to do it precisely.

That's another item for the to-do list.

[booster]

This is all very interesting, to be sure. You are doing an impressive combination of thoroughness and willingness to do the necessary research and time investment to fully understand the "hows and whys" of it all.


It will be of great interest to me to see how well the precise settings initially hold up over time do to the natural drift and operating condition variations, and what those changes, if any, have on the cells themselves for all the important parameters.


Once you find that out, you may be able to answer some of the questions that has always hung over the lithium battery installs, like how much control accuracy is really needed for good cell life and such. You will also know by then if the very precise setting up of everything in the beginning is worth the effort or if the system goes out of that precision quickly or not.



Excellent information for all of us. With the huge variation of systems available, from basic lead acid style controls being used up to very sophisticated and tightly (claimed) controlled and expensive systems this is the kind of stuff we need to understand better since the vendors all claim they work very well.

[Winston]

Originally Posted by markopolo

The 3.406 VPC (13.624V) this morning is less than 100% SOC but is it low enough for long term standby?

I'll be more comfortable using 3.350 VPC (13.40V) for long term standby.

We don't have good data at 13.6 volts . . . but our guess is that 13.624 volts is very close to 100% SoC.

Our tests have shown (for GBS cells) that 13.36 volts = 90% SoC. Since none of our charging sources can be programmed more closely than 0.1 volts, we've targeted 13.4 volts as a normal operating voltage.

Some argue that SoC's closer to 50% are best for long-term storage. You may want to take your 13.4 volt 'storage' number down a few notches.

[markopolo]

I think precision is more critical with the smaller size pack as there is less capacity to spare. I have to maximize the potential of what's there.

I haven't given much thought to what the out-of-commission storage voltage should be. Winston's posts will guide me on that.

I'm more interested in ready-to-go standby voltage and capacity than out-of-commission storage. In the event of a power outage I want to have the pack(s) ready and available at the highest capacity that does not affect lifespan too much. The system that the pack is currently connected to has an automatic transfer switch and an always on 1500W PSW inverter.

As Winston points out, 13.624V is very close to 100%. Probably only 200mAh less capacity than full for this pack. It's not worth any risk of cell degradation for only 200mAh capacity.

This chart from TI illustrates why 3.35VPC (13.4V) should/could be an effective standby voltage.

TI chart.JPG

At most, I'd be giving up 10% capacity according to the chart. That's close to Winston's observed 13.36 volts = 90% SOC.

[JohnnyFry]

Has anyone succeeded in accessing the online manual for the XiaoxiangBMS? If so could you post the url.
Thanks,
John

[markopolo]

I used this one yesterday for PC use: https://cdn.hackaday.io/files/162806...for%20v1.1.pdf

Looks to cover what I need.

Site: https://hackaday.io/project/162806-jbd-bms-protocol

Note: the firmware on that site is not for the 4S BMS' we're using.

[JohnnyFry]

Thanks for the fast reply. I do not have the same model but I was able to glean a lot of useful information from the site. I have ordered the precision voltmeter components you suggested...should provide something to while away a few hours during the "Shelter in place" period!

[markopolo]

Re: voltage accuracy - here's a comparison of what I have:

voltage_accuracy.jpg


Same image but open viewing:



It starts with the voltage reference and the measured voltage that came with it. The DDM645 is pretty accurate and I know which trimpot to adjust to dial it in but haven't done that yet. It is slow to settle on a voltage though. Kind of annoyingly slow when measuring 40 cells. The Aneng AN870 is fast but not as accurate. I think it's accurate enough though for most use. The test leads are nice and it came with bonus leads and clip ends and lug ends etc.

[markopolo]

Link to xiaoxiang 3.1.1015.23.apk - http://jiabaida.com/xiaoxiang.apk
It's the most recent version available I think.

Some folks on forums are using outdated three year old versions of the app and missing out on a lot of awesome features.

lots of features in app.jpg

one app multiple BMS.jpg



Edit: just adding another screen capture

Access to so many settings.jpg

[ikanode]

The link doesn't work
Link to xiaoxiang 3.1.1015.23.apk - http://jiabaida.com/xiaoxiang.apk

[markopolo]

I ran a load test today using a 780 Watt rated toaster and 650 Watt coffee maker from the van.

Pack 1 Test coffee maker and toaster.jpg

Powered the toaster no problem.

Pack 1 Test powering toaster.png

Powered the toaster and the coffee maker simultaneously no problem. There's a lot of power in this 18lb battery pack!

Pack 1 Test coffee maker and toaster simultaneously.png

Voltage did recover after the testing similar to lead acid batteries.

Pack 1 Test voltage recovery.png

[markopolo]

How to pass that 120A from the LiFePO4 pack to the van's AGM battery system circuit without allowing 100A or more to get to LiFePO4 pack from the van's alternator?

I keep coming back to using Schottky diodes in parallel. PDF: https://www.st.com/resource/en/appli...lectronics.pdf

Paralleling electrical items is generally avoided. If you must parallel then exceed the rate required. With diodes, it appears that exceeding the capacity by a large margin would be the only way to go.

Could seven of these 45V 30A Schottky diodes do it? I'll probably have to try it to find out.

Max load (microwave oven) is expected to be 90A so I hope 210A of parallel Schottky diodes can handle that. Any advice/comments from electronics experts will be appreciated.

They could be formed into a 1.1" diameter bundle with strain relief as shown. I'd insert the ends into molten solder in the cups provided by the lugs.

diodes.JPG

7 diodes.png

That's a project for another day.

I tested voltage drop on a Cooper Bussman typical RV isolator yesterday. Voltage drop was 0.7V even with only 0.25A load current. Next isolator test(s) will be to record current acceptance of the LiFePO4 pack when recharging with the 0.7V voltage drop through the SOC curve.

The DC charge controller I put together - https://www.classbforum.com/forums/f...tml#post103879 - can charge lead acid batteries from the LiFePO4 pack or charge the LiFePO4 pack from the lead acid batteries with precise voltage control but it is limited to 20A.

[markopolo]

Possible parallel LiFePO4 / lead acid solution when using Pack 1. Minimal user interaction required, maximizes output potential.

parallel lifepo4 and lead acid.jpg

Notes related to diagram:

Both DC controllers are boost/buck and have back-flow prevention. Precise output voltage and current control. 5A on shore power to LiFePO4 pack isn't much but I already have the controller so might as well use it. Both DC controllers can be set to power on with output enabled and could be set at 13.40V at startup for example. Both can be controlled by a phone App. Also, shore power would usually be at least an overnight so plenty of time for a small LiFePO4 pack.

Grid power charger (PD 45A can function as a power supply) effectively doesn't have a voltage reference in this diagram. I'll test to see how it affects this timer based charger as the lead acid batteries do need to see 14.4V from it.

I'm presuming that the LiFePO4 pack will supply all loads, even a 100A load, until load induced drop or discharged to around 12.8V when lead acid batteries start to share the load.

I've had a fully charged lead acid battery fresh off 13.5V float directly paralleled to the near full LiFePO4 pack for about 22 hours with very little capacity loss (0.01Ah according to the BMS). The LiFePO4 pack pulled some surface charge off the lead acid battery when first connected.

Solar: I'm trying to prevent a energy sucking loop of lead acid batteries charging the LiFePO4 pack which then charges the lead acid batteries. I think I can use two solar controllers, each rated for more than array total output. There are two bank solar controllers so that is another option but for now I want to use parts that I already have. I have a low cost unused solar controller that allows some adjustment of values according to the manual but if I can't control the voltage in all modes then I'll probably skip doing solar to LiFePO4 charging for now. I don't want the LiFePO4 unnecessarily held at 14.4V for example.

I did some testing using a traditional isolator on input and output. My alternator can drop as low as 13.9V when hot and with the fridge on inverter. As an input charging control source, the 0.7V drop caused by the isolator would leave 13.2V for charging the LiFePO4 pack. I could turn off the fridge etc. but it's not a great solution so I won't pursue it. Similarly, as an output control source I only measured 12.4V getting to the load so not pursuing that either.

I might not do any of this but think it's worth planning ahead. I might just use the 20A DC-DC charge controller to charge lead acid batteries from the LiFePO4 pack or charge the LiFePO4 pack from the lead acid batteries as needed.

For an example of a commercial lithium/lead acid pairing solution check this link: https://www.bos-ag.com/products/le300/


Edit: adding a link to Harry the hybrid pioneer's setup: https://www.classbforum.com/forums/f...stem-8526.html - and also JWBrown's setup - https://www.classbforum.com/forums/f...tall-9904.html

[JohnnyFry]

Well, I think that I have my problem solved. I have a Renogy 20 DC/DC charger connected theough the diode isolator. The charge enable line on the Renogy is energized by the ignition circuit but runs theough a NC contact on a DPDT relay powered by shore power so the Renogy is cut off as soon as the shore power is energized. A NO contact on the same relay energizes a 12V high current relay which parallels the starting battery when shore power is energized. This charges the starting battery through the TrippLite when on shore power and provides the voltage reference for the system after the LiFePO4 charging cuts out at 100%

[booster]

Marko, it will be interesting to see what the PD does on power supply mode. I had always assumed, not a good idea in general, that the PD would go to it's base voltage as if it had no Wizard on it, but I don't know that for sure. It may only have a way to internally reference at that voltage, which IIRC is about 13.6v.


Are the two solar controllers sharing the input from the panels?

[markopolo]

Originally Posted by JohnnyFry

Well, I think that I have my problem solved. I have a Renogy 20 DC/DC charger connected theough the diode isolator. The charge enable line on the Renogy is energized by the ignition circuit but runs theough a NC contact on a DPDT relay powered by shore power so the Renogy is cut off as soon as the shore power is energized. A NO contact on the same relay energizes a 12V high current relay which parallels the starting battery when shore power is energized. This charges the starting battery through the TrippLite when on shore power and provides the voltage reference for the system after the LiFePO4 charging cuts out at 100%

How many Ah is your new system Johnny?

Great that you found a solution. The Tripp Lite looks better than my PD charger for LiFePO4. The Tripp Lite, on high & wet cell setting transitions to float when current flow drops to around 10A. Float is around 13.5V on the Tripp Lite.

[markopolo]

Originally Posted by booster

Marko, it will be interesting to see what the PD does on power supply mode. I had always assumed, not a good idea in general, that the PD would go to it's base voltage as if it had no Wizard on it, but I don't know that for sure. It may only have a way to internally reference at that voltage, which IIRC is about 13.6v.

Well I learned a bit about Schottky diodes. You do see voltage on either side so the PD charger appears to operate normally in my brief testing.

Voltage drop is around 0.3V with a Schottky diode so 14.4V become 14.1V so not great for the lead acid batteries. Also, Schottky diodes do allow a small amount of backflow current.

Originally Posted by booster

Are the two solar controllers sharing the input from the panels?

Yes, that is/was the plan. Most of the plan doesn't work well so I have to rethink things.

[markopolo]

New plan diagram:

parallel lifepo4 lead acid new plan.jpg

Seems too easy .....

I have a 30A rated programmable relay that you set low(on) and high(off) voltage - model PS46A01 - https://www.google.com/search?q=ps46a01&tbm=isch

[@Michael]

Originally Posted by markopolo

New plan diagram:

Attachment 8980

Seems too easy .....

The loop from the lead acid -> relay -> DC/DC -> Lithium -> diode and back to the lead acid looks like a perpetual motion machine to me.

[markopolo]

Originally Posted by @Michael

The loop from the lead acid -> relay -> DC/DC -> Lithium -> diode and back to the lead acid looks like a perpetual motion machine to me.

Not perpetual.

For the loop to occur the lead acid system has to be above 13.2V. That can only happen while charging from any source (shore, solar, alternator). High LiFePO4 voltage will also cause the loop but only until LiFePO4 voltage drops to 13.4V (that's likely good for the LiFePO4 pack). The 0.3V drop across the diode pack means that the lead acid battery voltage will eventually be below 13.1V and the loop is broken.

The will be some drain from the LiFePO4 pack to the lead acid system but it will be very low, so low that it's probably not easily measurable based on my testing so far. I had the LiFePO4 pack directly paralleled to fully charged lead acid battery for almost 48 hours as a test and the energy loss was negligible. That test held the lead acid battery at 13.4V so holding at 13.1V and lower should take even less energy.

All of the above will happen automatically. The three main design goals were: 1. full use of the LiFePO4 pack up to 150A while paralleled, 2. base system being completely automatic and 3. precisely controlled charging of the LiFePO4 pack from every source (shore, solar, alternator).

There will be manual options.

I figured I'd add a battery switch that I have on the LiFePO4 pack output the but leave it on almost always. If the van was to be stored in my garage and not plugged in for some reason then I could turn that switch off. I can also disable LiFePO4 pack output via the phone app. If the van is stored for a while and plugged in then I could just dial the DC Charge Controller voltage down to 13.3V via another phone App. The lead acid system would be held at 13.2V by the PD Charger. Setting the Charge Control relay to turn on at 13.3V would also disable the DC Charge Controller as the PD charger float voltage is 13.2V.

That's the theory anyway !