question

elimac avatar image

Mixing lead acid and lithium

My Lead Acid OPzS battery bank is "becoming smaller" as I continue to load the system more an more. Initially I sized the system for 20% DoD, but now in next winter I am afraid it may reach 40 to 50% or even more.

I have now the chance to get a good priced set of Winston LiFePO4 90Ah cells and I was thinking to build a smaller independent system to offload some of the work of the main system.


I allways thought it would be not advisable to put lithium in parallel with lead acid, but the more I think of it, the less crazy it seems. My LA system is 24V based, the 8 cell Winston would be 25.6V nominal. I would source a 3rd party BMS to manage the lithium. Maybe the BMS can take care of the issues - disconnect in low and high side of the daily swings.

I actually have found a product to make hybrid battery banks (BOS - LE300) so maybe this isn't so crazy after all.


To install in parallel and monitor the 2 strings I was thinking if it would be possible to use a separate BMV on each string (Lithium and OPZS). Is this possible, using 2 BMVs on same Venus device?

BMV Battery MonitorLithium Batterybattery charging
2 |3000 characters needed characters left characters exceeded

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

6 Answers
mattybeshara avatar image
mattybeshara answered ·

I have been experimenting with mixing a 140ah fusion LifePo4 with a full river AGM 105ah.

The results are very interesting.

Using 2 x Bmv712 I can see the discharge between the AGM and LifePo4 accurately.

Both batteries are 100% SOC

When a discharge load of 80a was applied, 62ah came from the LifePo4 and the remainder from the AGM.

This was also replicated during a charge of 80ah.

The LifePo4 took most of the current all the way until 100% then the AGM was held in absorb stage until it reached 100%

I’m finding in storage the LifePo4 with a nominal voltage of 3.2v per cell (12.8v total) is holding the AGM at a float voltage.

As the load disconnects the LifePo4 spills current into the AGM (topping back up to 100% or just above 12.8v.

Interestingly the lithium discharged first and charges first. It also holds the AGM at 100% until the LifePo4 has depleted which then you will notice the AGM takes over noting the severe voltage sag when loads are applied.



2 |3000 characters needed characters left characters exceeded

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

elimac avatar image
elimac answered ·

@mattybeshara

Interesting and extreme coincidence - I have just taken the leap, 3 days ago, to connect my new 180Ah (2x 90Ah) new LiFePO4 batteries in parallel with my existing OpZS 600Ah battery.

I antecipated, and can confirm what you say: The Lithium charges and discharges first. And at ~3.4 V per cell, we don't need to have high absorption voltages for the Lead Acid, we can keep it float "almost" all the time - provided that all below is considered:


- I have looked at my overnight typical consumption and found it to be in the ~3 kWh range.

- I have added in parallel LiFePO4 just enough to cover this: ~4.6 kWh, -> at 80% this is 3.7 kWh

- I have set this profile in Victron MPPT. The reasoning is - if LA is not discharged, don't need high voltages which can overcharge it. And they're bad for Lithium too. So I changed to this:


1. Absorption voltage: 27.6 ( 3.45 per cell)

2. Absorption time: 30 mins

3. Float voltage: 27.0 (3.375 per cell)

4. Limited charge current in DVCC to 30A, this makes the charge take longer and avoid longer times in float. Might change in winter to higher value.

5. Disabled DVCC temperature compensation


I don't have any BMV, but I do have Venus GX and I installed on the LiFePO4 bank a EV type BMS that has basic protection and balancing and also monitoring via bluetooth. So I can see more or less the difference between the Venus OS current (global current) and the BMS current as being the Lead Acid current. I confirmed with a DC clamp that the Lead Acid current is almost zero most of time.


Tipical cycle of composite bank is now: 25.6 -> tipical lowest voltage in the morning -> 27.6 (30 minutes absorption) -> 27.0 V float from ~14h PM onwards till sundown.


I think is working really good. I had my BMS activating load disconnect once due to high current surge but inverters didn't go down due to the presence of Lead Acid bank :)

I am prepared to change the MPPT profile - raise absorption voltage and time in case the Lead Acid bank is used when in prolonged cloudy weeks.


Cells are Winston yellow, 90Ah, in 2P 8S configuration.



1 comment
2 |3000 characters needed characters left characters exceeded

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

Thanks to both of you for this. Now it's been well explained I can see how it works (I'm just slow).

Pioneering stuff even, and something I can envisage for my next batt investment. Please keep us informed of any developments, good or bad.

Thumbs up to you both..

Alistair Warburton avatar image
Alistair Warburton answered ·

That interesting, and I havent looked at cell voltages but I suspect you may have issues, long term, due to sub optmel charge and discharge characteristics for both packs.

Have you considered charging them seperatly? You would loose a bit of effiiciancy but gain total control.

A DC to DC converter could be used to take energy from one pack and push it into the other. This could be done in either direction, as required, and because you could buck or boost the voltage the state of either paak wouldnt limit your options. unless they were empty or full of course.

Since you would either be charging or discharging, as a whole, a single converter could be used and physically switched, in for out, depending on the transfer requirement.

Another option would be to charge the LA from the AC, physically disconnect that and connect the DC-DC to move energy from the LA to the LI, when discharging.

I am about to do the latter, without the DC-DC, so that I can keep my 12V system for lighting and backup power, having just installed a large BYD Lithium pack.

The thought process is that, being totally off grid, I can still have lights and a small amount of AC even if I take the main system completly offline.

If I need to charge the 12V system under those circumstances I can manually connect its charger directly to the generator. Note that I will not be connecting the AC side of things at all, other than a manual changover option for some AC circuits.

Al


2 |3000 characters needed characters left characters exceeded

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

elimac avatar image
elimac answered ·

From all that I've read, it seems this is still on its infancy. It is not tested, but it is not ruled out either!

I think that for the time being I will use separate systems, but in the medium term I really would like to join the Lithium and LA. There's some issues to resolve, like making sure the LA gets full charge and at same time it is not too much for LiFePO4 cells, and also that the FLOAT voltage is ok for both types.

2 |3000 characters needed characters left characters exceeded

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

ejrossouw avatar image
ejrossouw answered ·

The main challenge I see is with charging them as they will have very different characteristics. I see batteries as a "system" and therein lies my concern with trying to mix them. What size is the existing bank? With LiIfepo you can easily go to 80-90% DOD which is a lot more energy available, but without knowing your requirement cannot assess the true impact. Maybe you will only require an additional charger and can connect it as an additional DC power source? I think the added potential complexity and equipment required to manage it, for me will much negate the benefit.

2 comments
2 |3000 characters needed characters left characters exceeded

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

My current battery is about 14 kWh / C20, thats 600Ah. I' ve read some BOS documentation and they seem to advise rations from 1/4 to 1/3 for the lithium componnent. So, I am thinking about getting 4 kWh of lithium if the price is ok. I have now 2 kWh booked (90Ah). I may also have to get more panels, I have now 3200Wp.

I agree with, the charging is different, but also the discharge. I was antecipating the following:

a) Due to lower internal resistance, the LiFePO4 will charge first and also discharge first.

b) The dedicated BMS "should" disconnect below, say, 25V and above 28.7. That is, once the system reaches absorption, I would like to have just the lead acid charging a little more time

c) The system will only discharge from LA below 25V.


I would have to find a good BMS that allows this but I am worried if so much switching is ok for it.

@elimac Curiosity spurred me on and I found a RV installation (rvgeeks I think) and they used a dedicated charger for AGM and solar charger for the Lithium. These guys used a seriously professional company do their install so ... may well be the simplest way forward or start saving to go all out lithium.

Kristof G. avatar image
Kristof G. answered ·

hi,


iam thinking with you ;)

for the BMV , i think thats possible to connect 2 of them to the colorgx

and for the battery : maybe this?

https://www.victronenergy.nl/battery-isolators-and-combiners/argo-fet-battery-isolators


its just like a car, 2 batteries , but dont know how they will work together since its a diff. "style" of battery.


maybe a way to think?


grtz

1 comment
2 |3000 characters needed characters left characters exceeded

Up to 8 attachments (including images) can be used with a maximum of 190.8 MiB each and 286.6 MiB total.

We do how well they will work together - they work very well.