DIY Electric Car Forums banner

Floodie / Lithium hybrid battery experiment

40198 Views 93 Replies 16 Participants Last post by  dougingraham
I'm starting an experiment to test the effects of using lithium LiFePO4 cells in parallel with cheapo lead acid batteries.

Lithiums are 40 AH Calbs, and they'll be boosting 29HMs (formerly known as 29DC) and GC8s. My current frankenpack consists of 3 29HMs and 11 GC8s.

Phase 1 of FLHE will put 8 calbs in parallel with 2 29HMs with a JLD404 to monitor AH usage and control charging. The calbs will be charged in parallel with the floodies until approx entering gassing phase where the JLD will cut the lithium contactor while the floodies finish their cycle. 1 29HM will remain unboosted as a control.

Phase 2 will add 8 calbs in parallel with 3 GC8s and tested as above.

The experiment will try to gain info about how the current sharing works out as well as how lead cycle life can be prolonged with the aid of a lithium booster.

A successful experiment would demonstrate an increase in lead cycle life sufficient to justify the cost of the lithium booster.

Attachments

See less See more
61 - 80 of 94 Posts
Update: I had my new draft instrument panel nearly complete, but while trying to calibrate one of the 50V DVMs it crapped out, so I have to wait a month or so for more to float over from China.

In another setback: My 120V controller no longer likes to run at 140V. I had to pull one 8V batt from the floodie pack, which means I'll need to crunch some numbers and pull 2-3 cells from the booster pack (too bad it isn't 4!) and refigure the ideal charge cutoff point.
I've been running 40 cells in the booster pack for a while now. It works well enough, but not nearly as good as it did with 42.

With the charged resting voltages matched the boosters are charging the floodies significantly (while coasting or stopped) for the first mile, then just a little the next 2/3 of my commute, then the floodies are charging the boosters more than I'd like for the remainder. If I drive further than normal it increases until it's at .5C after ~20 miles, which is as fast as I'll let them charge, so I cut the booster at that point.

I might drop to 39 cells and play with the charge points a little, but still need to crunch the numbers.

One thing that's surprised me this winter is that the lithium booster pack is much more affected by the cold than the floodies, even though it's better insulated.

I'm pretty bad at keeping records, but for the days I managed to remember last month here's a comparison of outside temp vs load sharing for the booster pack. These are averages, so real-time numbers are a little more extreme.

I'd probably get better data if I compared the morning split with night-time lows and the afternoon split with daily highs.

Attachments

See less See more
One will be charging the other, always, until equilibrium is met.

Miz
A diode fixes that problem.
Pulled a cell and reconfigured the booster box on Sat.

I've been charging to 133.9 V for a resting 132 (3.3 VPC).

Now charging to 132.5 V to rest ~130.65 (3.35 VPC).

Today's drive went well. Boosters carried a slight majority of the load and were always being recharged when idle @ ~2A. Will be interesting to see how they measure next time I go further than usual.
Ziggy,

I have a few questions and wanted to see if I could get a summary of your experiment so far.

My understanding is that you have parallel FLA and Lithium packs that have the same resting charge voltage. When the EV is driven, energy is drawn from both packs at a current ratio determined by the internal resistance of the respective packs. Typically this means that more current is drawn from the Lithium side. When the vehicle comes to rest again, since the Lithium put out more energy, current flows from the FLA pack, recharging the lithium pack. Feel free to clear up any misconceptions.

The objectives are using the Lithium pack to buffer the heavy current draw from the FLA, which both extends the cycle life and reduces the Peukert effect on the FLA. On the other side the FLA serves as a buffer for the Lithium which will keep the Lithium pack from going into deep undercharge. Also the FLA power participation facilitates using a smaller Ahr lithium pack as opposed to a Lithium only pack. Again feel free to clear up any misconceptions.

So my questions:

1. What exactly happens in a heavy current draw situation such as accelerating from a stop? Does the current draw ratio stay the same, or does it shift towards the FLA pack as the current draw gets higher? Is there any protection for the Lithium pack to prevent exceeding the maximum pulse/continuous current rating of the Lithium?

2. You have a post that indicates that after a long drive that the charging rate reaches 0.5C from the FLA to the Lithium pack. I presume this is because since the total energy of the Lithium is smaller than the FLA that the Lithium is depleted? Also you stated that the Lithium pack is then cut off. Does that mean that you no longer charge the Lithium pack, or that you allow for it to recharge from the FLA via a diode, but not contribute energy to the controller via a contactor?

3. What are the effects if a booster pack is built with smaller Ahr cells than the 40 Ahr cells that you used for your test? Three small current draw cells are the A123 20Ah pouch and the Headway 38120HP 8 Ahr cell. Both of these allow for 20C continuous draw. There is also the headway 40152S 15 Ahr cell that can have a 10C continuous draw. With each with prices at $25 or less per cell, it would be possible to build booster packs in small chunks ($900 a slice for a 36 cell group for 120V nominal pack). I am unclear what happens when such a booster pack is heavily current driven in conjunction with the FLA pack.

Any clarifications or answers would be welcome.

Thanks

ga2500ev
See less See more
Yup, good summary. The way the cells are now (39 cells, high SOC for 120V nom). I get nearly even sharing with a slight bias to the lithium. If it's cold it's a slight bias to the lead. Previously is was heavily weighted to the lithium but the opposite when cold.

1) The ratio is nearly the same under low amps as high amps. The cells can't exceed current ratings. My controller only goes up to 400A max, which matches their 10C pulse rating. I normally only pull 220-240 while accelerating anyway, so the boosters see 3C pulse, 1-1.5 continuous. The most they've ever seen is 6C. I'll probably shoot for 4-5C regularly if I ever get around to installing my other meters.

2) Yes, with my previous setup the further I discharged the more out of balance the strings would get and I'd turn the contactor off when the balancing current reached C/2. That was always just for a mile or two so after letting the car sit for a minute I'd reconnect the booster and it'd charge starting ~16A and settle below 5 very quickly and I'd just charge like normal.

That doesn't happen anymore. The largest balancing charge I've seen in my current config is 3.5A around 2/3 into my commute, and it just drops off from there. Once I get more meters in I'll do a range test to see if that holds true past the 22ish miles max I've driven.

3) I don't know how things would work out with a smaller booster. That was originally my intention but I couldn't stomach paying over $1.50/ah for a booster string that wouldn't significantly contribute to my range once fully lithium. I think it could be made to work very well with a little tweaking.

I'm rather hooked on the whole booster thing so 5-10 years from now (after switching to full lithium and upgrading controller, charger, and motor) I expect I'll have a LiFePO4 pack with a LiFePO4 SuperCapacitor Hybrid Battery as a booster.
See less See more
I may very well do something similar if and when I build a road-going EV. But I plan to use SLAs instead of FLAs, and I intend to make a 250-320V battery pack with a three-phase 230VAC motor and VFD. I can get 12V 12Ah SLAs for about $20 each, including delivery, so that is about $0.14/Wh (although probably closer to $0.25/Wh including Peukert). They are 12 lb each (1 lb/Ah), compared to 0.6 lb/Ah for FLA:
http://www.tractorsupply.com/en/store/stowaway-battery-st27dc180

The cost for these FLAs is $85 or $0.067/Wh. For my purposes, I would need at least 20 of these for 240 VDC, which would be 26.4 kWh and about $1600. The weight would be 1300 lb which is a bit much but probably manageable especially if installed in a truck. If I can run at 350 wH/m at 40 MPH I would be pulling 14kW (18.6 HP) or 58A from the 240V pack, so with Peukert I would have a true 58 Ah or 1 hour of driving at 40 MPH or 40 miles (but probably 30 miles to be safe).

To get about the same performance with LiFePO4 I would need 80 x 60Ah cells which I might be able to get for $100 each including delivery, so $8000. They are about 7lb each for 560lb total. They would also need three times as many high-current interconnects, as well as a more sophisticated BMS, which adds maybe another 50 lb and $800. So 600 lb and $8600.

If I made a combination with 50 Ah FLAs and 30Ah LiFePO4, I would probably spend $4300+$800 or $5100, and weight would be about 650+300 = 950 lb. There is also the complexity factor with two different chemistries.

As for long-term cost, the Lithium cells might last 10 years, and the FLAs 2.5 years, so that makes the lead system $6400 vs $8600. If you factor in the recycle salvage value of lead at $0.27/lb, the cost of a set of FLAs drops $350 to $1250/set or $5000 for 10 years.
http://www.earthworksrecycling.com/prices/index.html

I think LiFePO4 needs to drop to about 1/2 its present cost, and then the numbers look a lot better. I don't know if it really makes sense to use the combination at this point, especially for a high voltage pack.
See less See more
Yup, good summary. The way the cells are now (39 cells, high SOC for 120V nom). I get nearly even sharing with a slight bias to the lithium. If it's cold it's a slight bias to the lead. Previously is was heavily weighted to the lithium but the opposite when cold.
I'm glad I've been reading this right.

1) The ratio is nearly the same under low amps as high amps. The cells can't exceed current ratings. My controller only goes up to 400A max, which matches their 10C pulse rating. I normally only pull 220-240 while accelerating anyway, so the boosters see 3C pulse, 1-1.5 continuous. The most they've ever seen is 6C. I'll probably shoot for 4-5C regularly if I ever get around to installing my other meters.
Is the 1-1.5C with the original setup of 40 Ahr cells? I was unclear when you got the second set of cells if you paralleled them with the original set or not (giving you 80 Ahr cells).

2) Yes, with my previous setup the further I discharged the more out of balance the strings would get and I'd turn the contactor off when the balancing current reached C/2. That was always just for a mile or two so after letting the car sit for a minute I'd reconnect the booster and it'd charge starting ~16A and settle below 5 very quickly and I'd just charge like normal.

That doesn't happen anymore. The largest balancing charge I've seen in my current config is 3.5A around 2/3 into my commute, and it just drops off from there. Once I get more meters in I'll do a range test to see if that holds true past the 22ish miles max I've driven.
Maybe you can clear up my confusion a bit. The action you described above is only when the car is stopped right? If the car is powered, then both strings are contributing energy to the controller/motor, and not to each other. So I'm just trying to make sure there isn't a situation when the FLA pack is both powering the motor and charging the booster at the same time.


3) I don't know how things would work out with a smaller booster. That was originally my intention but I couldn't stomach paying over $1.50/ah for a booster string that wouldn't significantly contribute to my range once fully lithium. I think it could be made to work very well with a little tweaking.
$1.50 an Ah seems to be about standard yes? Most of the sites I see has 40 Ah prismatics in the $58-$60 range which is just under $1.50/Ah. I've seen Headway 40152S 15 Ah cells for $23 which is about $1.53/Ah. I'm not so sure there's that big a price difference buying smaller cells. Consider if you were starting with 160 Ahr cells. They seem to be running about $210-$220 a cell. You can get 150 Ahr by tying 10 15Ah cells together. The cost at $23 each would be $230. So the total cost of a 36x160Ah stack is $7560 while the 360x15Ah setup is $8280. The total energy is 19 KWh vs. 17.8 KWh for a final $/Wh cost of $0.39/Wh vs. $0.46/Wh. But the second stack can be purchased as 10 individual $828 strings while the entire cost of the 160 Ah stack would need to be capitalized in the beginning.

I have access to 12Vx35Ah U1 wheelchair AGMs at $25 each. For $750 it's possible to put together a 12.6 KWh SLA stack in a 120Vx105 Ah 10S3P configuration at a weight of 750 lbs. Now I agree that adding less than 2 Kwh of Lithium isn't that helpful to range. But what are the effects in terms of buffering the deep discharge current when accelerating? What effect would that have on the range since the Peukert effect is lessened by Lithium string.
I'm rather hooked on the whole booster thing so 5-10 years from now (after switching to full lithium and upgrading controller, charger, and motor) I expect I'll have a LiFePO4 pack with a LiFePO4 SuperCapacitor Hybrid Battery as a booster.
As you can see I'm interested too...

ga2500ev
See less See more
Is the 1-1.5C with the original setup of 40 Ahr cells? I was unclear when you got the second set of cells if you paralleled them with the original set or not (giving you 80 Ahr cells).
At the beginning of the project I had a 124V nominal lead acid pack.

I bought 16 CALBs (SE40s). I put 8 of those in parallel with 2 12V 29HMs (Sam's replacement for 29DCs). This was my proof of concept just to make sure things went well. It did so I added the other 8 cells giving me 16 SE40s in parallel with the 2 29HMs and 3 8V GC8s.

I planned to run that for a long time gathering data, but ended up finding the cash to do the full booster pack, so I bought 26 more CALBs (CA40s). By this time I had replaced some old lead, giving me a 128V nominal lead pack (13 GC8s + 2 29HMs) and a 42 cell booster pack (16 SE40s + 26 CA40s).

My controller decided it didn't like being overvolted so I pulled 1 GC8 and dropped to 40 cells on the booster, then 39 to improve the balance.

I won't have a 80AH pack until the lead all dies in the next year or two and I buy another 40ish cells to put in parallel. So far I've only used series lithiums in parallel with lead.


The action you described above is only when the car is stopped right?
When coasting or stopped there's a balance current between the packs, sometimes one way, sometimes the other. Now that I've again found a good balance point it's very small, and is nearly 0 within a few minutes.


$1.50 an Ah seems to be about standard yes? Most of the sites I see has 40 Ah prismatics in the $58-$60 range which is just under $1.50/Ah. I've seen Headway 40152S 15 Ah cells for $23 which is about $1.53/Ah. I'm not so sure there's that big a price difference buying smaller cells. Consider if you were starting with 160 Ahr cells. They seem to be running about $210-$220 a cell. You can get 150 Ahr by tying 10 15Ah cells together. The cost at $23 each would be $230. So the total cost of a 36x160Ah stack is $7560 while the 360x15Ah setup is $8280. The total energy is 19 KWh vs. 17.8 KWh for a final $/Wh cost of $0.39/Wh vs. $0.46/Wh. But the second stack can be purchased as 10 individual $828 strings while the entire cost of the 160 Ah stack would need to be capitalized in the beginning.
I'd consider $1.25/AH to be a decent price for CALBs. Depending on the supplier the price/AH can be the same for small as large cells. Mine were cheaper and other prismatic brands are cheaper. Headways are more, and they've gone up instead of down; I wouldn't consider using A123 or other pouches because of the assembling hassle, same for non-threaded cyllindrical cells (some headways and many other brands).

While I think headways could make a great booster, I don't think you'd want to go that route for a full pack unless you really need C rates over 10. I think the few who have done full headway packs have said they wouldn't again. 360 cells would be a huge pain to connect, and I've heard they don't stay balanced as well as the prismatics, so that'd be another pain/expense.


I have access to 12Vx35Ah U1 wheelchair AGMs at $25 each. For $750 it's possible to put together a 12.6 KWh SLA stack in a 120Vx105 Ah 10S3P configuration at a weight of 750 lbs. Now I agree that adding less than 2 Kwh of Lithium isn't that helpful to range. But what are the effects in terms of buffering the deep discharge current when accelerating? What effect would that have on the range since the Peukert effect is lessened by Lithium string.
That could work. I'm not too familiar wth AGMs, but I doubt wheelchair batts are designed for much current. For comparison, my original pack was 10 125AH DC29s for ~$850 and lasted about a year.

I think my 12V batts are around a year old now, so unboosted they'd be about dead. I need to start doing an unboosted run every 1-3 months to see how they hold up on their own. Also for the reduced Peukert I need to get more meters connected and do a good range test.
See less See more
Yup, good summary. The way the cells are now (39 cells, high SOC for 120V nom). I get nearly even sharing with a slight bias to the lithium. If it's cold it's a slight bias to the lead. Previously is was heavily weighted to the lithium but the opposite when cold.

1) The ratio is nearly the same under low amps as high amps. The cells can't exceed current ratings. My controller only goes up to 400A max, which matches their 10C pulse rating. I normally only pull 220-240 while accelerating anyway, so the boosters see 3C pulse, 1-1.5 continuous. The most they've ever seen is 6C. I'll probably shoot for 4-5C regularly if I ever get around to installing my other meters.

2) Yes, with my previous setup the further I discharged the more out of balance the strings would get and I'd turn the contactor off when the balancing current reached C/2. That was always just for a mile or two so after letting the car sit for a minute I'd reconnect the booster and it'd charge starting ~16A and settle below 5 very quickly and I'd just charge like normal.

That doesn't happen anymore. The largest balancing charge I've seen in my current config is 3.5A around 2/3 into my commute, and it just drops off from there. Once I get more meters in I'll do a range test to see if that holds true past the 22ish miles max I've driven.

3) I don't know how things would work out with a smaller booster. That was originally my intention but I couldn't stomach paying over $1.50/ah for a booster string that wouldn't significantly contribute to my range once fully lithium. I think it could be made to work very well with a little tweaking.

I'm rather hooked on the whole booster thing so 5-10 years from now (after switching to full lithium and upgrading controller, charger, and motor) I expect I'll have a LiFePO4 pack with a LiFePO4 SuperCapacitor Hybrid Battery as a booster.
My bad Ziggy wish I read this thread orig instead startin the other.Where did you get your Lifepo4 cells from?
Like I sorta stated in other thread. I want to boost the 24vdc 80ahLifepo4pack with SLA or VRLA two 12volt 100ah which pair will give 24vdc 100ah SLA . Is same ah better or a little higher ah on the SLA side better. Also could you use a power diode between packs instead of contactor?
I am going to restate some my past issues (my apologies to OP)
I want to augment a Enginer PHEV RFE 24vdc 80ah Lifepo4 pack(this pack augments the OEM pansonic (Nimh 201vdc 7.2amp pack). Charge and balance of the RFE 24vdc 80ah Lifepo4 will be done outside of overall systems while vehicle unoperate(complex)However curious if I can apply a
24 vdc 90ah or 100ah SLA to boost the 24vdc 80ah Lifepo4 . Additionally thinking of using power diode to seperate the Lifepo4 from SLA and may possible use solar panels with charge control on SLA. Thanks ziggy for your posts
Multiple threads on the same topic is great. It promotes further discussion and makes things easier to understand.

I think it works best when the lead is significantly larger than the lithium, otherwise the lithium will be doing nearly all the work and an imbalance will form the longer you discharge.

If things don't work as well as you want though you could always use one, then the other.

I'm not aware of diodes that can handle over 100V+ @ 100A+. If they exist and are reasonably affordable it could help but you'd need separate monitoring on each pack. You need a contactor (or relay) anyway to break the connection during charging.
As long as you don't over or under charge the lithium cells you will be fine. I had a disaster with my Prius booster pack to my lead acids. Prius cells over charged. I did not have a smart charger which would be important in the case of using different batteries. Lithium charge is nearly the same as lead acid except the balancing cycle and float cycle.

You going to use a lithium charger or lead acid charger?
I have found and seen a lot of this on Lifepo4 packs where someone thinks you can trust a BMS to do its function on charge (plug,charge,and leave) However the reality is Lifepo4 get overcharged even with BMS. I am of the thinking that timed charge,voltage checks, bottom balance is a better solution.
I am hoping maybe a SLA booster pack to a PHEV Enginer Lifepo4 will work.
Multiple threads on the same topic is great. It promotes further discussion and makes things easier to understand.

I think it works best when the lead is significantly larger than the lithium, otherwise the lithium will be doing nearly all the work and an imbalance will form the longer you discharge.

If things don't work as well as you want though you could always use one, then the other.

I'm not aware of diodes that can handle over 100V+ @ 100A+. If they exist and are reasonably affordable it could help but you'd need separate monitoring on each pack. You need a contactor (or relay) anyway to break the connection during charging.
My electronic and EV history is limited. But was thinkin maybe these could work
http://www.ebay.com/itm/DD104N16K-I...564?pt=LH_DefaultDomain_0&hash=item3a7ed4060c
http://blog.evtv.me/store/proddetail.php?prod=diode
Those could work. They just look rather weenie to me compared to my other connections.
Multiple threads on the same topic is great. It promotes further discussion and makes things easier to understand.

I think it works best when the lead is significantly larger than the lithium, otherwise the lithium will be doing nearly all the work and an imbalance will form the longer you discharge.

If things don't work as well as you want though you could always use one, then the other.

I'm not aware of diodes that can handle over 100V+ @ 100A+. If they exist and are reasonably affordable it could help but you'd need separate monitoring on each pack. You need a contactor (or relay) anyway to break the connection during charging.
I am thinking IGBT may work too. So you could not solar panel charge the SLA side while connected . I have heard and seen relay however comin up with how to diagram or wire in the relay not comfortable with. In your application you used a Tyco kilovac
The EVTV dual diode may be better because you can use both of them, while the one on eBay is two diodes connected as a half-bridge. The problem with diodes is that they conduct only in one direction, so each battery pack can drive the controller, but they can't charge each other, and regen will not work. It might be better to use a pair of big MOSFETs or IGBTs so the intrinsic freewheeling diode can steer the current out of the packs, and the gates of the devices can be activated for charging as desired. With MOSFETs, when the gate drive is applied, current can flow in either direction so it may be more efficient at lower current levels where the voltage drop from RdsOn is less than the diode drop. ;)
The EVTV dual diode may be better because you can use both of them, while the one on eBay is two diodes connected as a half-bridge. The problem with diodes is that they conduct only in one direction, so each battery pack can drive the controller, but they can't charge each other, and regen will not work. It might be better to use a pair of big MOSFETs or IGBTs so the intrinsic freewheeling diode can steer the current out of the packs, and the gates of the devices can be activated for charging as desired. With MOSFETs, when the gate drive is applied, current can flow in either direction so it may be more efficient at lower current levels where the voltage drop from RdsOn is less than the diode drop. ;)
I kinda want a one way feed so that would be SLA to Lifepo4 to Nimh.The Lifepo4 Enginer (middle feed) could be switch off at any point during operation. Thinkin maybe high voltage power diode,IGBT, golf cart relay, isolated solenoid relay. possible choices
Lead along side NiMH is not a good idea. Yes it can be done but introduces a high level of complex issues.

I agree that people see the illusion of saving money. It is that illusion that stops EV's from becoming main stream. I am for breaking the illusion.

Electric is it and Lithium is king. Cost IS less.
Lead has drawbacks to Lifepo4 in both weight and energy density.
.
61 - 80 of 94 Posts
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top