Author Topic: Solar BMS a solar charge controller for LiFePO4  (Read 57266 times)

0 Members and 1 Guest are viewing this topic.

Offline donovanpl123

  • Newbie
  • Posts: 1
  • Country: jm
Re: Solar BMS a solar charge controller for LiFePO4
« Reply #75 on: April 09, 2016, 12:40:21 am »

 
I have no electrolytic in my design so 25 years no problem all is solid state there is also no PWM or so absolutely no noise (electrical noise or interference with other devices)
 
So, if there is no PWM in the BMS, then are you just connecting the PVs to the batteries on and off?  If this is so, how do you achieve the various segments in the charge cycle?
 

Offline electrodacusTopic starter

  • Super Contributor
  • ***
  • Posts: 1858
  • Country: ca
    • electrodacus
Re: Solar BMS a solar charge controller for LiFePO4
« Reply #76 on: April 09, 2016, 02:16:40 am »
So, if there is no PWM in the BMS, then are you just connecting the PVs to the batteries on and off?  If this is so, how do you achieve the various segments in the charge cycle?

Lithium cells require only cc and cv charging that is if you want to take advantage of the battery entire capacity and do not care as much about the cycle life.
For long cycle life you can only use one stage the cc ( bulk ) charge.
The most common Lithium cells used in most portable electronic devices and recently electric cars will only charge up to 85% if you are using just one stage cc charging (you stop the charging as soon as the voltage gets to in this case 4.2V).
If you do that you almost double the battery life cycle while you are losing about 15% of the stated capacity. Since in mobile electronics energy density is important that 15% is way more important then cycle life that is anyway good enough for this type of applications.
I recommend and use LiFePO4 batteries for stationary solar energy storage systems and those will charge to at least 95% with just a single stage charging cc and have a much better life cycle than the more common Lithium-ion mentioned above so cost amortization over the life of the battery is way better.
Typical LiCoO2 cells (usually found as small 18650 cylindrical cells) will last 500 cycles at 100% DOD and if charged with just cc up to 4.2 they can do 700 to 1000 cycles while just around 85% of capacity is available cost in volume is at around $220/kWh
So for a 1kWh battery 500 cycles x 100% x 1kWh = 500kWh over the usable life of the battery $220/500kWh = 44 cent/kWh stored (this is the theoretical amortization cost of this battery charged with cc to 4.2V and cv until current drops to 0.02C)
If the same 1kWh battery is charged with cc only 800 cycles x 85% x 1kWh = 680kWh better than first case $220/680kWh = 32cent/kWh

LiFePO4 costs more at around $400/kWh but you can find large cells hundred of Ah capacity and can last 3000cycles at 100% DOD (even 6000cycles for quality Sony cells) and around 7000cycles at 70% DOD for Winston cells that I usually recommend for solar storage
So same 1kWh capacity will be 2x as heavy as a LiCoO2 so not good if energy density is important but 7000 x 70% x 1kWh = 4900kWh so $400/4900kWh = 8 cent/kWh and the life of the cell is probably 20+ years in typical solar application.
In real life you will not be able to use 70% of the battery capacity every day so over 20 years (about 7000 days with one cycle every day) with an average of 30 to 40% DOD per day you can not realy get to use 4900kWh but still it will be close and probably real life cost of storage will be around 20 to 30 cent/kWh still will be way better than LiCoO2

Sorry for the long explanation the idea is that in stationary offgrid storage cost of storing each kWh over the life of the battery is important and for that you need to only use cc charging so stop the charging as soon as the cell gets to the set level 3.55V for LiFePO4 or 4.1V or even 4V for long life using LiCoO2 the PWM or other methods of constant voltage charging will be detrimental to battery life.

Since last time I wrote here the new SBMS model was done. Here is a video I just uploaded about the last version if you are curios
« Last Edit: April 09, 2016, 02:18:22 am by electrodacus »
 

Offline scttnlsn

  • Newbie
  • Posts: 9
  • Country: us
For LiFePO4 you should only use one stage charging just the bulk part as soon as any cells get to your set limit say 3.55V the charging should be completely stop and only start recharging again if voltage drop below say 3.4V or so.

I've tried this approach with my own LiFePO4 setup and noticed that it tends to toggle the charging on and off.  Once the max cell voltage hits 3.55V then the charging is cutoff and the cell voltages relax.  Often, only a few minutes later all the voltages dip below 3.4V and the cycle repeats.  Is this bad for the cells?  I assume it would eventually stop but wasn't sure if this is too much stress for the cells.  Curious how you chose 3.4V?  Should I be picking a value specific to my cells?  Something just above where they tend to fall?
 

Offline electrodacusTopic starter

  • Super Contributor
  • ***
  • Posts: 1858
  • Country: ca
    • electrodacus
For LiFePO4 you should only use one stage charging just the bulk part as soon as any cells get to your set limit say 3.55V the charging should be completely stop and only start recharging again if voltage drop below say 3.4V or so.

I've tried this approach with my own LiFePO4 setup and noticed that it tends to toggle the charging on and off.  Once the max cell voltage hits 3.55V then the charging is cutoff and the cell voltages relax.  Often, only a few minutes later all the voltages dip below 3.4V and the cycle repeats.  Is this bad for the cells?  I assume it would eventually stop but wasn't sure if this is too much stress for the cells.  Curious how you chose 3.4V?  Should I be picking a value specific to my cells?  Something just above where they tend to fall?


Yes you should select this values based on your configuration and battery characteristics for best results.
Of course if you have a Load connected then battery will discharge below the set point and charging will start again. If there was no load and charging started again only a few minutes later then maybe you are charging the battery at very high current more than typical 0.25C or you have a battery with high internal resistance.
My new battery made of A123 cells (8s10p) about 185Ah usable capacity has a very small internal resistance and the limits for this where set at 3.5V charging cutoff and 3.35V charge recovery voltage.
See below a screenshot from the system.


Notice the battery charge current graph with green the 1h graph (one minute interval) there was a charge current of about 44A for about 4minutes (that is about 2.9Ah put in to battery) then at current load of about 2.835A average (blue graph 1h) it will take just over 1h before the charging will start again for another 4 minutes or so.
Of course if Load will increase there will be more frequent charge cycles
That 2.9Ah charge discharge cycle is about 1.5% DOD and since that repeats about once an hour it means there will probably be 5 or 6 of this cycles per day so not a big deal since at such low DOD the battery can perform few hundred thousand of this cycles so absolutely no problem.
With this quality LiFePO4 batteries used in offgrid energy storage the most significant degradation is with aging and not cycling.

Offline scttnlsn

  • Newbie
  • Posts: 9
  • Country: us
Re: Solar BMS a solar charge controller for LiFePO4
« Reply #79 on: August 06, 2017, 11:03:05 pm »
Thanks, that's very helpful info.  I'm using 4 of these cells http://www.batteryspace.com/lifepo4-prismatic-module-3-2v-20-ah-10c-rate-64-wh-6-0---un38-3-passed-dgr.aspx in series and charging with up to 10A from a PV panel.  The cells are fairly new so I would have expected a fairly low internal resistance (though I have not measured...probably something I should check).  Even at 5A (0.25C) I see rather quick charge toggling so I'll play around with the voltage thresholds a bit.
 

Offline asimlink

  • Newbie
  • Posts: 4
  • Country: pk
Re: Solar BMS a solar charge controller for LiFePO4
« Reply #80 on: August 22, 2017, 04:48:37 pm »
Hi, I have seen SBMS100 schematic mentioned in SBMS100 manual pdf. while looking at the Balancing MOSFET i find that Qb6 MOSFET node after the series 12+12 resistors seems to have repeated BAT5 label. Shouldnt it be BAT6? All of the next Balancing MOSFETS have shifted net lable issue. For example GCB7 MOSFET QB7 has BAT6 and so on.

Please have a look at the attached snapshot of the SBMS100 scheme:



Can someone suggest if it is a mistake or am i unable to understand the scheme the way it is implemented?

 

Offline electrodacusTopic starter

  • Super Contributor
  • ***
  • Posts: 1858
  • Country: ca
    • electrodacus
Re: Solar BMS a solar charge controller for LiFePO4
« Reply #81 on: August 22, 2017, 07:01:51 pm »
Hi, I have seen SBMS100 schematic mentioned in SBMS100 manual pdf. while looking at the Balancing MOSFET i find that Qb6 MOSFET node after the series 12+12 resistors seems to have repeated BAT5 label. Shouldnt it be BAT6? All of the next Balancing MOSFETS have shifted net lable issue. For example GCB7 MOSFET QB7 has BAT6 and so on.

Can someone suggest if it is a mistake or am i unable to understand the scheme the way it is implemented?

There is no mistake in the schematic. You can notice that last 3 mosfets are also p-channel and this has to do with how the ISL94203 is build internally.
 

Offline asimlink

  • Newbie
  • Posts: 4
  • Country: pk
Re: Solar BMS a solar charge controller for LiFePO4
« Reply #82 on: August 23, 2017, 08:14:56 am »
You have done a wonderful job by making this open source SBMS100. Thank you for the explanation and pointing out that the last 3 mosfets are P type.

Kind Regards
 

Offline asimlink

  • Newbie
  • Posts: 4
  • Country: pk
Re: Solar BMS a solar charge controller for LiFePO4
« Reply #83 on: August 25, 2017, 07:03:36 am »
Can you please suggest what are part numbers for following components?

I dont find any part number in sch, pcb, bom or in pdf files for these components:

Page 42 or SBMS100 manual:
D17, D4 schotky diodes
D9-D12  schotky diodes
Qb1-Qb8

Page 43 or SBMS100 manual:
TVS1-TVS2 zener diodes


And can you explain where do you have connected CSI1/PWR_FLAG in the schematic i dont find any connection of CSI1/PWR_FLAG to BAT+?

Thank you
 

Offline electrodacusTopic starter

  • Super Contributor
  • ***
  • Posts: 1858
  • Country: ca
    • electrodacus
Re: Solar BMS a solar charge controller for LiFePO4
« Reply #84 on: August 25, 2017, 08:15:33 am »
Can you please suggest what are part numbers for following components?

I dont find any part number in sch, pcb, bom or in pdf files for these components:

Page 42 or SBMS100 manual:
D17, D4 schotky diodes
D9-D12  schotky diodes
Qb1-Qb8

Page 43 or SBMS100 manual:
TVS1-TVS2 zener diodes


And can you explain where do you have connected CSI1/PWR_FLAG in the schematic i dont find any connection of CSI1/PWR_FLAG to BAT+?

Thank you

All diodes are 60V 1A except for D15 and D16 that need to be 100V rated.
Q1 to Q5 are n-channel Q6 to Q8 are p-channel and the rating will depend on what sort of current you want to use for cell balancing. if you use the same 24Ohm then they should be rated at least 300mA
TVS1 and TVS2 as seen in schematic are 36V and the pulse power ration will depend on what sort of current you expect to be using on the PV input in my case I used 1500W TVS
CSI1 is connected to Battery+ trough a short 2m max 16mm2 cable (Is the P3 high power connector on SBMS labeled Batt+) not the same with BAT+ in the schematic that is just connected to Battery+ (cell8 positive terminal) trough a thin cell balancing cable.

I hope you will not try to build an SBMS since that will be extremely expensive. I usually spend well over $1000 when I build a prototype with big part of that for all the PCB's needed but also the components since they are in low volume.
Maybe you are just building a different device based in ISL94203. 

Offline asimlink

  • Newbie
  • Posts: 4
  • Country: pk
Re: Solar BMS a solar charge controller for LiFePO4
« Reply #85 on: August 25, 2017, 05:20:32 pm »
Thanks for the explanation.

No I am not rebuilding SBMS100. I am working on a charger which will use single 300W panel and will have Li-ION batteries. I was studying ICL94203 chip and its reference design (AN1952.PDF) and i came across SBMS100. SBMS100 looks great reference design to me. I also liked the use of Aluminium PCB for heat dissipation.




 

Offline electrodacusTopic starter

  • Super Contributor
  • ***
  • Posts: 1858
  • Country: ca
    • electrodacus
Re: Solar BMS a solar charge controller for LiFePO4
« Reply #86 on: August 25, 2017, 05:29:57 pm »
Thanks for the explanation.

No I am not rebuilding SBMS100. I am working on a charger which will use single 300W panel and will have Li-ION batteries. I was studying ICL94203 chip and its reference design (AN1952.PDF) and i came across SBMS100. SBMS100 looks great reference design to me. I also liked the use of Aluminium PCB for heat dissipation.

ISL94203 is great except for the very small TQFN package with just 0.4mm pin pitch. Good luck with the charger.

Offline fourtytwo42

  • Super Contributor
  • ***
  • Posts: 1183
  • Country: gb
  • Interested in all things green/ECO NOT political
Re: Solar BMS a solar charge controller for LiFePO4
« Reply #87 on: August 27, 2017, 08:00:47 pm »
Excuse me if this is slightly off topic but how is this open source hardware if you are charging people money to obtain the details of it ?
 

Offline electrodacusTopic starter

  • Super Contributor
  • ***
  • Posts: 1858
  • Country: ca
    • electrodacus
Re: Solar BMS a solar charge controller for LiFePO4
« Reply #88 on: August 27, 2017, 08:08:35 pm »
Excuse me if this is slightly off topic but how is this open source hardware if you are charging people money to obtain the details of it ?

I do not charge any money this are fully open source for anyone. I do sell the SBMS since is less expensive than you building just one or two SBMS for your own use because of the low volume (and that is not including the work and equipment needed just the parts). 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf