charging LiFePo4 with solar cells

[commanda]

With large format LiFePo4, I really think you're over-thinking the problem by trying to add active balancers. My scooter uses 16s 40 AHr TS LFP, and after almost 2 years of use, the time from any shunt low to all shunts low is still only several minutes (at about 0.8 A). So long as you are reaching full charge regularly (possibly weekly in a PV situation) the cells track very well. I have considered adding a stopwatch to time any shunt low to all shunts low, as a measure of pack balance.

The idea of doing a weekly balance charge really is a hangover from the lead-acid days.

 

[Mike B]

Amanda, sorry for any confusion, but I thought you had brought up the end of charge issue:

Once you get any shunt low, you need to limit the charge current to whatever your shunts can handle.
I KW of solar will make 70 - 80 amps at 12 volts. Not trivial to handle.

Your comments point out the problem well I thought. With larger systems the end of charge shunting approach would require some big resistors and switching. You will have the same problem with a 4 kw PV system at 48V.

Automatically balancing a 2 to 3 kwh battery bank for years on end? It needs a simple, robust approach. Switching some big capacitors back and forth makes sense. Add an independent monitoring circuit that compares cell voltages and alarms if any are more than 0.0X volts difference.

How to switch them back and forth?

Anyway, that would eliminate the need for shunting at end of charge, and the PV controller would just taper off normally. simple.
Mike

 

[myzter]

@texaspyro passive balancer with a single cap switching between cells.. please share ?

 

[texaspyro]

@texaspyro passive balancer with a single cap switching between cells.. please share ?

The balancer that I am playing with has one cap per cell (well N-1 caps per N cells) plus a P channel and N channel FET per cell. It is essentially the same circuit in the Atmel ATA6870 app note http://www.atmel.com/dyn/resources/prod_documents/doc9184.pdf

I am doing a capacitively coupled gate drive with independent clock lines for the N and P channel fets. This eliminates the possibility of energy sapping "shoot-through" where both fets are on at the same time. The gate drive signals are generated by a small 8-pin micro.

The unit is exceedingly dumb. Generate clocks and the caps are switched between adjacent cells. It makes no measurements, has no precision components. Current flows into the cap from the higher voltage cell, out of the cap to the lower voltage cell. It can run when the pack is charging, discharging, sitting idle, or watching TV.

Current is I = C * dV / dT The bigger the voltage difference or capacitor, the more current can be transferred. There is a practical limit on cap size. You want the cap to fully charge when it is switched across the higher voltage cell. Too high a switching frequency wastes energy in switching losses and gate drive capacitance. Still remains to be seen what the best cap size and switching freq is.

The unit can be built to operate totally independently, but I was going to integrate it with a BMS that has a small microprocessor on each cell. The micros talk to a central controller via an opto-isolated serial bus. I was going to clock the serial bus synchronously with the cell balance clocks.

There are also inductive balancers that can do much higher currents. They are more complicated and failure prone... some failures wind up putting a inductor (i.e. dead short) across the cells.

 

[commanda]

Amanda, sorry for any confusion, but I thought you had brought up the end of charge issue:

Once you get any shunt low, you need to limit the charge current to whatever your shunts can handle.
I KW of solar will make 70 - 80 amps at 12 volts. Not trivial to handle.

Mike

Mike,

The two lines you quoted were not meant to be read as one. I can't imagine anyone would try to shunt at 80 amps. My current pv setup is 12 volt 180 AHr, 2.3 KWhr. Bulk charge is whatever the pv can generate, about 7 point something amps. Any shunt low switches off the bulk charge switch (implemented with a solid-state relay) and leaves a constant current source of just under one amp. Time from this point to all shunts low is still measured in minutes. In the pv setup, all shunts low doesn't actually do anything. It stays in constant current mode indefinitely. The shunts only have to carry this limited current.

I'm still not seeing the need for any sort of active balancer. When charging, they all reach full charge at pretty much the same time. When discharging, the voltage curve remains essentially flat until the last 10% of capacity, at which point it falls pretty quickly. Even if one cell had less overall capacity, it wouldn't really become obvious until it was almost gone. It would take some very serious charge switching to prop it up. I think the easier approach is to be conservative, and not try & use 100% DOD.

Amanda

 

[hollisterco]

how bout those new solar roof shingles DOW is comin up with? i dont know when theyre starting the rollout but im sure that will be definitely cheap, or those inkjet printed solar panels compared to the silicone ones, theyre trying to get it at $1 per watt, dont know how far along they are, anyone heard anything new from nanosolar? they where pretty loud a year ago about all this but havent heard much since, they might already be selling but only to industry, no word on prices AFAIK
nanosolar.com

 

[texaspyro]

I did a little test with a simple switched capacitor balancer. First I fully charged two 2300 mAh A123 cells to 3.670V. Next I drained 1000 mAh out of one cell. It dropped to 3.261V. I switched on the (highly unoptimized) balancer. It started at around 40 mA, quickly dropped to 27 mA with the low cell at 3.30V After 3 hours the cells were at 3.36V and 3.31V with 4 mA of balance current. After 8 hours they were at 3.35V and 3.31V and 3 mA of current. At 10 hours the cells were at 3.353V and 3.313V and 3 mA of balance current. An initial 400 mV voltage imbalance was reduced to 40 mV.

After 10 hours, I disconnected the cells and discharged them at 5C (12 amps) on my battery analyzer. Around 118 mAh of charge had been transferred from the high cell to the low cell.

This test was with no load on the cells. LiFePO cells with rather different states of charge can have very close open circuit voltages and a capacitive balancer balances the cell voltage. Once the cells are at the same voltage, balancing stops. You would get a better SOC balance (not just voltage balance) by running the balancer while the pack was being charged or discharged where the cell voltage better represents state of charge.

 

[commanda]

Test is not very well designed. The 3.67 volts is hot off the charger. This will drop closer to the nominal 3.2 volts once a small amount of charge has been used.
Try this. Charge 2 cells fully. Drain 1000 mAHr out of one cell. Drain 500 mAHr out of the other. Now repeat your test.

In terms of cost/benefit ratio, I don't see the point of any sort of balancing, other than the full charge shunt. In the pv off-grid scenario, we're talking BIG batteries. My current setup is 180 AHr, and that really is a small setup for this application. So long as you can reach full charge periodically ( I have suggested weekly), track your remaining capacity with an accumulating AHr meter, and stay away from 100% DOD, you should be sweet. A periodic full charge brings all the cells to the same SOC. They track really well.

 

[texaspyro]

Try this. Charge 2 cells fully. Drain 1000 mAHr out of one cell. Drain 500 mAHr out of the other. Now repeat your test.

In the test that I did, the surface charge was drained in the first half hour and the cells were then at their nominal voltage.

 

[texaspyro]

I did another test on the switched capacitor balancer. This time with it connected during a charge. Very interesting results... http://endless-sphere.com/forums/viewtopic.php?f=14&t=20864&start=0

 

[Jeremy Harris]

I've been testing the solar charge system on my electric boat this week. I used four 50 watt nominal Sunflex panels (http://www.sunflexsolar.com/product/sunflex-50w-flexible-solar-panel-50-watts-free-shipping-1), connected in parallel (through isolating ideal diodes) to a Fatmax maximum power point tracker (http://solarmppt.com/product_info.php/cPath/23/products_id/33). The output from the MPPT feeds directly to the 12.8V nominal, 80Ah, LiFePO4 pack (4S 8P of 10 Ah Headway cells). The BMS is my own version of the original Goodrum/Fechter shunt regulator, using big power transistors driven by the LM431. The 'any shunt low' line from the optoisolators drives a FET that normally shorts a 5 ohm, 10 watt, resistor that is fitted to the charge ground line. As soon as one shunt activates, indicating that one bank of cells has reached 3.65V, the FET turns off, so the MPPT tries to drive current into the pack via the 5 ohm resistor. This limits charge current to around 1 amp maximum, enough to balance the pack reasonably quickly, yet keep the shunt power dissipation within limits (the shunt transistors are mounted on a big heatsink). With all the shunts activated the charge current is effectively limited to less than 1 amp and flows primarily through the shunts, not the cells.

So far this system seems to work pretty well. The MPPT gets about 10 to 15% more usable power from the solar panels than a direct connection, which is very worthwhile, particularly in overcast conditions. On a very cloudy day last week the panels (which are mounted horizontally on a canopy above the boat) were delivering over 60 watts, almost enough to provide cruise power. In bright sunshine I'm expecting well over 100 watts, more than enough to cruise without worrying about battery depletion.

Its taken some experimentation to get this far, but I'm generally pleased with the solar performance.

Jeremy

 

[RTA]

Hello,

For my Electric tadpole bike:
- 250W minimotor (goldenmotor)
- 36V lifepo4 15 aH (pingbattery)

I'm looking valuable solar power system for charging the battery during ride. How about 2 pcs 10 w semiflex monocrtystal panel?, I can be upgrade quantitys if needed. Problem is solar power controller, what kind of? How i should do it cost effecienfy? Is there harm for charging LiFePO battery during riding?

 

[commanda]

Hello,

For my Electric tadpole bike:
- 250W minimotor (goldenmotor)
- 36V lifepo4 15 aH (pingbattery)

I'm looking valuable solar power system for charging the battery during ride. How about 2 pcs 10 w semiflex monocrtystal panel?, I can be upgrade quantitys if needed. Problem is solar power controller, what kind of? How i should do it cost effecienfy? Is there harm for charging LiFePO battery during riding?

RTA;

There are a number of issues related to carrying panels with you.
Weight
Fragility
Pointing them at the sun whilst travelling
The sheer area of panels required to be meaningful (with attendant wind drag issues)
In practice, your panel will put out significantly less than the sticker rating might lead you to believe.

Your battery pack contains (36 * 15) 540 WattHrs. So, 2 x 10 watt panels would take 27 hours (plus) to charge it completely.

Amanda

 

[RTA]

Hello,

For my Electric tadpole bike:
- 250W minimotor (goldenmotor)
- 36V lifepo4 15 aH (pingbattery)

I'm looking valuable solar power system for charging the battery during ride. How about 2 pcs 10 w semiflex monocrtystal panel?, I can be upgrade quantitys if needed. Problem is solar power controller, what kind of? How i should do it cost effecienfy? Is there harm for charging LiFePO battery during riding?

RTA;

There are a number of issues related to carrying panels with you.
Weight
Fragility
Pointing them at the sun whilst travelling
The sheer area of panels required to be meaningful (with attendant wind drag issues)
In practice, your panel will put out significantly less than the sticker rating might lead you to believe.

Your battery pack contains (36 * 15) 540 WattHrs. So, 2 x 10 watt panels would take 27 hours (plus) to charge it completely.

Amanda

Thanks Amanda,

I have an idea to charge same time when i'm cycling long distances not cycling only with panels. But i need to think about this solar panel once again.

What kind of converter/regulator solution is sensible from 12 Vdc solar panel to 36 Vdc LiFePO4 battery?

 

[commanda]

Use panels in sets of 3 in series to charge 36 volts directly.

Amanda

 

[RTA]

Use panels in sets of 3 in series to charge 36 volts directly.

Amanda

So i can use example following solar converter: Solar Converters PT 36-5A, 36V MPPT Charge Controller, and this directry connected to battery charching plug.

http://www.altestore.com/store/Charge-Controllers/Solar-Charge-Controllers/MPPT-Solar-Charge-Controllers/Solar-Converters-Charge-Controllers/Solar-Converters-PT-36-5A-36V-MPPT-Charge-Controller/p1380/

 

[myzter]

I think U have two options:

1) Constant Voltage Regulators -Up Converters
ie. 12 V panel to 36 V battery @ 4 amps

2) depending on battery chemistry.. LIPO or LiFePO4
charge only two cells at one time from a small onboard PV array (4v - 4.5v x2) / minus diode loss of 0.3v
this would require a control panel with SPDT rocker/toggle switches

However, the choice #1 is easiest by far because you can use off the shelf parts.. wire it up and done.. Without building your own panels as in the choice #2 you will be forced to absorb the issues such as weight etc....

I have good results with my VIPV (vehicle Integrated Photovoltaics) only because I was able to build them super lightweight, and integrate them as aerodynamic wind cutters. The panels themselves are fairly durable (aluminium & polycarbonate) but because of the thin film topsheet my PV cells are not protected. Lucky cells are pretty cheap $3 per 4 watt cell (delivered to Canada) that Im not to worried. I choose not to protect the cells by using a heavier top sheet not just because of weight but PV output is reduced by atleast 1 amp by using standard polycarbonate which is typically 88% light transparent. I did test some 94% material but I still measure a big drop in Amp output, so that why I use a 96-97% thin film that is practically weightless.

.. the pic is old and does not show recent modifications.

 

[RTA]

I think U have two options:

1) Constant Voltage Regulators -Up Converters
ie. 12 V panel to 36 V battery @ 4 amps

2) depending on battery chemistry.. LIPO or LiFePO4
charge only two cells at one time from a small onboard PV array (4v - 4.5v x2) / minus diode loss of 0.3v
this would require a control panel with SPDT rocker/toggle switches

However, the choice #1 is easiest by far because you can use off the shelf parts.. wire it up and done.. Without building your own panels as in the choice #2 you will be forced to absorb the issues such as weight etc....

I have good results with my VIPV (vehicle Integrated Photovoltaics) only because I was able to build them super lightweight, and integrate them as aerodynamic wind cutters. The panels themselves are fairly durable (aluminium & polycarbonate) but because of the thin film topsheet my PV cells are not protected. Lucky cells are pretty cheap $3 per 4 watt cell (delivered to Canada) that Im not to worried. I choose not to protect the cells by using a heavier top sheet not just because of weight but PV output is reduced by atleast 1 amp by using standard polycarbonate which is typically 88% light transparent. I did test some 94% material but I still measure a big drop in Amp output, so that why I use a 96-97% thin film that is practically weightless.

.. the pic is old and does not show recent modifications.

Thanks myster,

So I think that example following regulator is suitable for 12 Vdc converter. Even its quite expensive.

http://www.altestore.com/store/DC-to-DC-Voltage-Converters/Constant-Voltage-Regulator/Up-Converter-Constant-Voltage-Regulator/Solar-Converters-Cv1236-2-12-36V-2A-Voltage-Reguator/p1241/

Still i'm wondering to do 3x12Vdc panel to series, is it work at all ? Of course weight is one thing what need to think. Without solar system my tadpole trike weight will be 25 kg + my weight 73 kg = 93 kg (213 lb) + solar system ?

 

[myzter]

Do you have the dimensions of the panels ?
what is the output of the panels - volts and amps ?
have a web page for them ??

The 12 Vdc converters on that website have 1 and 2 amps output..
so if your battery pack is 2Ah it will take 1 hour basically to charge.. right ???
larger packs like 10-20Ah would take 1-2 days !!

I was looking into them aswell last year, but the output is just 2 low of amps...
The manufacturer can do custom converters but it adds $100 to the cost...

The panel orientation is a key factor.. all three need to be very similar in sun exposure..

 

[RTA]

Do you have the dimensions of the panels ?
what is the output of the panels - volts and amps ?
have a web page for them ??

The 12 Vdc converters on that website have 1 and 2 amps output..
so if your battery pack is 2Ah it will take 1 hour basically to charge.. right ???
larger packs like 10-20Ah would take 1-2 days !!

I was looking into them aswell last year, but the output is just 2 low of amps...
The manufacturer can do custom converters but it adds $100 to the cost...

The panel orientation is a key factor.. all three need to be very similar in sun exposure..

If I have small panels i can not assume that i use solar power for moving forward. It's just a extra for battery and some extra kilometers for trip.

Something like this:
http://cgi.ebay.com/10-Watt-Solar-panel-10W-12-Volt-Battery-Charger-Mono-D-/110651017745?pt=LH_DefaultDomain_0&hash=item19c3508a11
or
http://cgi.ebay.com/2pc10W-total20w-solar-panel-solar-battery-charger-12v-/170624223571?pt=LH_DefaultDomain_0&hash=item27b9ff0953

One panel optimum operating current(IMP):0.56A. I think that i dont need more than 2 amp converter.

Anyhow, i dont have experiense of solar panel charging even electrics so there might be misunderstanding.

Do you have an idea?

 

[myzter]

One panel optimum operating current(IMP):0.56A. I think that i dont need more than 2 amp converter.

that wont really help you much at all. The amps are just too little to be useful during the day light hours.
for half the price of that converter you could buy a huge powerful battery pack..
(pm me if u need help finding the right battery)

my PV panels do not send power to the motor because the voltage is just too low 9V. But I have good enough amount of current to charge a spare battery while riding - 7-8amps.

 

[RTA]

One panel optimum operating current(IMP):0.56A. I think that i dont need more than 2 amp converter.

that wont really help you much at all. The amps are just too little to be useful during the day light hours.
for half the price of that converter you could buy a huge powerful battery pack..
(pm me if u need help finding the right battery)

my PV panels do not send power to the motor because the voltage is just too low 9V. But I have good enough amount of current to charge a spare battery while riding - 7-8amps.

Thanks,

I think that it should be at least 80 W panel, light and flexible, should be also roof.

Example:
http://cgi.ebay.co.uk/80W-Mono-Flexible-solar-panel-Car-Boat-12V-charge-/270726542370?_trksid=p4340.m263&_trkparms=algo%3DDLSL%252BSIC%26its%3DI%26itu%3DUCI%252BIA%252BUA%252BFICS%252BUFI%252BDDSIC%26otn%3D8%26pmod%3D200582034734%252B200582034734%26po%3D%26ps%3D63%26clkid%3D8591740878223342704

and CV12/36-4PV solar converter works fine with this?:
http://www.altestore.com/store/Charge-Controllers/Solar-Charge-Controllers/MPPT-Solar-Charge-Controllers/Solar-Converters-Charge-Controllers/Solar-Converters-CV1236-4PV-4A-Charger/p1244/

Do i connect this directly to current 36 15aH lifepo battery what i have or should i have spare battery as you mentioned?

 

[myzter]

nice find!
should work with direct connection.. other than possibly a bypass diode but the converter probably has this built in??
and BMS for battery monitoring

 

[efried]

Hello,

here is my ride -
http://www.effiziente.st/funelecsolar.jpg - the third panel has moved to the front.
I'm willing to use a 36V charger.
http://www.altestore.com/store/Charge-Controllers/Solar-Charge-Controllers/MPPT-Solar-Charge-Controllers/Solar-Converters-Charge-Controllers/Solar-Converters-PT-36-5A-36V-MPPT-Charge-Controller/p1380/
Even if it is not for LIFEPO4, the cut off voltage might be sufficient for large batteries.
what do you think?
I know that 3 MPPTs for 12 each forming 36V in the end would be better but thats not feasible without a major project on the topic.
thanks

 

[myzter]

nice kit efried
So the output from each separate panel is different from one-another ?
Im not sure at all.. about placing three MPPT's in series for higher voltage if this even works, have you asked anyone from altestore.com ?

PV cells are like capacitors in a-way that they will not pass current... I tried to series the output from a small PV panel with a small bank of caps and the voltage added up, but the current from the caps simply dispersed into the PV cells.