Load Google Translate Hi all
I'm trying to figure out what the best solution for charging a 6V lead acid battery from a 9V peak power / 10.8V O.C circuit solar panel. On cloudy days, the 5W solar panel will put out anything from 20 - 60mA at 6V (tied to the battery). In the winter in some locations, these conditions will persist for weeks at a time, so making the most of the current all of the time rather than making the most when it is sunny is more important.
These are the shortlist of parts:
Linear LT3652 LT3652 - specifically for solar applications
TI BQ24650EVM-639. BQ24650EVM-639. - specifically for solar applications
TI BQ24450 - shunt regulator for lead acids
TI BQ2031 BQ2031 - switching regulator for lead acids
The Linear LT3652 LT3652 cannot be used because although the supply current for the chip is listed as 'typically 2.5mA'. It also mentions that the BOOST supply current is typically 20mA, meaning 22.5mA will be wasted before anything gets charged.
The BQ24650EVM-639. BQ24650EVM-639. can't be used because of the current required for it to operate in charging mode. It is listed as 'Adapter supply current' and says typically 25mA. This means on a cloudy day, lots of the energy will be eaten up by the chip.
Am I misinterpreting either of these?
The BQ24450 looks promising. It is a basic shunt regulated design and I have a few questions:
1/ I assume it would require a minimum input voltage of:the battery charging voltage + the diode drop + the transistor / FET drop + the sense resistor(s) drop. I'm planning on charging at no more than 0.8A (with 2 x 5W panels). Can anybody think of any other reason why the BQ24450 would not be suitable for this application?
The BQ2031 BQ2031 also looks like it may be a solution. I have read the application note for MPPT with the bq2031 bq2031- http://focus.ti.com/lit/an/slva378/slva378.pdf
1/ Apart form the 2mA current draw + transistor drive current, are there any other hidden power requirements from the charging circuit (without MPPT adjustment).
2/ What kind of efficiencies should I expect in my application?
Which of these would you suggest? What other alternatives are there?
I don't expect using a switching design vs a shunt design to give a large advantage - Assuming full sunlight, a shunt system this utilises approximately 68% of the energy (vs MPP) from the panel charging directly (running at 6V would mean 0.275V / cell vs MPP of 0.42V / cell, so reading off approximately from the chart in this datasheet http://focus.ti.com/lit/an/slva378/slva378.pdf and comparing 0.42 x 0.036 = 0.0152 vs 0.275 x 0.038 = 0.0104 per cell). Switching designs would have to exceed this efficiency figure across the range, especially at the low end.
I am open to other solutions to this problem and any help would be appreciated.
Many thanks
Oli
Hi Oli
Not sure I'm really understanding your problem but do have some experience with low power 12V installations. My understanding is that the solar cell (within the bounds of it's maximum O/C voltage) behaves as a constant current generator whose output is proportional to the light intensity. Therefore, connection to a lead acid type battery may simply achieved using for example a LDO low quiescent current linear regulator to limit overcharging, crude but simple. For example I've used an LP2952 with output blocking diode (shottky). Losses are the volt drop in any non return diode(s), the LDO forward voltage drop, the LDO quiescent current.
A far bigger problem fo me has been over discharging of the battery. This prompted a more elegant solution using back to back P channel MOSFETs controlled by a window comparator, disconnecting the battery from the charging and discharging circuit, preventing both overcharge and discharge. The window comparator and voltage reference/sensing are powered from the solar panel side of the MOSFETs, use low power devices, and achieve less than 1mA waisted in the circuit. Losses are largely confined to the comparator circuit quiescent current and reverse blocking diode in the panel and a very small MOSFET bias when the battery is connected to load/panel.
Hope this helps, good luck
James
Hi James, thanks for your reply. The real issue is using a LDO on it's own lacks several things:
1/ Temperature compensation - this is required so that it is not overcharged in hot weather or not charged adequately in cold weather. It can cause gassing if overcharged and may permanently affect the batteries capacity.
2/ Quick charge and float charge. If the LDO is set to just the float voltage, then when there is lots of sun available, it won't take full advantage of it as it may not charge as quick as it could. Charger IC's normally fast charge at a higher voltage until charged and then drop back to the float voltage. I believe a charge above the float voltage occasionally can help rebalance the cells internally if they are normally charged at float voltage too.
3/ Pre-conditioning charge - If the battery has been discharged more than optimally, then charging it at the full current may just cause heating. Trickle charging it until it reaches a certain potential will avoid this problem.
I need to look after the battery as much as possible, so a charge controller IC is preferable. Although I could roll my own solution, it will take significantly more time and effort and I just don't believe there isn't a solution out there already!
Could you provide more detail on your undervoltage lockout implementation?
Thanks
Oli
Hi Oli, did say the LDO was a crude solution, ...no temp' compensation, which to be honest hasn't troubled automotive applications of Lead Acid batteries over the years, sealed Lead Acid are less forgiving. Quick charge/float charge not an issue for me as large capacity battery, so the 5W panel never generates enough current to be classed as quick.
Bear in mind the LDO will still pass current even though it's not reached the float/set voltage.
Might take a little while on the other circuit. The trick with under voltage is to bias the back to back P MOSFETs via a high value pull-up resistor on the battery side. The window comparator is powered from the other (load/charge) side. For the battery to be connected to load/charge, the dual P MOSFET switch gate is grounded via open collector, small current from battery through bias resistor. The window comparator is sensing the battery voltage on the load/charge side, when out of limits the open collector ground becomes open, the two MOSFETs gates go high to battery volts, then the window comparator circuit is isolated and no power is being drained from the battery (other than negligable leakage through the MOSFETs). The whole arrangement is then effectively latched off until the charging/load supply rises above the lower threshold. Note this may also be attempting to supply the load, which if greater than the panels current capacity will never occur, in which case you'll need an additional control circuitry to disconnect the load. In my case not a problem as the load is only a few mA.
A similar issue arises at overcharge, where the charge/load supply can suddenly rise to the panels open circuit voltage as the battery is disconnected, something to be aware of if your load is less than the panels current capacity, may require a dump load/clamping circuit to protect equipment connected. Not an issue if the load is greater than the panels output, as all current will be going to the load anyway and some from the battery.
Temperature compensation and variable charge rates are not considered as before, although could be. I'm generally using SLA batteries of 7.2Ahr to 19Ahr with 5-6W panels and very light but continuous loads, 4mA typically and tolerant of 18V input voltage. Installations have run without maintenance for several years, the last failure was due to water damage of the panel.
Hi Oli,
perhaps you can use UC2906 from TI:
http://www.ti.com/product/uc2906
http://www.ti.com/lit/ds/slus186c/slus186c.pdf
It has temperature compensation, quick charge and float charge (called Dual Level Float Charge) and pre-conditioning charge. Supply current 1.6mA typical.
In case of bq2031 bq2031 you should take into account +5V regulator and op-amp supply currents. I have not found any information about efficiency but I think you can expect 80+ % as output voltage is low and diodes are used in buck converer. If there would be synchronous converter, you can expect better efficiency.
bq24450 is nearly same as UC2906. In standard application power switch driving current (DRVE pin) flows to load. Instead of reverse protection diode you can use p-MOSFET (lower voltage drop).
Yes, I tested with the BQ24450 (it is the same part as UC2906 - rebranded after TI bought unitrode) and it seems to work as described with p type mosfet. Thanks for your reply anyway.
© 2009 Premier Farnell plc. All Rights Reserved
Premier Farnell plc, registered in England and Wales (no 00876412), registered office: Farnell House, Forge Lane, Leeds LS12 2NE