Solar Freaks

Hi there,
Can someone tell me howit is possible to set a voltage level for a pv circuit. I.e. When I want to use a charge controller I want the charge controller start working when the voltage delivered from the solar cells is at least higher then a specified level. I want to avoid that at low light conditions the controller is maybe in a "semi powered" state.

On MPPT controller I'm working on (11.55V 4.8W panel charging 6V SLA) I didn't used voltage level, instead I measured input power while solar panel was connected directly to the battery. I used several low power states. If PV power is below 300mW and higher that 250mW, runing MPPT is not viable so configure mosfet bridge to connect solar panel to battery but don't go to sleep (actively monitor power change). if Power is below 200mW, do the same but go to sleep. If PV power is 0W shutdown mosfet bridge and shunt monitors and go to sleep. During sleep MCU wakes every 8 seconds to check PV power in direct connect mode and changes state if conditions are met.

This way during zero power state current consumption from battery was 200uA.

P.S. Hi everyone! I'm new to this forum.

I want to use a DC-DC converter to charge a battery with the solar cells. For this I'd like to use a buck converter wich needs higher voltage than the output voltage. Powering a 4 cell Lipo battery I'd need at least 16.8 volts minimum from my solar cells, so I need something to set this level before entering the voltage into the converter as the converter is set to this voltage level.

By the way, how is it possible to tell the converter to let the solar cells output a specific volatge (i.e. The MPP) and output a static set voltage? If I change the duty cycle of the dcdc converter the ratio of input to output voltage will change, but if i want a static voltage of 16.8v at output and the solar cells deliver maybe 24v but the MPP would be at 18v; how can I achieve that the converter will work with 16.8v output and 18v input instead of 24v from solar?
Or is there another way to realize thise?

Word of advice, don't use Li-Poly batteries unless absolutely necessary. They are way to sensitive to be charged by a DIY device. Charging Li-Poly batteries in series is also a problem. Since Li-Poly (Li-Ion batteries in generals) are very sensitive to overcharge you would need to balance individual cells during charging in order to avoid catastrophic failure. Take this for example, you are charging two cells in series, one cell is 3.8V, other is 4V, charger will run until voltage is risen to 8.4V, but since cells weren't balanced, one with higher starting voltage will end up being overcharged, damaged and at risk of exploding and setting everything around it on fire. Charging Li-Ion cells in series should be done with dedicated charger ICs.

As for the charging with DC-DC converter the answer is no. You can't have fixed voltage on DC-DC converter output while maintaining MPP voltage on solar panel. As long as battery can accept current from converter without its voltage crossing certain threshold, charger can operate in pure MPPT mode. When the voltage reaches that threshold charger must run voltage regulation, instead of MPPT algorithm.

I'll give you example with my charger (here's short video on YT http://www.youtube.com/watch?v=pm8U1--0Bnw). If the battery (6V 4.5A VRLA battery) is drained, at full sun charger will manage to funnel as much as 0.8A into it (from 4.8W solar panel it manages to extract 5.7W). Charge voltage is temperature compensated but lets assume it is 7V. As the battery is charged, its voltage will steadily rise until it reaches 7V. When that happens charger will switch from MPPT to voltage regulation and maintain voltage at that level. MPPT will kick in only if the battery voltage drops below 7V, in order to bring panel to the MPP voltage or as close to the MPP voltage, where voltage regulation can take over. As the time passes voltage will be maintained at 7V, but current that goes into battery will steadily fall until the point where won't drop any further. VRLA battery can be maintained at that level indefinitely (this one has floating life rated at 5 years).

If you want fixed voltage for Li-Polys, you can put dedicated 3 cell Li-Ion charger/regulator IC at the output of DC-DC converter in order to charge batteries (relatively) safely. But if you put linear Li-Ion charger IC, MPPT running on DC-DC converter will be pretty much useless since charger IC will dissipate extra power as heat. As a solution DC-DC Li-ion charger would be better solution. Linear Technologies for example has DC-DC buck Li-Ion charger with fixed point MPPT but only for 3 cells http://www.linear.com/product/LT3652. They are good but can be a real pain to solder (QFN-esque package).

Thanks for the great reply!
So as far as I understand the DC-DC converter can not be set on a specific output voltage when doing MPPT, but as long as the output voltage from the converter is higher than the bat voltage the battery will be charged, correct? And the output voltage will change with lightening conditions as the MPP will change too. When reaching the bat voltage of 16.8v (4 cell Lipo) I'd have to change from MPPT to voltage regulation so that the converter will output 16.8v and no current will flow from the solar panel to the battery, right? And I'd also have to shut down the converter if the panel voltage is less than the bat voltage because of low sun.

I'm aware of the problems using Lipos, so I was planing to check not only the bat voltage but also the cell voltage of each cell. So I can see when a cell is reaching the max cell voltage of 4.2v and switch the operating state of the converter to voltage regulation.

Could you actually recommend an operating voltage for a solar panel if using a battery with max 16.8v? I haven't bought any solar cells yet and I don't know how much voltage a panel should provide for a efficient use with such an MPPT.

By the way, is there a reason why you chose a incremantal conductance algorithm for your charger and not a P&O method?
Thanks for your help!

When charging battery you don't have to think about output voltage, because battery voltage will be output voltage. As long as the input voltage of the solar panel is greater than battery voltage DC-DC step down converter will charge the battery. Basic equation is as follows: Pout=Pin*Efficiency, output power will be equal to input power multiplied by efficiency factor (0-0.99).
For example, if input power is 0.5A x 12V = 6W, output power is 6Wx0.95 = 5.7W, but since battery voltage is lower (6V) than solar panel voltage, charging current will be 5.7W/6V=0.95A. Basically what DC-DC step down converter will do is to convert higher voltage and lower current to lower voltage and higher current, vice versa for step up converter. If the 6V battery is discharged its voltage will be lover, hence charging current will be higher. As the battery charges, voltage will rise steadily, but in the same time charging current will fall. When battery voltage reaches charging voltage, charging process is not yet finished. As the time passes and battery voltage is maintained at charging voltage, and charging current will fall. When it drops below certain level battery is fully charged and process can be terminated.

In theory input voltage can be many times higher than output voltage, but that is not true in practice. Rule of the thumb for step down converters is to have desired output voltage at 50% duty cycle. What it means that input voltage should be twice as high as output voltage. So for your application you should chose solar panel with MPP voltage of 14.8V x 2 = 29.6V. Note that I used discharged Li-Poly cell voltage to calculate input voltage.

As for the MPPT algorithm, I implemented and tested both P&O and incremental conductance method, and later worked waaaay better.

P.S. Be sure to google for Li-Poly balancing.

But what is charging voltage then? You said output voltage is battery voltage.
I thought converter output voltage needs to be higher so current can flow from converter output to battery. When battery voltage is sensed as max voltage the output of the converter will be set to this voltage by a change of the Duty cycle, therefor no current will flow anymore and charging is stopped until the battery voltage will fall again and the converter output can be set higher again to charge, right?

If the panel needs MPP voltage of ~30v, Voc needs to be a little higher ~33v i think.
Can you tell me what switching frequency you are using for your converter, as I've read that lower frequency is more efficient as higher.
Thanks!

There is no fixed charging voltage at converter output, MPPT algorithm will vary duty cycle so that maximum power can be extracted from solar panel. Voltage at output will vary and it will be higher than battery voltage (if MPPT algorithm is working well it has to be), but you won't be able to measure it because all excess voltage will be converted into charging current that flows into the battery. Healthy battery has very low internal resistance, lets say that at some point during charging it is 100mΩ, so even 50mV above battery voltage will drive 0.5A of current into it. When charging voltage is reached (fore example 4.2V for Li-Ion), current that is flowing in that moment into the battery will be significant. You will have to regulate that voltage by lowering duty cycle which will decrease converters output voltage. Again, you won't be able to measure it, but you will see that charging current decreased, which means that output voltage is lower.

As for the frequency, my converter is operating at 250kHz. I've chosen higher frequency so that inductor value is lower and thus it is physically smaller. Measured efficiency is 92-93%, but more that half of that 8-7% loss is dissipated on schottky reverse blocking diode at the converters output. Without it, efficiency would be 96 to 97%.

Yes, I think we mean the same now, so I understand the principle now - by setting the converter output voltage to the same voltage as the battery got there will be no difference anymore, therefore the current will be zero and battery is stopped being charged.

Have you thought of changing The diode with another mosfet to increase efficiency? I've also read that operating with higher duty cycle is more efficient than lower. You think battery to panel voltage ratio should be more like 70-75% then?
As I'm using this application for rc purposes I probably won't use such high switching frequency to lower rf disturbances created by the LC; maybe 50kHz will be adequate.

Best place for that diode is at solar panel out, before DC-DC converter. Solar panel current is lower than DC-DC converter output current thus voltage drop and power loss is lower. But since I'm using synchronous buck topology, I put it there as a safety measure. Because of some bug in the code or MCU failure I might end up with mosfet bridge entering unwanted state. For example during normal operation high side mosfet is turned off while low side mosfet is turned on, if MCU for some reason get stuck/locked up in that state my battery will be connected directly to the ground and shorted. In my books that rates as catastrophic failure. This way battery and some other system that runs on it (I intended this charger to be autonomous power supply for some remote sensor station or something like that) is safe and operational until battery discharge and die.

You shouldn't go under 50% duty cycle, anything above is OK. According to the data logged during testing of my device, duty cycle varied from 53% to 77% during DC-DC operation (MPPT and PID voltage regulation) and during that time efficiency remained pretty much constant, 92-93%.

I thought u use the diode as a nonsynchronous converter, how do you place the diode at the panel out then? Afaik if you are running the converter in synchronous mode there is no protection of a mcu failure, so I don't understand your setup exactly.

I've found this http://www.tmenet.com/products/lithium-charger/xtrema-balancer as balancer. I think this should do the job of keeping the cells at same voltage level, what do you think?

It's configuration like this:


Diode D2 is the reverse blocking diode I mentioned. It prevents battery from being shorted via low side n channel mosfet and in the same time keeps battery from being drained through panel via high side p channel mosfet body diode.

I hope you'll manage to integrate that balancer into your circuit. I didn't charged Li-Ion in series nor had any need to do it so far. I'm just reminding you to take everything into consideration when dealing with Li-Poly batteries.

Great, thanks!

But why did you keep D1?

Schottky D1 is kept because of intrinsic body diode on low side n-mos Q2. During dead time, when when both mosfets are turned off, circuit will behave like non synchronous buck converter because of body diode on Q2. That body diode is very slow switcher, and if allowed to conduct it would introduce additional losses into equation. D1 will counter that by not allowing body diode to conduct.

Some mosfets made just for DC-DC converters have built in schottky diode, so you don't have to worry about that, but they are pretty rare.

BTW, DC-DC converter I made uses p-mos n-mos push pull configuration because of low solar panel operating voltage. Since you will be using panel around 30V Vmp, it is recommended to use n-mos only push pull with half bridge mosfet driver (like IR2109 or LTC4444).

Thanks for all that Information MilanL.
I guess thats all I had to know =D
Greetings