# Efficient use of power by MPPT charger

Discussion in 'The Projects Forum' started by Billy4184, Jun 24, 2014.

1. ### Billy4184 Thread Starter Member

Jun 20, 2014
31
0
Hi, I am designing an MPPT solar charge controller to get more efficiency out of my system. At the moment I am at the conceptual stage and I have some questions about MPPT battery charging in general.

First of all, the way I understand it, the implementation of MPPT works via varying the resistance of the circuit such that panel output voltage * current drawn are at the maximum point on the power curve.

Then the battery charger section takes whatever voltage is 'given' by the MPPT section and drops it to within the battery charging voltage band (e.g. 13.5-14.5V for lead acid), maybe using a buck converter. Further, the way I understand it, most MPPT battery chargers use the constant-current bulk-charge method followed by a constant voltage topping method (perhaps followed by a constant voltage float charge).

Now what mystifies me is how the charger is supposed to run a constant-current bulk-charge algorithm while using the power coming to it efficiently. Because, if the charger simply regulates the output voltage via a sense resistor such that the output current remains at some predetermined constant level, that voltage is dependent on the battery characteristics and charge state, so essentially the battery charging power curve is fixed for a fixed current. So what if the MPPT suddenly gives more power? Does it get wasted? What if the MPPT produces less power? Does the constant current become not so constant?

Lastly, if there is such a thing as a constant-voltage MPPT charger (which makes more sense to me for this situation) how would transition from bulk charge to topping charge be carried out during changing conditions (changing current). Plateau detection for instance would be impossible.

Thanks very much for your help.

Billy

2. ### Billy4184 Thread Starter Member

Jun 20, 2014
31
0
Bump. Don't be afraid to break the news.

Nov 12, 2008
3,719
2,842
4. ### ErnieM AAC Fanatic!

Apr 24, 2011
7,978
1,841
Well no. Start by understanding a solar cell is a current device, and you control the current being drawn for the max power point. Follow that by understanding a lead acid battery wants a current to charge it.

So to put the two together you make a current regulated converter where you vary the current drawn from the cell (and thus delivered to the battery) to get the max power from that cell.

Obviously you move from the max power point when the cell is being topped off which you sense by the voltage across the battery. There will be times when the solar cell cannot produce as much current as the battery can take (but that current can be maximized), and also possibility that the cell is capable of producing more current then the battery should get (when you are technically wasting the power you cannot use).

So for the most part stop thinking of voltage and resistance and start thinking in terms of current.

Someone on this board who has actually built some of these mentions you gain no efficiency in using MPPT over a simple constant current approach when you are on a 12V system: the built in inefficiencies are equal to any gain you may see.

Personally for a single use device I wouldnt design anything, Id just go out and buy one. It is far cheaper. Id only build if I was either learning something or producing a version for production.

Papabravo likes this.
5. ### THE_RB AAC Fanatic!

Feb 11, 2008
5,430
1,311
That was me. MPPT on a 12v panel->12v battery is barely worth the effort unless your buck runs 95% efficiency or so.

A 24v panel->12v battery can be well worth the effort.

6. ### Billy4184 Thread Starter Member

Jun 20, 2014
31
0
I'm not sure if I made myself clear. I don't think my problem has to do with battery charging per se. @ronv that site says clearly that the bulk charging (<= 70%) is done at constant current. Since it would seem that the output current is kept constant by regulating the charging voltage (depending on the battery state etc), this suggests that the power going into the battery is the (constant) current multiplied by whatever voltage makes the battery draw that current, and is rather independent of the power being provided by the MPPT section. Is this correct?

Now unless the system leaves a considerable margin of error (wasting a lot of power and nullifying MPPT gains) when the charge current is determined, the changing availability of power from the panels will mean it is NOT constant-current, since there will be times when there is simply not enough power to maintain the current level. How does a battery charger deal with the feedback info from the battery when voltage and current are changing due to power availability? I assume that the battery takes a while to stabilise before info is meaningful.

Thank you both for you responses. @ErnieM that is interesting about MPPT not being useful when solar panels match batteries. I wasn't certain about this but was planning 24V solar and 12V batteries for that reason, as I would like maximise power on rainy/overcast days.

Edit: @THE_RB cheers

7. ### Billy4184 Thread Starter Member

Jun 20, 2014
31
0
By the way, an MPPT charger here in Aus costs \$150+...the buck converter section of mine will cost \$10 and another \$10 for the Atmega328p which I will use for MPPT tracking. Add another \$10-20 for other stuff and the savings are massive. I already have panels and batteries but the chargers are cheap and simple.

More importantly, I will be able to fix it, since everything these days comes with multi-layer boards and microscopic components that would probably vaporise if you put a soldering iron near them.

8. ### THE_RB AAC Fanatic!

Feb 11, 2008
5,430
1,311
No. The job of the MPPT is very simple; to provide the MAXIMUM current possible into the battery, for any level of insolation (sunlight level). So it runs the solar panel + buck converter in a way that give max current into the battery. That is its sole purpose in life.

When the battery voltage gets high (as it gets more charged) the circuit might ALSO perform voltage regulation. That is NOT an MPPT function, it's a battery protect function. So when the battery voltage gets >13.8v the circuit reduces the charge current as needed, to stop the battery voltage going over 13.8v.

In that situation the buck is not running as a MPPT but is running as a simple buck voltage regulator. There is no MPPT power gain at this point, because (obviously) the charging has been reduced, and the solar panel has excess capacity now that is not being used.

A. The MPPT action is simple, the circuit adjusts the buck PWM duty to provide the maximum current into the battery.

B. The voltage regulation is simple, if battery voltage >13.8v, the PWM duty is reduced (which reduces charge current).

If you want to make an arduino circuit to perform those two actions it should be a straightforward project, and we can help you with suggestions. But as mentioned please understand that if this is a 12v solar panel and a 12v battery, it is probably not worth doing because the gain from MPPT is small under those conditions.

If you do plan to go ahead then the design of a high efficiency buck converter capable of ?? amps is the main problem. The control algorithm is the easy part. It will help if you have experience with high current high efficiency SMPS design.

9. ### Billy4184 Thread Starter Member

Jun 20, 2014
31
0
Thanks for your comment. So if the MPPT provides all of the available current to the battery, am I to assume that most MPPT solar chargers are constant-voltage (for the bulk-charging phase)? I seem to have ran into more constant-current ones in my online searching. This is mainly why I asked the question, and as I said in the thread post, constant-voltage makes more sense to me for fluctuating power, as then the battery is surely taking all the power that it is able to, assuming charger is outputting the maximum reasonable charging voltage.

Of course, I assume that all chargers are either constant voltage or constant current (at least throughout each of the charging phases), since otherwise decent feedback would be difficult to get.

The system I'm planning is 24V solar to 12V battery.

I have made a few buck converter circuits. Lately I've spent a lot of time calculating switching losses, conduction losses and shoot-through losses for various types of buck converter. I have tentatively decided on using a high side p channel (to make switching easier) and low side n channel, since flyback diode losses are too much for such high currents. However I will have to experiment with shoot-through. My understanding so far is that the best way is to switch the high side fet a teeny bit slower than the low-side one (which is of course a trade-off with conduction losses) and try to put a thin margin of dead time around the switching interval.

Both fets will probably be driven by totem-pole BJT gate drivers to enhance the switching current which will be probably done by a TL494 PWM controller.

10. ### ErnieM AAC Fanatic!

Apr 24, 2011
7,978
1,841
Yes it was. I did not source the comment in case I misstated it, then all the blame for the error would befall on myself.

For the n-th time you say voltage when we say current.

11. ### wayneh Expert

Sep 9, 2010
16,099
6,210
Maybe you're hung up on the word "constant". The MPPT doesn't supply "constant" current, as RB has taught, its job is to supply MAX current. On a sunny day, this current level may indeed be nearly constant during bulk charging but the control scheme is not "constant current", it's MAX current to the battery.

If there's too much power available from the panel for what the battery needs at that moment, the control scheme comes off MPPT and becomes constant current or constant voltage, as required to protect the battery. At that moment it ceases to be MPPT. Any scheme that is not delivering maximum charging current to the battery is not MPPT.

Last edited: Jun 27, 2014
12. ### Lestraveled Well-Known Member

May 19, 2014
1,943
1,236
Perhaps a different viewpoint might help with understanding.

1. The MPPT electronics just monitors and controls the battery charger. There are not two power conversion devices. There is a buck mode power converter/regulator and a MPPT control device.

2. The mode of charge changes for the charging mode needed. Constant current for bulk and constant voltage for absorption and float. The buck power converter simply changes the feedback point to accomplish this.

3. If your panels are producing significantly more power than what is required by the charger, then the MPPT function has little or no effect on the operating point of your power converter. You simply are not using all the power that is available from your panels.

4. The MPPT function becomes useful when the power produced by the panels is from slightly above, to below, the power required by the charger. This region is where the MPPT does its best work. This is why:

Solar panels are great devices but they are dumb.
Buck power converters are also great devices, but they too are dumb.
These two dumb devices work great together when there is more power available than what is needed. A problem occurs when the power converter wants 20 watts and the panels are only producing 15 watts. The power converter is dumb and will increase the load on the panels in an attempt to get the 20 watts it wants. The result of this dumb tug of war is that the power converter will not even get the 15 watts that panel can produce because it will overload the panel.
OK, here is where the MPPT comes in to save the day. The MPPT is smart, it recognizes when the panels are not producing enough power to support the demands of the charger. When this occurs the MPPT tells the charger to reduce its output so that its input requirements match the maximum output of the panel.
So, for example, the charger is set at 10 amp charge current with no MPPT. The panels can not support the wattage needed for this and the charger ends up only putting out 4 amps into the battery due to the V-I miss-match. The MPPT will reduce the chargers output from 10 to 7 amps, which the panels can support and maximum efficiency is achieved.

Mark

13. ### ErnieM AAC Fanatic!

Apr 24, 2011
7,978
1,841
I suggest you re-read RB's post. No where does he say MPPT supplies a constant current. In fact, he doesn't even use the word "constant" anywhere in this thread.

14. ### Billy4184 Thread Starter Member

Jun 20, 2014
31
0
@ErnieM, the way I understood it, each phase (bulk, topping, float) is constant-something. Commonly constant current for bulk, and constant voltage for the rest.

Correct me if I'm wrong, but I thought that since the battery tends to be sluggish to respond with relevant feedback information, since there is 'stuff' going on between the two terminals that cannot immediately be seen at the terminals, it is better to maintain a constant-something (e.g. current) so that the non-constant characteristic (e.g. voltage) has time to respond and stabilize and be useful as a feedback information source.

Otherwise, if there was no delay, there would be no need for constant-anything and you could simply get the charge state by plugging momentary current, voltage and temperature values into some formula and getting a charge state result.

IF you use constant-voltage, then a drop in solar power may make it look as if the battery has had enough (if you use a series sense resistor), unless the charger somehow checks the available power and compares to it. Perhaps disengaging the battery momentarily and testing the current would be a way for the charger to do this.

Now IF you use constant-current, my problem is that unless you have a margin of error where the solar array generally gives more power than the battery requires, you will be riding a fluctuating power curve (due to changing irradiation) and it would be impossible to maintain constant current (since of course one cannot put the same current into a battery at different voltages). And then this means that power is getting wasted in that margin. Of course, some problems (such as this perhaps) are not easily fixed.

Last edited: Jun 26, 2014
15. ### Billy4184 Thread Starter Member

Jun 20, 2014
31
0
@Lestraveled thanks for the comment. Please see my reply to ErnieM above. I thought something had to be constant for feedback to work.

16. ### Lestraveled Well-Known Member

May 19, 2014
1,943
1,236
Billy
The heart of a MPPT is the buck power converter. It takes in high voltage at low current and converts it to low voltage at high current out. The input current has nothing to do with the output current. A buck converter can take 100 volts at 1 amp in and convert it to 1 volt at 100 amps out. (minus losses). In other words, the output current of the panel has got nothing to do with the output current other than calculating the wattage. Power in equals power out minus loss.

Mark

17. ### THE_RB AAC Fanatic!

Feb 11, 2008
5,430
1,311
Thanks for the info. We can get a better idea of your skill level now.

You said "such high currents"... Can you give details of the max current (solar panel total wattage will do) and solar panel voltage, and battery system voltage? Those 3 specs will help us understand the size and scope of what you are attempting.

The PFET high side driver will likely cost you more efficiency than you will save by using sync rect on the lower device. But again we need to see those 3 important specs to know more.

18. ### Billy4184 Thread Starter Member

Jun 20, 2014
31
0
Hi Mark, I understand that the extraction of maximum power from the panels (MPPT module) is separate from the battery charging (charger module).

When designing a system with an MPPT module and a battery charger module, is it not true that something in the charger module must be maintained constant (current or voltage) through any of the charging phases, and then the MPPT module must be able to provide a buffer margin of excess power, so that whatever must be constant can be constant even through small changes in weather conditions (e.g. cloud passes over sun)?

19. ### Billy4184 Thread Starter Member

Jun 20, 2014
31
0
Hi The_RB, panel voltage will be 24v, battery voltage 12v, max current 20A (edit: I only have 10A of panels at the moment if I will be putting them in series for 24V but will probably add later).

Are the p mosfet losses due to relatively high Rds(on) compared to n mosfets? I've found some reasonably low ones (~0.020 Ohms). I'd be interested to know how a high side n-mos can be switched with a positive Vgs without having two separate power sources and not adding too much complexity and further losses.

Thanks

20. ### wayneh Expert

Sep 9, 2010
16,099
6,210
The "you" I was referring to was the OP, not RB.