MPPT vs regular SMPS for solar charge controller

[Red Squirrel]

I've been reading up a bit on solar charge controllers and basically there is PWM, which is not that great as only a fraction of the power is put into the batteries so you don't get as much out of the panels . Then there's MPPT, which sounds rather complex.  What I'm wondering, is there a reason why one can't just make a SMPS circuit that takes the solar panel voltage (which varies based on load/sun etc) and create a set float voltage for the batteries?  Wouldn't that be simpler and still take full advantage of the solar panel?  Of course the batteries would only take the current they need, but the lower the battery voltage the more current they'll take, so won't they charge at a decent rate that way?  Or is MPPT simply going 1 step above that by acting more like a "smart" charger and doing a bulk charge first, then float after?

Bought a small solar panel and cheap charge controller for a project but it looks like it's PWM, so I might just build my own, or buy a better one.   Though I still need to do some testing as maybe the PWM will be fine for my application anyway.  It's basically just for an emergency/camping power pack to charge flashlights, phone, etc.

[mikerj]

An MPPT system uses a switch mode regulator, but the control scheme varies the input impedance of the regulator (by adjusting the output voltage) in order to match the output impedance of the solar panel.  When source and load impedances are equalised, you maximise power transfer.  A simple switching regulator will not do this by itself.

However, MPPT is fairly complex and usually only justified when you have to extract every last bit of power from the panels.  A regular switching regulator will still do a perfectly good job for most applications.

A tracking system to make the panels follow the sun will get you a lot more energy throughout the day than a static panel + MPPT.

[Lightages]

MPPT chargers aren't that much more expensive in the big scheme of things that I wouldn't install one in any system.

As far as trackers go, the cost and maintenance of a tracker is not worth it anymore now that solar panel prices are so low. It is cheaper and more reliable to just install another panel or two to make for the difference a tracker would make. This depends on if you have the space of curse.

[suicidaleggroll]

What I'm wondering, is there a reason why one can't just make a SMPS circuit that takes the solar panel voltage (which varies based on load/sun etc) and create a set float voltage for the batteries?  Wouldn't that be simpler and still take full advantage of the solar panel?

No, it wouldn't.

What happens when you tell your SMPS to output 14.2v with a current limit of 2a, but due to overcast conditions the solar panel is only able to provide 10w?  Your SMPS will drop its input impedance to try to increase the current draw from the source so it can provide the requested 25-30w (assuming your battery is charging, not floating), but all that's going to do is short out the solar panel and give you -ZERO- power.  The voltage will drop below the minimum required by the SMPS, the SMPS will shut off, and the entire process repeats.

Running a SMPS with a fixed voltage/current limit will only work if the panel can provide at least as much power as you're asking at all times (in other words, it only works if you're operating the panel above the maximum power voltage).  In that case, it will work, but any additional power the panel is capable of will be wasted.  In the cases where it can't provide the requested power, you'll get zero power instead.

An MPPT always runs the panel at the peak power point, whatever that peak power is.  If the panel can only output 10w, you'll get 10w, at whatever voltage and current combination are required by the battery.

An MPPT is just a SMPS with some intelligence.  Rather than fixing the output voltage and hoping for the best, it adjusts the output voltage so that the power delivered to the load never exceeds the power available from the panel, and the panel is always running at its peak potential, even if that isn't as much power as you might hope for (and design your fixed SMPS for).

[Seekonk]

You can do it and it can be dumb as a stump.   It is very easy to set the output as a fixed maximum voltage and then just OR input voltage as an override to crank down switching duty cycle.  On the small scale one of those 2596 modules can be controlled by lifting the #5 on/off pin and switching it with a transistor.   Power point is fixed to temperature.  Just adjusting the voltage four times a year for each season is close enough or you can attach a couple diodes to the back of the panel and let it track itself.  I have a system using an arduino with a very simple programm a FET, capacitor bank, and a 60hz  transformer out of an old UPS.  People are nuts not using power point.

[johansen]

give it some time and i think we'll have the mppt brains on ebay for $7, as we see happened with sine wave inverters.

http://www.ebay.com/itm/New-Pure-Sine-Wave-Inverter-Driver-Board-EGS002-EG8010-IR2110-Driver-Module-/311128100008
there is a three phase version of ^that as well and its 15$ iirc.

[Red Squirrel]

Oh so by using a straight up SMPS if you keep trying to get the voltage to be 13.5, once the solar panel voltage goes below a certain point, then it shuts off, while with MPPT you can still sorta charge the battery with a lower voltage?  Some might also do a voltage boost I'd imagine.

If I was installing a large scale system I'd definitely go MPPT though, but I was mostly just curious why it works the way it does instead of just try to hold a steady 13.5v.   Though I guess you could probably use a much higher voltage solar panel or two in series and probably get away with more using a basic SMPS topology.   As for PWM, does it actually give the battery the direct voltage from the panel?  I was testing with the cheap one I got for sale at Canadian Tire and I was getting a reading of like 20 volts at the battery.  Isn’t that very dangerous?  Especially for a sealed gel cell.  You don't want to be around one when it decides to vent.

That pure sine inverter board is pretty neat.  Pre made pure sine inverters tend to be crazy expensive.  Something I will want to experiment with later on.

Oh and what do you mean by dropping input impedance?  isin't impedance basically resistance, but for AC? How would that be controlled and how does that come into play for solar panels?

[Seekonk]

"once the solar panel voltage goes below a certain point, then it shuts off, while with MPPT you can still sorta charge the battery with a lower voltage"

Yes and no. You do turn it off and on, but 50,000 or more times a second.  Many SMPS power supplies operate this way.  PWM is rather old school.  When you operate at 150K it is not realistic to slice and dice a pulse into smaller sections.  Be assured the SMPS keeps working at low levels and you don't need to use the on/off pin. If you bother to read the 2596 data sheet it shows the low voltage shut down.  t is damn fast.  If a 339 or open collector transistor with pullup the sense pin can be faked into thinking the output voltage is too high.  the pullup feeds voltage to the input through a diode.  When transistor is saturated no current can flow and it looks like nothing is there to the sense pin.  Once collector is higher than sense pin, current flows through diode making output voltage seem higher.  A TL494 does this the same way with dual inputs, purely analog.  I haven't invented anything.  This method works quite well.  There are a dozen ways to make AMY SMPS operate at power point.

At 12V the gains in power are rather small in good light.  Lower light levels (real life conditions) the gains are greater.

[suicidaleggroll]

Oh so by using a straight up SMPS if you keep trying to get the voltage to be 13.5, once the solar panel voltage goes below a certain point, then it shuts off, while with MPPT you can still sorta charge the battery with a lower voltage?  Some might also do a voltage boost I'd imagine.

...

Though I guess you could probably use a much higher voltage solar panel or two in series and probably get away with more using a basic SMPS topology.

No, it has nothing to do with which voltage you select or what your panel voltage is, the problem is the way DC/DC converters operate.  They are all designed with the assumption that the source is CV (constant voltage).  A solar panel is only ~CV to the right of the maximum power point (Vpanel > Vmp).  To the left of the peak a solar panel is CC, and DC/DC converters do not work correctly with a CC source.

Oh and what do you mean by dropping input impedance?  isin't impedance basically resistance, but for AC? How would that be controlled and how does that come into play for solar panels?

resistance is for DC
reactance is for AC
impedance is the combination of resistance and reactance.  It can be DC or AC or both.  It's a more generic term than either resistance or reactance.

In this case, he's referring to the relationship between voltage and current (remember ohm's law) at the solar panel output (SMPS input).

Assume a SMPS is at steady state.  Steady input voltage, steady input current, steady output voltage, steady output current, Pin = Pout (assuming 100% efficiency), everything is happy.  Now what happens if the load increases slightly, even a milliwatt.  Well Pin != Pout anymore, the SMPS needs to pull more power from the source in order to deliver this extra power to the load.  How does it do that?  All SMPS do this by reducing their input impedance, which means for a CV source, you get the same input voltage, more input current, which gets you that extra power.

This does not work with a CC source.  With a CC source, lowering the input impedance gets you the same input current and LESS input voltage, which means LESS input power.  This causes the SMPS to lower its input impedance even more, getting it even LESS power, and this feedback loop continues until the panel voltage drops below the SMPS minimum input voltage and the SMPS browns out.

All of this means that a regular SMPS running off of a solar panel cannot run the load when the requested power is more than the panel can provide, and when the requested power is more than the panel can provide, you only get the requested power.  It's the worst of both worlds.  With an MPPT you always get whatever the panel can provide.  With other less sophisticated panel controllers, such as PWM, or the one that Seekonk has been describing, you might not get the full power that the panel can provide, but you'll be relatively close.  Anything is better than a SMPS with a fixed output voltage/current connected directly to the panel.

[Seekonk]

"With other less sophisticated panel controllers, such as PWM, or the one that Seekonk has been describing"

Why, I think I've been insulted.
Its like trying to teach labradoodles.

[theoldwizard1]

Check the  Sunraiden 1000 watt inverter/charge controller for $200 US !  There are very few technical details, indicator lights or controls on it.

There is one "under the hood" video that I am sure Dave would like !  He does a very poor knock off of DaveCAD and in a couple of cases his circuit analysis is just plain wrong (see the comments after the video), but this is the best info I have seen.

There is no info on battery charging so that is a crap shoot.

[Julez]

How much power does your system need to be?

For low power application, I found these little controller to be quite nifty:

http://www.ebay.de/itm/MPPT-Solar-Panel-Controller-5A-DC-DC-Step-down-CC-CV-Charging-Module-LED-Display-/191693901453?hash=item2ca1d8ce8d:g:RBAAAOSwVL1V-2bV

A good approximation to MPPT is to simply not allow the solar panel output voltage to fall below a certain point. The voltage of a solar panel does not change much dependent on solar intensity. What is changing is the ability to provide current.
So what this little controller does is to increase input current until the voltage is dragged down to the set treshold point (15-17V for normal "12V" panels), and then the voltage is kept there.
The maximum output current and voltage can be set as well, so you have a charger and "sort of" MPPT tracker in one, for less than $13.
Be advised that the output must not be conneced to voltage if there is no input voltage present. I circumnavigated the problem by soldering a schottky diode on the bottom side where they installes a simple piece of wire instead.

[djQUAN]

Nice little modules. Might try and get some to play with.

I circumnavigated the problem by soldering a zener diode on the bottom side where they installes a simple piece of wire instead.

I think you meant normal Si or Schottky diodes?

[Julez]

Schottky, sorry. I guess Si would work as well, but with slightly higher losses.

[Seekonk]

That looks cute, a slight reworking of a common power supply. If you think a moment.
it isn't very useful. It is still a LM2596 variant with a better inductor and diode.
3A is on the optimistic side as it doesn't have enough heatsinking.  Heat actually
has to conduct through the board to the heatsink.  OK for a very small system, but
the LED display may be a significant draw.  Wording implies that it has to be
connected to the battery all the time so the LED display will run all night and loss
of power may mean that stored settings are lost. Or these setting may be set by a pot.
I always have to buy one to figure out what this stuff actually does.  Just doesn't
look practical.

As to the original topic, any SMPS can be turned into a MPPT controller.  The feedback
pin of the controller chip can be faked to think the output is over voltage.  This can
be done with a pullup resistor (25K or more) feeding a diode to that pin. a transistor
or LM339 keeps it pulled low until panel voltage drops below the desired power point
voltage.  The reference voltage for the comparator can be created by a few silicon diodes
in series. These can be tucked inside the panels junction box to sense the panels temperature.
Panels after all are just a bunch of diodes and the power point voltage tracks the sum
of the stacked forward voltage.  Numerous buck converter assemblies would be suitable.
Even multiple LM2596 modules could operate in parallel for higher current.  I build simple
buck converters and control them with an Arduino.   

[Julez]

It's not that bad actually. 60W is too high though (I tested with 12V 5A), but one can reduce the current to whatever one likes, 3-4A seems reasonable. The LED display only shows parameters, the settings are made with trimmers, and are thus non-volatile. As for the current the display needs, this is only drawn from the input side. There is also a tiny slide switch which allows the display to stay off. For a dollar cheaper or so one can also get the very same module without display at all.

With the 4 trimmers, one can basically set the minimum input voltage, the maximum output voltage, the maximum output current, and the current at which an LED starts to shine, to indicate that charge is finished at a very low current for example.

The precision is within reason, and the LED display PCB has 3 potis on the back which allow to adjust in/out voltage and output current accuracy.

Heres a test video:

[station240]

The basic idea/operation of MPPT is maximum power (watts) is only available at a certain voltage/current, which is less than the maximum voltage. What this voltage is the MPPT controller has to find out itself, as it varies almost constantly due to sun angle, heat, clouds etc. The MPPT controller scans occasionally to graph what current it gets for various voltage 'loads'.

One solar panel data sheet shows it outputs 250W at max voltage, but 265W at around 90%.

[Seekonk]

Nice production quality on the video. Still, nearly impossible to tell what is happening on the display as they just flash.  Certainly an option for small systems when building something isn't an option.  It still rates an avoid.  Setting a fixed power point several times a year as seasons change is a fairly reasonable option given the simplicity.  20C change in panel temperature doesn't change the power output that much if set at an average condition.  It allows some pretty creative diversion methods.   I can pull power off the raw solar at various points and change priority just by changing the power point a tenth of a volt.  100% of potential solar power always is used.  I divert most excess power to heating water when there are not other needs.  The future of home solar will prioritized loads and scheduling.  Ironically, that technology will be driven by the new smart grid power meters that will eventually slice and dice what you pay as power costs vary during the day.   

[Antscran]

Hi all,

A late reply but this might prove of some use.  I wrote a tutorial on a MPPT project build, covering the electronics, software and some considerations regarding efficiency.  It's based on a Texas Instruments C2000, which has some advantages as it allows the ADC to be triggered by the PWM, so reducing the noise on samples.

http://coder-tronics.com/c2000-solar-mppt-tutorial-pt1/
http://coder-tronics.com/c2000-solar-mppt-tutorial-pt2/
http://coder-tronics.com/c2000-solar-mppt-tutorial-pt3/
http://coder-tronics.com/c2000-solar-mppt-tutorial-pt4/

There are some videos demonstrating the Volts, Power and PWM sampled data as well (they are pretty boring, but show the system operation)




Cheers,
Ant

[Morphin]

Hi
I have one of the 5A MPPT.

In the manual it says:

"1.Set Output Voltage: Rotate the CV potentiometer and adjust the output voltage according to the float charging voltage value of the battery to be charged or the voltage needed to light multiple series or parallel connected LEDs.(Adjust when there is NOT a load)
2.Set Turn Lamp Current: After setting the output voltage, anticlockwise rotate the CC potentiometer for about 20 cycles. Then short the output ports with a 20A multimeter. Rotate the CC potentiometer and adjust the turn lamp current to the value you need.Then rotate the potentiometer, which adjusts the turn lamp current, until the LED turns Green."

What is the meaning of "Then short the output ports with a 20A multimeter" ? I don't really understand the second step. I think it is for adjust the minimum current?

Anybody has a clue?

Thank's
Andy

[Seekonk]

You are adjusting the maximum current from a minimum setting. The meter provides a short and this causes the supply to only go to the maximum current set by the pot. You turn the pot till you get the desired current limit.  The LED is green when the current is under that setting.

[Morphin]

THank's. I think i understand more now

How do you short with a multimeter?  Is this a stupid question?

Andy

[Julez]

Just set it to current measure, put the positive wire to the positive contact of the MPPT PCB, the negative wire to the negative contact, and power it up. Now the electrons can flow freely through the DMM.

[Morphin]

Thank's a lot 

[Julez]

You're welcome.
However, you should not set it to the max 5A; mine blew up at that value. I recommend 4A max, better 3.5A.