Workings Of A MPPT Controller

[jimindenver]

No matter what, the MPPT program is only active in bulk. I believe it is so because from a strictly battery charging point of view the program is no longer needed to provide maximum power to a battery that has already come up in and is likely tapering the acceptance rate. There was something about simple power point tracking vs MPPT in the NAW&S page I posted. I can't say it has much effect on the output I see because if I see 80 amps in bulk, I can see it in absorb and float too.

 

[John61CT]

80A in Float?

 

[jimindenver]

Yes sir. Even of I crank on the microwave a Morningstar controller stays in float for 30 minutes. Actually the peak is somewhere between 80 and 85 depending on time of year, voltage of the bank and if the trucks panel is tilted or not. I expect some increase after I switch the three 250w to parallel. Individually they are good for a peak of 18.5a each and I am only seeing 50a or slightly higher in series. The truck panel is always good for 30-32 amp when tilted.

 

[John61CT]

Ah, so that's carrying loads, I read as being accepted by the batt. Is there a BM shunt in there informing the SC?

 

[bonvanroulez]

No matter what, the MPPT program is only active in bulk. 

If you’re saying what I think, that the MPPT program is NOT active in absorb stage, I don’t believe that’s always true based on an experience I had. I hope I can relate this clearly. Last winter in Arizona, early afternoon, overcast conditions, system had reached absorb stage and was at about 14.5 volts. Just goofing around I disconnected the solar panel connection to the mppt controller and tied it directly to the batteries. No loads at the time. With the panels connected directly to the batteries, the voltage measured was down to about 13.4 volts.

The batteries were still accepting enough amps that the panels could not produce at the battery voltage of 13.4 volts, especially as the panel’s power drops off sharply at voltages below Vmp.

Reconnecting the panels to the mppt controller I watched the voltage quickly climb back to 14.5 volts. I don’t recall now what the specific current readings were, or what the panel voltage was going into the mppt, but I do remember thinking; Cool, the mppt is bucking up the limited amps because of a cloudy day to get the batteries up to absorb voltage and keep them there, where the direct connection could not (such as with a pwm controller).

The manual for my controller only says that, when the battery reaches the absorb setpoint, constant voltage regulation is used. I can’t find any mention that the mppt function is bypassed in absorb stage, and the scenario above would bear that out.

I believe that little test showed me that my mppt controller has great benefit besides the early bulk stage in getting a full charge in very common cloudy winter conditions.

 

[John61CT]

Well high amps acceptance is maintained well into Absorb.

In using dino juice sources in the morning to get to say 85% full, where in the SoC curve the V transition takes place depends on both bank chemistry and the charging current rates.

So maybe MPPT is maintained until the acceptance rate drops to certain level rather than just hotting CV?

 

[frater secessus]

No matter what, the MPPT program is only active in bulk. I believe it is so because from a strictly battery charging point of view the program is no longer needed to provide maximum power to a battery that has already come up in and is likely tapering the acceptance rate.

I am not convinced, and it seems your own observations argue against it.  My observations are just like yours, by the way: the MPPT controller can kick out max power in any charging stage if enough load is added. 

Consider these questions:

1.  How does the controller find the maximum power point in Bulk so it can crank those 80A?  (we will all likely agree on this one)
2.  How does the controller find that same power point in Float or Absorption so it can crank 80A?
3.  How does the controller find any non-maximum power point in Float or Absorption?

The first answer won't be controversial;  it's something like "the controller uses a power point tracking algorithm that finds Vmp."  P&O, IC, whatever.

I am suggesting the controller does the same thing to find the power points in questions 2 and 3.  The only difference is it is looking for an arbitrary power point rather than the maximum power point.   Need 5A?  Run the panels at this power point  Need 23.2A?  Run the panels at that power point.   Need max power?  Run it at this other power point.  Occam's Razor, wat wat.

This is why I say the M in the MPPT name is causing needless confusion for off-grid folks.  The name makes better sense in the grid-tied world where the controllers run the panels at Vmp constantly.  I'm looking through open source MPPT implementations to see if I can suss out what the power point tracking algorithms are actually doing.  I'll report if I find something either way.

Explanatory note for any nearby forum owners:  I am trying to help people by heading off arguments caused by imprecise terms and poor quality technical documentation.  It's the manufacturer's fault but I think we can route around the damage here on the forum.

 

[frater secessus]

So maybe MPPT is maintained until the acceptance rate drops to certain level rather than just hotting CV?

After acheiving Vabs the MPPT controller starts moving the panels off Vmp [usually] toward Voc in order to reduce current.  We can sit and watch panel voltage increase as Absorption progresses.

This is why overpaneling can make underloaded controllers run hot. <-- I know that sounds backwards.

 

[John61CT]

So my speculation is that the reduction of Amps demand is driving that scaling down process, not the CC-CV transition per se.

Which explains both why the process kicks in some time after that transition, and scales back up even after Float when high amp loads come online.

 

[jimindenver]

So lets talk about MPPT then and why it does it. The program is what decides how to bring up the voltage of the battery the best way in bulk. PART of that program senses the battery voltage and adjust the array in relationship to it. Since the voltage of the battery is constantly rising the controller has to constantly adjust the array to match. That is tracking the Max power point. A controller that has fast tracking makes the adjustments more often.

Since the voltage is constant in absorb and float there is no reason to constantly track a new point every few seconds because you are sitting on a power point. It may not constantly be perfect but it will be close enough that you and I would never know it. What is a bit off when it matters very little and the battery is self limiting the acceptance rate anyways.

So how can I see all that power in absorb and float if the program isn't boosting it any longer? That is one of the myths of MPPT...there is no boosting going on by the controller or program. The power is always there, the buck converter makes it accessible. The answer is in the formulas Bob was using. I do not the actual numbers off of my panels but I know the roof runs 750w at 105 volts and around 10 amps. The buck converter allows it to be seen as 750w at 12 volts and 50 amps. Lower the voltage and the amps come up naturally. It does not take a program to do what happens naturally. The programs is all about control and the battery has control in absorb and float, hence the program is not needed.

So the battery has control of the voltage and acceptance rate plus the power point is constant in absorb and float. The buck converter is still seeing the same power coming down so it is still available IF a load is present that can use it. The panel did not change because of the mode the controller is in, neither does nature, the power is always there.

Taking the panel off the controller and connecting it directly did nothing more than prove why the MPPT controller is needed. When you pulled the panel off the voltage of the battery dropped. With the battery controlling the panel at a lower voltage you only saw partial output from the panel and not enough to raise the battery back up quickly. Plug the controller back in and you saw full output from the panel which was enough to raise the voltage again. It was not a program that gave you access to the rest of the power, it was the buck converter.

There is a real time way of seeing the program in action. With a Eco-worthy you have to take the cover off and test a component to see the voltage going up and down as it attempts to track. With a Morningstar controller I can connect a computer and watch the various values change in real time. The only time I saw the controller constantly trying to track was when I was doing mixed panel test and the controller seemed confuse and not able to settle on a power point. With either way of testing you should be able to see the tracking stop in absorb and float.

So now I have fried the two remaining brain cells that I had left and on a Sunday morning no less. I know that it will too vague for some and too deep for others. Like I said the answers are already here in the thread with the link to NAW&S and the formulas Bob posted.

 

[John61CT]

OK thanks I think I'm clearer.

I wonder why they don't put DCDC buck conversion in PWM?

 

[jimindenver]

I had a friend on another forum buy a buck converter and tried to make it work, it didn't. I think what he was missing was a PWM controller to send the power into. A PWM unit is in every MPPT controller, it is what sends the power to the battery after the buck converter. The difference is the PMW is receiving all of the available power at a acceptable voltage instead of limited to what the battery dictates in a normal PWM system.

So the next step would be to add the PWM controller and see if the voltage set by the battery would at least crudely control the buck converters output. Honestly there has to be some reason it will not or you would see none MPPT controllers with buck converters selling for less because the need of the electronics and program is gone even if it meant slightly less efficiency or very slow tracking.

 

[frater secessus]

This is, I promise, my last post on this matter. 

I'll distill my reply down to one issue here;  the longer (!) response I deleted will go on my blog.

So how can I see all that power in absorb and float if the program isn't boosting it any longer? That is one of the myths of MPPT...there is no boosting going on by the controller or program. The power is always there, the buck converter makes it accessible.

That might be the critical piece of bad information that is throwing everything else off.

The MPPT controller can (and does) stop unneeded power from coming into the controller by running the panels at a power point that makes less power.  This is knowable, observerable stuff.   We can see it with a multimeter or a controller display.

I just went outside and observed my system in Float. A bit of of shade on the panels but the solar was providing the ~18W it took to hold float voltage and run what little load there was.  Panel voltage was observed to be 39.2V.

Wait, why was the panel running way out at 39.2v when panel's max output is at 36.5v?  Because at 39.2v on my panels' power curve the panel puts out ~.5A, which gives us ~19w. 
Either that's a massive coincidence or the controller found the power point that would make the correct amount of power.

Then I added a load. 

It took a minute to settle down but when it did the system found equilibrium it took ~90W to hold float voltage and carry the load.  But now the panel voltage was observed to be 36.7v!

Why was it now 36.7v when a minute before it was 39.2v and max power is at 36.5?  Because at 36.7v on my panels' power curve the panel is putting out 2.5A, which gives us 91W. 
A second coincidence?

Not a coincidence; a pattern.  To paraphrase Dave Chappelle:  "that's what [M]PPT controllers do;  that's what they're for."


Summary for onlookers

Off-grid MPPT controllers manipulate solar panel output to match demand.



/crickets
seriously, that's it


Test it for yourselves

I encourage everyone to test this for themselves rather than believing anything I say.  If the summary above is correct the solar setup's behavior should be predictable and repeatable along those lines.

So either watch the display or use a multimeter at the PV inputs on your controller to check your panel voltage (Vpanel)  under various charging an load conditions.  Then check  where Vpanel lies on your panel's power curve (usually in panel specs).  Multiply Volts X Amps at that power point on the curve to get power in Watts.  Compare this to the real charging and load demands that were happening at that time in your camper.

After watching the patterns of behavior for a while, switch gears.  Now think about what the controller WILL do if you change loads or charging setpoints.  Make your prediction then make the change.  See if the prediction was correct.  Do it a few times with variations to see if it is repeatable.  For added rigor (and cheap amusement) make the prediction in front of someone and have them verify the result.  

Then relax with your dog and have a cold beverage, because you grok the the workings of an MPPT controller.


Love and respect to all, especially Jim who I admire greatly.  I now return you to your normally scheduled program.

 

[jimindenver]

I was going to mention that I have caught the Eco-Worthy controllers slacking off in bulk before. The twin 245w systems were tracking the sun yet not pushing full output into the bank even though it was not up to absorb voltage yet. I didn't see full output until the Mr Coffee pulled the banks voltage down and then both controllers went to full output and stayed there until absorb was hit.

So on its own the Eco-Worthy MPPT program will take a kinder, gentler approach to charging unless a load kicks it in the butt and makes it work for it.

I think the conversation is wonderful.

 

[tx2sturgis]

It's all 'smoke and mirrors'...

 

[Weight]

You will have to explain to me how the controller affects the panel voltage. The controller can affect the charge voltage out of the controller. That is if the panel voltage is high enough. But then, my pwm controller does the same.

 

[John61CT]

The voltage on a wire is "negotiated" between the loads and sources.

Think a low SoC batt's resistance controlling the actual voltage while charger pumping in amps striving to get to V Absorb.

 

[tx2sturgis]

You will have to explain to me how the controller affects the panel voltage.

Its very simple, lets see if the guys here get it right!

(lurking)

 

[tx2sturgis]

A certain member here PM'ed me...but I cannot send a reply due to being blocked. 

Lets not do drama, ok?

Either unblock me or don't send PM's.

Thank you.

 

[jimindenver]

Its very simple, lets see if the guys here get it right!

(lurking)

I found this in a quick search. I have no idea if they are talking out their....

The MPPT presents a variable impedance load to the solar panel to keep it operating at the optimum point.   Typically this is done with some sort of switcher to an intermediate DC bus.


http://www.eham.net/ehamforum/smf/index.php?topic=80516.0