MPPT vs PWM test

[jimindenver]

Today Bob and I set up two 200 watt portables, one with a ECO-worthy 20a MPPT ($97 on their site, $85 on ebay), the other with a Solar30 PWM controller ($25). Both have displays so no remote meters or fancy dongles needed. While it wasn't the extended time lapse test I had hoped for, it did prove two concepts.

First we both were baffled when the MPPT system was not able to beat the PWM system. What had happened was the four panels were not all the same and the two panels in series we had hooked up to the MPPT controller were not a good match. That not only weakened the stronger panel, as I have said before the mismatched panels voltages kept the controller from doing its job properly. Once we swapped the panels around to give the controller a matched set it started doing what was expected. The parallel panels on the PWM system have no effect on each other and anything above battery voltage is shed anyways. The PWM controller showed the same amps output as it did before.

Now for the good part. The MPPT system easily out did the PWM system by 3 amps when the battery was in the low 12 volt range (11.5a vs 8.5a) and by 2 amps at 13 volts (11.5a vs 9.2a.) That's 3 amps a hour on 200 watts when your battery needs power the most. 2 amps when your battery is not near to being full. Those extra amps add up over a days time, just imaging if we were talking 400 or 600 watts. Not only is a considerable amount of power, PMW is keeping your battery from being charged.

 

[Motrukdriver]

I wonder if that is because of the specific brands of controllers. I'd be interested in seeing a side by side comparison of two Morningstar controllers like the TS-45 and the TS-MPPT-45... but who has the money to shell out for that kind of test. That's nearly $600 for both of those controllers.

 

[frater secessus]

In normal systems there is only one MPP; in mismatched systems there can be more than one.  Some MPPT controllers are programmed to look for multiples.  The downside of doing this is the system could be wasting time (and therefore harvest) sniffing around for other peaks.

AltE has some neat videos about the effects of using mismatched panels.

 

[John61CT]

Thanks for doing that, excellent test!

I bet the difference would be greater using a higher quality / newer tech MPPT.

And of course less with a better PWM unit.

In some unusual conditions a great PWM (like Bogart's run via their Trimetric) will actually put out more than an MPPT, but that was a few years ago.

For me the best aspect of MPPT SCs is greater flexibility in choosing panels to maximize the wattage you can fit on a little roof.

And newer SCs giving lots of charge profile customization, to help your specific bank last longer, making sure they get to 100% Full via endAmps.

 

[B and C]

First we both were baffled when the MPPT system was not able to beat the PWM system. What had happened was the four panels were not all the same and the two panels in series we had hooked up to the MPPT controller were not a good match. That not only weakened the stronger panel, as I have said before the mismatched panels voltages kept the controller from doing its job properly. Once we swapped the panels around to give the controller a matched set it started doing what was expected. The parallel panels on the PWM system have no effect on each other and anything above battery voltage is shed anyways. The PWM controller showed the same amps output as it did before.

A friend of mine is going to put on more panels but he can't get anymore of the panels he originally bought.  I am curious about the mismatched panels.  Not same voltage, watts?  

I guess I am asking what makes a good match?

 

[Weight]

I have a Morningstar TS-45. PWM. My batteries are charged. I think we need a more controlled and scientific test.

 

[tx2sturgis]

So, my conclusion from the test is that MPPT controllers are finicky about the panels being used, and the PWM controller is more much flexible with mismatched panels...

PWM 1

MPPT 0

Another win for P-W-M!



(just a little humor that will probably be deleted for not being helpful)....

 

[GotSmart]

Hmmm

I had PWM and 200w of panels, and 448 ah of batteries. 

My batteries were always fully charged. I never drained anything even though I had a CPAP , roadpro oven, lights and computer. 

By careful use anything works better than nothing. 

MPPT is obviously more efficient, but why pay more if you have no need?

Brian, your friend should set up the new panels to their own controller.

 

[Motrukdriver]

I would also add, is the price difference between MPPT and PWM worth the advantages? Does one have better battery health management over the other. If you could get longer life out of your batteries because of better charging/management I could see paying more for a MPPT controller. Batteries ain't cheap.

 

[tx2sturgis]

I bet the difference would be greater using a higher quality / newer tech MPPT.

And of course less with a better PWM unit.

Good points...

I would also like to see a 'control' in the test: ie., a straight connection from a set of those panels to a battery, and/or thru a shunt controller, to get a baseline for input from the panels with no controller in the way.

Notice I didn't say leave it there with no controller and unattended for hours...it could be monitored manually (and disconnected when acceptance of charge current is dropping) so that the batteries are kept healthy.

 

[DLTooley]

The latest Victron 75/15 has Bluetooth built in. $120 at pkys. I think the Victron has better charge state mgmt, not sure how you would test that under varying environmental and use categories. Another question I have is the benefit of higher voltage panels/arrays in low light conditions. 

Thanks!

 

[user 423]

Are PWM controllers any more reliable? Are controllers in general free of failures?

 

[Weight]

Rod Collins did a controlled careful study between MPPT and PWM controllers. He was using liFePo batteries so the results are not directly related to lead acid. He found a better than 20% difference. https://marinehowto.com/mppt-vs-pwm-solar-controllers/

 

[wayne49]

Good points...

I would also like to see a 'control' in the test: ie., a straight connection from a set of those panels to a battery, and/or thru a shunt controller, to get a baseline for input from the panels with no controller in the way.

Notice I didn't say leave it there with no controller and unattended for hours...it could be monitored manually (and disconnected when acceptance of charge current is dropping) so that the batteries are kept healthy.

Please do this with your panels and battery and report back.

Renogy 100 watt panels put out 20.1 VDC.

 

[tx2sturgis]

Please do this with your panels and battery and report back.

Renogy 100 watt panels put out 20.1 VDC.

lol....that's open circuit voltage...but when loaded down, pulling current, all panels show a drop in voltage...this is exactly how MPPT controllers vary the voltage of a panel...by loading it down. And PWM controllers apply the full panel voltage in short (or long) pulses. 

One of the small panels on my old pickup truck maintained my starting battery for years with no controller, only a blocking diode.

 

[wayne49]

I wonder how the panels would like battery current flowing in.

Please have separate videos of the battery and of the panels during this smoke test.

I have measured 20.1V per panel when hooked up to a controller hooked up to a battery bank. Not an open circuit.

 

[jonyjoe303]

years ago I ran a similar test with the same ecoworthy 20 amp mppt controller and a wincong pwm controller but I used seperate panels a 240 watt 36 volt mono and a 120 watt 21 volt mono. Panels were connected separately no series/parallel. Panels lying flat on roof on my van.

120 watt panel kyocera 21 volt (about the max I got out of panels with lead acid)
mppt = 6 amps
pwm = 6 amps

240 watt panel sharp 36 volt 
mppt = 12 amps
pwm = 6 amps

But more recently I been using lithium with the 240 watt panel/ecoworthy mppt and I got following results
11.1 volt li-ion battery 
mppt = 15 amps
pwm = 6 amps

12.8 volt lifepo4
mppt = 12 amps
pwm not tested but I figure it would max out at 6 amps.

With lifepo4 I have to adjust the bulk voltage on controller to 15.4 volts to get 12 amps, at the recommended 14.6 volts the controller will only put out less then 6 amps.

 

[Trebor English]

I would also like to see a 'control' in the test: ie., a straight connection  . . .

Notice I didn't say leave it there with no controller . . .

Maybe the March 1984 Mother Earth News controller would do what you want.  It is as simple as they get.  It always charges through a diode and a 5 ohm resistor.  When the solar panel makes power it turns on a relay shorting the resistor.  When the battery gets to the set voltage the relay goes off.  

https://www.motherearthnews.com/diy/solar-charge-controller-zmaz84mazale  

There is a schematic, printed circuit layout, all you need to make one.  As an experimental control it ought to be good.  It has the obvious deficiency of all that relay coil current being wasted but that can be measured and accounted for.  This is not your typical pulse width modulated (PWM) controller but rather a straight on / off controller.  Removing the 5 ohm resistor might make it a better experimental control.

 

[B and C]

The really small one that I got at Harbor Freight to keep the Geo battery "maintained" is only 7 watts and I don't use a controller. It rains too much here to measure voltage but it starts right up.

Rain, rain, go away, come again another day.....

 

[tx2sturgis]

I wonder how the panels would like battery current flowing in.

Please have separate videos of the battery and of the panels during this smoke test.

I have measured 20.1V per panel when hooked up to a controller hooked up to a battery bank. Not an open circuit.

The blocking diode I mentioned, and often incorporated in solar controllers, blocks reverse current at night, or anytime the panel output voltage drops too low, but is not normally needed during daylight. 

Of course there is a small voltage drop across the diode. And you will not see 20.1 across that panel if you measure it with a seperate meter when the batteries are at a lower state of charge and being actively loaded down by the controller. 

BTW, shunt controllers apply full panel voltage and current up to the point that the shunt is applied, then they short the panel output completely. Simple: ON (full panel voltage to the battery) or SHORTED (panel voltage clamped).

If you read my statement again, I plainly stated it should be a monitored and manually controlled test, just to read the current with no controller in the way, not an unattended experiment over days or weeks. 

I knew it would happen that someone would have a negative response, that's why I put in the disclaimer. 

Thanks for your input tho...