36 Volt Solar Panel , 250 watts - Can I Connect to 12V Solar Regulator ?

[goboatingnow]

Originally Posted by LeaseOnLife

Dave

I might have the tools and parts to do a test like it. Morningstar MPPT controller and a display for it, it can show quite a few values, telling me what the controller is doing. Logging would be by pen and paper. Verifying some key data by multimeter.

But I am not sure how to define the test parameters... thinking out loud.... I can tilt the panel and change the amount of sunlight coming in. Then track Watts, Volts, Amps on the solar side, and then? Compare with datasheet?

Let me know what would be some easy to reproduce steps to confirm functionality (or quirks) of a MPPT

Dirk

The main issue I'm finding is that the controller doesn't track the MPPT point that well. Certainly taking the published PV curves. Reading some recent technical papers of Mppt algorithms , it's quite involved to do it right. What I was thinking, was trying to compare the theoretical mppt point with the actual. It's really one of those smoke and mirror products , you can't really tell if you are at the mppt point at all, just what the controller is telling you.

I was think of rigging up the Mean Well DC to DC to a PC and using it to characterise the panels into a known load. This would determine the real PV curve which could be useful to compare real mppt controllers. It's complicated by the fact that the PV curve also moves with sunlight. I'm beginning to have my suspicions.

Dave

[LeaseOnLife]

For a quick check, I re-read my own post

http://www.cruisersforum.com/forums/...tml#post918783

Output of the MPPT controller: 13.2, input voltage (solar panel): 17.5

This is close to the spec Vmpp of 17.7. I never programmed the controller, it finds this Voltage by itself, and is not too far off. (I don't remember time of day or clouds...). I am not too suspicious

I will do a few trials around midday, currently in Miami, should be enough sun to get the spec panel wattage. But this will only give a few datapoints, not really good statistics, and only for one brand MPPT controller....

Dirk

[goboatingnow]

Originally Posted by LeaseOnLife

For a quick check, I re-read my own post

http://www.cruisersforum.com/forums/...tml#post918783

This is close to the spec Vmpp of 17.7. I never programmed the controller, it finds this Voltage by itself, and is not too far off. (I don't remember time of day or clouds...). I am not too suspicious

I will do a few trials around midday, currently in Miami, should be enough sun to get the spec panel wattage. But this will only give a few datapoints, not really good statistics, and only for one brand MPPT controller....

Dirk

I beleive the most common algorithm is "perturbate and measure " ( !!) . The controller moves the load and then measures if the watts delivered improves. It keeps going until it finds no improvement. I've found some evidence to suggest that some arrays have a series of bumps in the PV curve and the controller is stopping at these, but I've not enough data to really make a useful comment. Then there is the mess of mppt when you have multiple panel, I'm convinced that's a factor I often think you really need a controller per panel.

Dave

[LeaseOnLife]

Pertubate and measure: This is what I want to see, I want to see the Voltage go up with more sunlight, down with less sunlight, back up with more sunlight. If the MPPT does this, at least this simple test is passed.

I agree, a MPPT per panel is probably best, I went this way for 3 of our panels, the fourth panel still has the cheap batt-charge-contoller (will be changed soon). But I do hesitate to recommend this solution, one can buy a lot of extra solarpanel area/watts for the cost of the MPPTs.

Originally Posted by LeaseOnLife

.....
Disclaimer: I don't think either solution (parallel or in series) is good. A controller per panel make sense if the panels see different light.

....

Dirk

[goboatingnow]

Originally Posted by LeaseOnLife

Pertubate and measure: This is what I want to see, I want to see the Voltage go up with more sunlight, down with less sunlight, back up with more sunlight. If the MPPT does this, at least this simple test is passed.

I agree, a MPPT per panel is probably best, I went this way for 3 of our panels, the fourth panel still has the cheap batt-charge-contoller (will be changed soon). But I do hesitate to recommend this solution, one can buy a lot of extra solarpanel area/watts for the cost of the MPPTs.

Dirk

Yes but the controller shouldnt let the voltage go above the Vmpp either.

Dave

[LeaseOnLife]

Originally Posted by goboatingnow

Yes but the controller shouldnt let the voltage go above the Vmpp either.

Dave

Sure, but it should "know", because the current drops off (rather rapidly) and the MPPT should tune back to a lower voltage. Thinking out loud again, it really has to constantly "wiggle" on the voltage, to find out if there is a more optimum point above or below the current setpoint. I will play with it, I guess we have a few parameters I can check on, with a simple test.

Dirk

[hellosailor]

Dave-
From what I've gathered, there are three MPPT algorithms out there, plus whatever tweaking each vendor has done on the one they've chosen. Each apparently has advantages under different conditions, i.e. cloud shading versus full and slowly changing sunlight.

I suspect the only way to really compare them would be to set up multiple identical panels (same vendor, same lot) each on a different MPPT controller, each to one battery (also identical lots), and then see not only which one passed the most post on to the battery, but which one recharged a battery fastest, or most thoroughly. (i.e. to 95% fast, or 100% a bit slower?)

Anything else would mean constantly analyzing panel outputs, switching controllers, analyzing outputs from them, and simply switching controllers might interrupt the logic each was using. To do the comparison really well, let alone accurately, will probably require some formal funding. But then again, that's the kind of test that might qualify for grant money or funding, to see which MPPT scheme actually IS best for what conditions.

[goboatingnow]

I was wondering what would be the difference between a ordinary DCDC converter and an mppt one. It's easy to knock up simple switchers these days.

Dave

[hellosailor]

Dave-
"the difference between a ordinary DCDC converter and an mppt one. "
You could contact Linear Technology and ask an engineer. They have sample kits and premade MPPT sets available "to the trade".

From what I've been able to glean (because no one in the business wants to really share their technology, can't blame them) an MPPT controller is basically two DC-to-DC converters being controlled by one microprocessor or varying ability.

The "input" side is allowing the solar panel to charge an inductor (a coil or coil and capacitor) to some state that in theory matches the impedance at which the solar panel produces maximum power output. There's nothing "active" being done, except that whenever the impedance of that central toroid goes above or below the MPP of the solar panel, THEN the microprocessor pulses power from the toroid to the battery. And in this case timing is everything.

It is very much like having a canal with one lock in it. You open the input side, water gushes in. When you decide that's enough, you adjust the opening at the input side, maybe close it, maybe not, and at the same time open the exit gate. Again, maybe all the way maybe not.

The trick is to keep the impedance of that central "charge accumulator" matching what the solar panel wants to see, while simultaneously sucking out as much power as you can, in pulses of the right height, duration, and frequency, to charge the battery.

So now you're dealing with multidimensional algorithms on both sides. How to tweak that impedance, how often to tweak it, how much time to lose in retweaking it, or just let the microprocessor do some sleeping from time to time to get better efficiency. And then having it fiddling at least three variables on how it sends pulses to the battery--based on a logic and lookup table which has been loaded with battery capacity and optimum voltage, etc. by the user, and the charging algorithm that the company selects.

A LOT of little balls in the air, all at once, for the microprocessor. Of course that's all donkey work and even a tiny donkey can juggle those many balls all at the same time, if you've got a good program for it to run.

It may not always be happening that way, but I have reason to think that's the way it is being done, again with variations by each vendor.

DC-to-DC into a pool, DC-to-DC out of the same pool. And the rest is just a simple job of monitoring and controlling all the variables on both sides.

Compared to a modern engine control computer in a car...childs' play.

Figuring out which man behind which curtain is the real wizard among the bunch, ah, that's a bit harder to guess.

[noelex 77]

That's a good description of how the better MPPT regulators work, but there is a lot of cheap stuff out there.It is difficult to know exactly how each system operates, but the cheapest way seems to be to periodically measure the open cell voltage of the solar panel and then deduct a set voltage say 3v and assume this is the Vmpp. The panels are then run at this voltage. As the Voc rises and falls it is assumed the Vmpp rises and falls by the same amount.
Not very efficient. if you combine this with only charging the batteries at only one voltage (no separate boost and float voltages) you have a very inefficient regulator, but if it says MPPT on the box, people think its good

[goboatingnow]

Hmm I don't think theres any need for two converters.

Looking at the LT3562, that really what I'm doing with the mean well DCDC unit. Setting a fixed Vmpp and regulating to that. However the LT3562 is in reality not a tracking mppt unit.

As to the need for two converters no its not needed. All a tracking mppt is is a simple switch mode regulator, with input sensing , what the switcher does, by varying the energy conversion,( by PWM) in effect changes the load the regulator places on the solar panel. Vin,Iin are measured, then the controller will slightly change the panel load and see if the watts in improves, the algorithm then chases in the direction of increasing the watts in, until no further improvement is found.

There no storage "pool" as that's not really feasible in electronics. ( other then what happens in the inductor in a switch mode setup)

Theres no mystery about the mppt algorithms, they have been written about extensively, nor , with a micro, is it difficult to do. ( ps I'm a design engineer, so in the trade) . What I was wondering is how good the existing mppt implementations are. I also believe that the current costs are very high, but I suspect that will change when more integrated solutions exist.

It's a nice implementation the LT chip, and in my experience LT is a good IC house, might knock up a little module and play with it ( especially the parallel output version) its not a tracking unit as you have to program the Vmpp, hence you cant just throw any panel at it, but still nice enough. It's designed for applications where the panel is known in advance, like parking meters etc.

The NXP Mpt612 IC from NXP ( ex Philips) is a true mppt controller , containing "patented mppt software". More complex to rig up. cypress semi also has a SOC ( system on a chip ) similar solution.

silicon Chip magazine http://www.siliconchip.com.au/cms/A_112335/article.html has a nice article on mppt.

And TI have a good module description at http://www.ti.com/ww/en/analog/solar...ications.shtml

Dave

[Chkm8]

Hi Guys, I would like to run 3 - 240 watt, 39 volt panels on the arc to a 24 volt Battery bank .. controller suggestions 1 controller or 3 and then how to tie into battery bank ... Thanks for thoughts .. Paul

[LeaseOnLife]

Originally Posted by LeaseOnLife

Sure, but it should "know", because the current drops off (rather rapidly) and the MPPT should tune back to a lower voltage. Thinking out loud again, it really has to constantly "wiggle" on the voltage, to find out if there is a more optimum point above or below the current setpoint. I will play with it, I guess we have a few parameters I can check on, with a simple test.

Dirk

To follow up on the simple test, I learned a bit by just observing the voltage of the solarpanel: This MPPT controller sweeps the Voltage every 183 sec, from just above battery voltage to max voltage of the solarpanel. Each sweep is rather short, about a second. Then the MPPT controller controls to a specific voltage for the next three minutes, sweeps again, controls the panel to a new voltage.

I can't say it will find the absolute optimum of the panel, but the Vmpp seems to hover around the manufacturers spec, for a warm panel, midday sun in Miami. I am happy with the results. A simple controller, as Dave proposes, with a fixed Vmpp, will probably be cheaper and very close to the performance of a good MPPT.

Dirk

[noelex 77]

Originally Posted by LeaseOnLife

To follow up on the simple test, I learned a bit by just observing the voltage of the solarpanel: This MPPT controller sweeps the Voltage every 183 sec, from just above battery voltage to max voltage of the solarpanel. Each sweep is rather short, about a second. Then the MPPT controller controls to a specific voltage for the next three minutes, sweeps again, controls the panel to a new voltage.

I can't say it will find the absolute optimum of the panel, but the Vmpp seems to hover around the manufacturers spec, for a warm panel, midday sun in Miami. I am happy with the results. A simple controller, as Dave proposes, with a fixed Vmpp, will probably be cheaper and very close to the performance of a good MPPT.

Dirk

This sort of system is common. It works well for house type systems, but there are some problems on a boat where shadows are encounted and can fluctuate rapidly as the boat swings or rocks. A shadow will reduce the voltage contribution of the whole string of solar cells that share a common bypass diode so even a small shadow will reduce the Vmpp considerably.

[deckofficer]

Originally Posted by Chkm8

Hi Guys, I would like to run 3 - 240 watt, 39 volt panels on the arc to a 24 volt Battery bank .. controller suggestions 1 controller or 3 and then how to tie into battery bank ... Thanks for thoughts .. Paul

From the posts, there are a lot of folks with more knowledge on this than me, so wait for their reply. Me, I'd wire the (3) panels in parallel with (1) quality controller with smarts.