define MPPT "as" marketed?

[bob g]

at the risk of stirring up all sorts of vigorous argument, and while i understand what i understand mppt to actually be, i am somewhat at a loss as to the difference in the theoretical definition and what the actual as deployed definition might be.

while i understand the theory behind mppt or maximum power point tracking, i am left to wonder if the definition is so broad , that maybe it turns out that it isn't exactly mppt in a broad application sense but maybe at best a booster within very narrow parameters as built and sold by the industry.

allow me to explain what i think i know, and then i will follow with a question.

theoretically mppt ought to take a higher voltage source and via a buck converter and some programmable control reduce the voltage while increasing the amperage with the result being more power available to charge the battery bank.

this higher voltage can be in some cases as high as ~150volts dc from the source and the battery bank could be anything less, typically 12,24, 48volt nominal delivered to the batteries.

while the products sold by the big companies will operate within these parameters, it appears that they are optimized for use with normal nominal panel voltages, so that maximum efficiency is seen when using a 12 volt panel with perhaps an open circuit voltage of 17volts. (24 and 48 at their correspondingly higher open circuit voltages) or in some cases they reach best efficiency at 24volt nominal or 48, or whatever,, the point being often they are not as efficient if fed 150vdc
as they would be if fed the 17vdc (for a 12volt system)

i am using 12volt system as an example only, not a suggestion.

what i am wondering is this

do the manufactures engineer and tailor their products to provide maximum efficiency for what they perceive as being systems most in demand?  that would make sense from an economic standpoint but when applied to systems made for higher voltage inputs why would we tolerate lower efficiencies for the broader width (between open circuit and battery nominal voltages?)

i thought at first it was a product of the buck converter section, but this is clearly not the problem, a converter can be made as high in efficiency at wider width as it can be for narrower width (between supply and load voltages), so... what is the culprit?

is it within the programming, wherein the effort to maximize efficiency for a narrow parameter for a specific nominal voltage at common supply voltages results in a compromise(s) the result of which mandates lower efficiency at other supply and load voltages?

hope i am making myself clear

if this is the case, (which i firmly now believe) , then there is a real good argument to be made for either of two things to happen

1. the oem ought to allow entrance to the programming of their mppt controller so that the user or installer can adjust the programming to optimize efficiency for other voltage battery banks and other input voltages.... or

2. the drive to design and build a controller that is open source in the programming side of things is certainly high in my opinion. as difficult as it might well be.

yes building high power converters is a tough prospect, so that is a significant hurdle, once crossed the cost to produce would be reasonable relatively speaking.

the programming side of the thing, while complex is probably the most doable in that programming generally does not eat up parts, just lots of man hours, of which there are lots more programmers than there are power electronics guys out there. probably 10k programmers for every qualified power electronics engineer.

having said that it would appear to me that one (or group) ought to clearly define the operation of an mppt, specifically what parameter is it to operate under. then lay out some basic code architecture, enough to get an idea of what needs to be monitored and what needs to be controlled and how best to do each both separately and together. 
then start work on the power electronics end of things.

i am thinking that for instance it might be possible to do the following

start with a switch mode power supply, such as a heavyduty puter supply some of which like 90-270volt ac input,  and have outputs in the 30 or more amps at 12volts dc...

hack into the control circuitry of the converter and instead of a flat dead on regulator, insert a microprocessor wherein one could take various inputs such as input voltage, temperature, windspeed (if for wind) and whatever else to crunch he numbers and vary the circuit to
make the output voltage anything one wants to charge a 12volt battery.

puter supplies are mass produced, most are about bullet proof to start with, if bought in quantity are relatively inexpensive so the converter section is  about 90% there...

which leave the control side of things, the microcontroller and its coding. i can do the bs2 stamp but there are lots of folks that can do much better with a pic chip for a fraction of the cost of a bs2.. and there are lots of folks that could program one to do just about anything we can imagine...

so why not?

but first we have to define the problem (what specifically is mppt)?, under what parameters will it be asked to operate (input and output voltages)? and can it be made so that it can be optimized for specific
installations with the goal being the ability to take full advantage of the converter.

so there you are, any thoughts guys?

bob g

[boB]

Bob, I would have to say that MPPT does not necessarily mean it must take a higher voltage and buck it down to a lower voltage.

In fact, some grid tie inverters take a lower voltage and make AC voltage out of it which of course varies from zero
volts to a couple hundred volts.

MPPT, in my mind just means that the "tracker" finds or runs the electric power source at its maximum power point voltage,
which could be any voltage.  How it converts that input to its output can be done in different ways as you very well know.
The usual way of using a buck converter is because it just makes more sense to reduce the voltage because, for battery charging,
the output voltage is normally fairly low and the installation will usually benefit from having a higher voltage at lower current from a power source
that could be hundreds of feet away from the batteries.  This reduces the cost and size of the wire that has to run some distance away.

Also, by boosting instead of bucking the input to output voltage, that voltage range will be limited because the MPP voltage can soon bump up against the battery voltage and that's as high as it gets.  Bucking it down will give a wider range of input voltage and lower input current.

Buck/Boost can do even a wider range but then, the converter gets more complicated and possibly less efficient than a straight buck converter.
It usually doesn't make a lot of sense to bother with that when it's so easy just to wire things for a higher input voltage for battery charging.

Yes, the hardware efficiency usually gets worse as the input voltage goes up but because the voltage is  normally higher to reduce wire losses and shading issues, the gain made up in efficiency of the system as a whole will normally more than make up for that converter efficiency loss.

my 2 cents on those aspects.

boB

[bob g]

thanks boB

that clears up my thinking on the subject

the reason i asked was based on the variety of posts here and elsewhere that have mppt defined as being all sorts of things, some of which don't make a lot of sense to me.

at least i know now that the track i am on is in the right general direction

thanks again
bob g

[birdhouse]

is there a standard to be able to call your CC a mppt version? 

i recently did an off grid solar install with three 245W panels (30.7Vmp) series wired into a tristar mppt-45.  24v battery.  the tilt angle was slightly above latitude for better snow shedding/winter charging, but i was a little dissapointed with the results.  granted it's been hot temps, and imperfect angle, but the max he's seen from the panels is 600W!

i've got a rogue power 30A mppt on two series 12v evergreen 195W that i've been very pleased with.  24v battery.  the display comes with the unit, and it blows the pants off the MS display.  lets you really know what the controller is doing.  @$325.00 it's a deal...  though it has a Oc 50v max!  if the guy who makes those could up the OC voltage to say ~100v or even 175v, with the same price point, he'd be on top of the hill! 

my two series evergreens many times run at 34-38V with that rogue controller.   it tells you on the display exactly the amps per non mppt-d versus what it's putting into the battery and it's amps.  far better than the MS display, which basically doesn't tell you much of what's really going on, unless you chuck a laptop to it....

just my thoughts based on experience with two different mppt controllers

adam

[bob g]

unless you have verified the screen reported numbers via some external metering, i would not necessarily trust the numbers...

reason being, a well known mppt controller we tested a couple years ago, displayed volts and amp input, along with volts and amps output
three of which were fairly accurate, one was way off!  turns out the amps output was calculated rather than actually measured!

calculated?  really?  wtf!

if you display a calculated number for output amps you can make a controller look as good as you want it to be.

to be fair, we were using the unit on a 12 volt battery and feeding it with less than 250watts of panels, it might well have been more accurate on higher voltage batteries and with more input panel wattage,,, however it doesn't change the fact that rather than actually measure the amperage they chose to calculate it and report that number on the screen..

most of us would expect a number on a screen to be reasonably accurate and the product of a measurement.

point being, before you condemn a unit, or get all excited about one, it ought to be cross checked to see if the numbers reported on the screen have any correlation to reality.

bob g

bob g

[boB]

is there a standard to be able to call your CC a mppt version? 

i recently did an off grid solar install with three 245W panels (30.7Vmp) series wired into a tristar mppt-45.  24v battery.  the tilt angle was slightly above latitude for better snow shedding/winter charging, but i was a little dissapointed with the results.  granted it's been hot temps, and imperfect angle, but the max he's seen from the panels is 600W!

i've got a rogue power 30A mppt on two series 12v evergreen 195W that i've been very pleased with.  24v battery.  the display comes with the unit, and it blows the pants off the MS display.  lets you really know what the controller is doing.  @$325.00 it's a deal...  though it has a Oc 50v max!  if the guy who makes those could up the OC voltage to say ~100v or even 175v, with the same price point, he'd be on top of the hill! 

my two series evergreens many times run at 34-38V with that rogue controller.   it tells you on the display exactly the amps per non mppt-d versus what it's putting into the battery and it's amps.  far better than the MS display, which basically doesn't tell you much of what's really going on, unless you chuck a laptop to it....

just my thoughts based on experience with two different mppt controllers

adam

 600 watts from a 735 watt PV array is about 81% of the nameplate rating.  That is not far off of what you will usually get.
You hardly ever get the actual nameplate rating.
You won't get 600 watts on a 12V battery with a 30 amp controller but I bet that Marc is working on that as well as the 50V limit.
He should probably get that controller UL or ETL listed though at some time if he wants to sell them into the larger US market.

What was the Voc and what was the MPP voltage it was running at ??  If it was somewhere around 75% to 82% or so of Voc,
then it can't do much better than that  (typically, unshaded).

Bob G...
 Yes, you cannot always trust the metering on a controller.  When comparing, one should probably use a 3rd
party meter.  But even then, with  good metesr, measuring efficiency is very hard to do accurately.

The MX60/FM60/FM80 you tested does not have an input current sensor and so it has to calculate (guess)
at that current based on input and output voltage and output current measurement.  The reason that doesn't work
great is pretty much because of the efficiency loss of the CC which is not taken into account on that controller.
(unless they changed that recently)


boB

[Mary B]

The tristar doesn't even need a meter, mine is talking to my computer giving me any data I want to see and a bunch I don't need!

[bob g]

boB

you be one sharp dude, and you certainly don't miss much do ya?

i didn't want to use a product name in this discussion, but i do remember some time ago having a similar discussion on this forum which you took part in...

your memory is very good my friend!

bob g

[boB]

your memory is very good my friend!

bob g

Thank you.

I have been taking my Ginkgo Biloba !

boB

[thirteen]

So can one mppt controler be used to regulate power from two different sources. One would be 725 feet away and will be a micro hydro system using high voltage and the other will be a split solar system one will be 25 ft away and 275 feet away.  24 v  system
Or would I need to get another mppt for each system. and one for the solar panels 275 away?

[birdhouse]

boB-
sorry for the super delayed response!  so have you had coresponace/relations with marc with rogue?  i think his controller is top notch!  i've been running it for almost a year now!  aside from the obvious gains from mppt, marc's/rogue controller is far better than the ts-45/60mppt from morning star in the display, and the fact that everything is included in the checkout price. 

like i said before, if he could raise his VOC to 120, or oven 175, he would have the market!  especially if UL listed. 

a ts-45 mppt with a junky (IMHO) display runs about $500...  if marc could get the VOC up, he could easily sell his controllers for $400 a pop.   his controller already stores 30 days of info without even chucking a computer to it, and has the best/most informational display out there!

adam

[clockmanFRA]

bob g, Here in France i installed a Tristar MPPT for my Solar Trackers, the main reason was getting the panel voltage as high as possible, and saving the installation of thick cables becuase of my 200 meters cable run to my 48v 1000ah battery bank.

However, Your Topic  "define MPPT "as" marketed?" for me is very appropriate as in my observations and hereing from others nearby, 'The MPPT has gone to Market to early'.

Several Tristar Owners have reported seroius issues with thier controllers, from blowing up to regestering permenant faults.
Mine tells me every fault possible on its display, even though i am well within the operating parameters.

I also have 3.7m diameter Hugh Piggot design Wind turbines and I will never/ not install a Tristar MPPT controller on them. I note that for a brief 1/2 second, my MPPT controller just drops the input voltage when it likes, okay for Solar no issue, but for a Wind Turbine it would be scary for loading to be stopped started. 

So for me Tristar MPPT went to market to Early, so i await the magic blue smoke. 

[Yyrkoon]

A few years ago I read a research paper from some young woman. The topic of course was of MPPT devices. As I recall the gist of the whole 30+ page document was that MPPT can be up to 25% more efficient. Granted, there was also a temperature delta that brought that efficiency down greatly. Once passed a certain temperature ( I am remembering 70F , but could be wrong ), the efficiency gain dropped sharply. So what I seemed to take away from the paper was that it was only good during winter months.

Now, from the algorithm side of things. The most common software approach for MPPT is the perturb and observe method. Also known as a hill climber method.  What this means is that every test cycle, the algorithm takes voltage and amperage samples from the output stage.  Then calculates the output power, steps up/down the voltage. Then recalculates again. This is done constantly, forever in search of a better "power curve". Which means that half the time, the charge controller could be working at a less than best efficiency. That is, possibly. It really all boils down to how a given company implements the software.

In my mind, the ideal MPPT controller should have isolated input/output stages. Input/output voltage/amperage sensing. Output should be both peak voltage, and amperage controllable. Then this MPPT should have an MCU, and software capable of recalculating / applying power changes between PWM time slices. Also, the design should use buck/boost topology. For best possible efficiency, This is not impossible, but can be expensive to implement. Which of course is going to be passed onto consumer. However, this also explains why the perturb and observe method is used most often. Changing software algorithms is easier and cheaper to implement.

So, to answer the original question. Yes, and no. The concept is solid, but sometimes( almost always ? ) poorly implemented. If not, then the cost is going to be too much. Also, the power lost through using a regular PWM versus a MPPT design . . . could be made up by adding another PV panel to your array. probably for less than the cost of a good MPPT controller.

Do keep in mind that I am only a hobbyist embedded systems designer. I have no degree in electronics engineering, or power electronics design. But do have a strong programming background. So I do understand the software side of things fairly well. Also, having lived completely off-grid now for 4 years. I have a fairly good idea of what should work well in such a system. With that said, I do not know everything . . .

[Yyrkoon]

1. the oem ought to allow entrance to the programming of their mppt controller so that the user or installer can adjust the programming to optimize efficiency for other voltage battery banks and other input voltages.... or

2. the drive to design and build a controller that is open source in the programming side of things is certainly high in my opinion. as difficult as it might well be.

#1 will never happen. Too many issues to keep up with. Least of which is maintaining IP on their source code. Another big one would be warranty issues. Now if you mean being able to adjust certain parameters, then sure. That could work.#2 would be very cool, and is possible. With creative commons licencing, companies can dictate how the hardware, and software can be used. Single/ personal use, etc.

start with a switch mode power supply, such as a heavyduty puter supply some of which like 90-270volt ac input,  and have outputs in the 30 or more amps at 12volts dc...

This is impractical. Not only does a computer PSU have a minimal load requirement on the 5v regulation circuit. But the components most likely will not match those needed for a wide input voltage to single(adjustable) output stage. Then there is PFC circuitry, etc that is not needed in a DC/DC converter. Again, I am not an electronic engineer. But I do have a friend who has been an EE for 30+ years. That and I have gone over this with him several times in the past. . . I wish it were that simple too.

I think the real challenge would be to take a wide range of DC input, and output a single ( again adjustable ) output. So say, if you have a panel that outputs 30V @ 1A, but has a max series voltage of 300 volts. You might want to ideally have a wide input value of 30V-300V. Then on the output stage, you could have 12,24 or 48 nominal output regulation. *This* could work, although with current technology I am not sure that wide of an input range could be made to happen.

Power electronics is indeed complicated. From a software / MCU selection perspective however I think it is fairly simple. One note though . .  . a single core 8-16bit MCU wont cut it. Watch some of the demonstration / application videos for TI's Piccolo MCU to understand what I mean.

[bob g]

i was not suggesting a dc to dc scheme by using the computer psu
but rather ac to dc conversion via the psu

reason being the windgen produces AC anyway, so why not take it and convert it and let the psu do the conversion and rectify for use.

i am not suggesting using it as designed, but rather as an experimental base to modify and control via a MCU

it has long been my position that mppt has more to do with the use of a microcontroller and the code, and little to do with the converter section.

i would assert that any good converter that can be made to accept control from a MCU via proper coding can be made into an mppt controller,,, in other words mppt is all about the mcu and coding and really very little to do with the converter used.

or,, put another way a converter is a converter, but it cannot be an mppt controller without some form of control usually done by a mcu and programming.

for me the most difficult part is the converter, that is  why for experimental/learning project i would tend to use an off the shelf solution  to the problem even if it is not optimal... its just hard to beat
something that is available and mass produced and cheap as dirt surplus.  not saying one could be easily modified, or even that it can be, just saying it is likely worth the effort to give it a look to see if it could be reverse engineered to the point of determining if the unit can be made to accept control from a mcu and programing.

for instance many of the converters have a trim pot to tailor the output over a reasonable range which covers a nominal battery voltage charge range, 12-15 volts is not uncommon.

replace this trim pot with one of the stepper resistors and control the stepper resistor with the mcu and with programming and some feedback loop to the mcu, lookup tables, and programming one could alter the output of the converter to pretty much exactly what voltage he wants to charge the battery with..  this is one way i can think of

another deeper hack might be the  replacement of the modulator section of the converter (pwm) and use the mcu to provide the pwm signal to control the converters output voltage, again with feedback, lookup tables, programming etc to get accurate output for charging.

if we use an ac to dc converter like a puter psu, some of which are able to auto switch from about 90vac to 270vac we can then wind the wind gen stator to cut in at say 90vac and cut out at maybe somewhere close to 270volts ac... this certainly reduces the need for heavy transmission cables... also many of these converters can also operated on dc input current so we could rectify the windgen and send the product on two wires and do the conversion down line as described.

efficiency might not be good at all? i don't know, some converters psu's are very efficient some not so much. however from a learning standpoint i think there might be something to be said for it, and if it works even if not as efficient it might still prove to be very useful in applications where the factory made mppt are short on their promise.

having a converter that is programmed with 100% open source, so you can go in and tailor ever aspect of its operation, make changes to lookup tables, include feedback and sensors that might not be commonly used, etc has a certain attraction in my opinion

also i have been looking at some 5volt psu supplies that don't have the other voltages, some are very high amperage output units, 3 of which if isolated and put in series could be used to do a 12volt system.


bob g

[Yyrkoon]

Well I wont argue with you about DC/DC converters being complex . . . they certainly are. I've been trying to wrap my brain around some of the concepts now for a year or more. The wonderful thing though. Since in our usage we're just charging batteries. We can get away with a lot more. Such as we do not have to worry too much about noise, and  super tight regulation. etc. 

With that said let me give you a little information that may help out. Microchip's application notes AN1114, and AN1207 cover switchmode power supply topologies. Google should find these no problem, and they are downloadable as a PDF.

MPPT for Wind power is going to be tricky though. I remember having a couple discussions with Flux on these very forums a couple years ago. On the very same topic.  Just not all what we talked about. I do however seem to recall someone mentioning computer PSU's for such a purpose.

[boB]

1. the oem ought to allow entrance to the programming of their mppt controller so that the user or installer can adjust the programming to optimize efficiency for other voltage battery banks and other input voltages.... or

2. the drive to design and build a controller that is open source in the programming side of things is certainly high in my opinion. as difficult as it might well be.

#1 will never happen. Too many issues to keep up with. Least of which is maintaining IP on their source code. Another big one would be warranty issues. Now if you mean being able to adjust certain parameters, then sure. That could work.#2 would be very cool, and is possible. With creative commons licencing, companies can dictate how the hardware, and software can be used. Single/ personal use, etc.

A decently designed "solar" MPPT charge controller ~should~ already be highly optimized to give the best performance already,
but with some of today's controllers that have communications (modbus for example) and allow a set input voltage to be commanded, a programmer
~could~ write and control their own algorithms to run remotely on that controller from a PC or other embedded system.

boB

[Yyrkoon]

A decently designed "solar" MPPT charge controller ~should~ already be highly optimized to give the best performance already,
but with some of today's controllers that have communications (modbus for example) and allow a set input voltage to be commanded, a programmer
~could~ write and control their own algorithms to run remotely on that controller from a PC or other embedded system.

Seems like a warranty nightmare waiting to happen. But now that you mention it. I seem to recall reading a story from someone online a while back. Where they had such a controller, or something similar, and they blew their output stage out of the controller.

[boB]

A decently designed "solar" MPPT charge controller ~should~ already be highly optimized to give the best performance already,
but with some of today's controllers that have communications (modbus for example) and allow a set input voltage to be commanded, a programmer
~could~ write and control their own algorithms to run remotely on that controller from a PC or other embedded system.

Seems like a warranty nightmare waiting to happen. But now that you mention it. I seem to recall reading a story from someone online a while back. Where they had such a controller, or something similar, and they blew their output stage out of the controller.

Yes, absolutely could be a problem if they are not careful as well as the controller enabling weird and dangerous values to be programmed in.

Hopefully the controller being commanded doesn't let the user abruptly change the voltages, but should smoothly, safely and somewhat slowly ramp up or down to the commanded voltage as well as set limits on the voltages that can be commanded.  Those limits should also include not going over the units max current limit settings and things like that.

boB

[DamonHD]

Such a system would be interesting, but for me to use it I'd have to stop being cheap an actually get the interfacing kit to my (Linux) server and make sure that just running the interface didn't eat too much juice.

In any case, I could surely use the monitoring values to keep a track of Wh in and out in software as a bonus...

I would follow the boB school of thought that it should not be possible for me to fry the equipment our cause a fire with bad values (unless perhaps I use a 'super advanced' setting that tells me that I'm all by myself if the magic smoke escapes).

Rgds

Damon

[boB]

Such a system would be interesting, but for me to use it I'd have to stop being cheap an actually get the interfacing kit to my (Linux) server and make sure that just running the interface didn't eat too much juice.

In any case, I could surely use the monitoring values to keep a track of Wh in and out in software as a bonus...

I would follow the boB school of thought that it should not be possible for me to fry the equipment our cause a fire with bad values (unless perhaps I use a 'super advanced' setting that tells me that I'm all by myself if the magic smoke escapes).

Rgds

Damon

Yes, and try not to write 'ones' to "reserved" registers and bits.

boB

[DamonHD]

Also good.

(Apologies for several typos in that previous post!)

Rgds

Damon

[bob g]

its no wonder to me why the oem's of mppt's keep their coding proprietary and not allow access to just any joe smo, everything from product warrantee would be at risk to product liability and other issue would come flooding in via lawsuits.

i know i wouldn't allow anyone access to control code under any circumstances, no matter what his level of expertise might be! 

my point is this, as it relates to experimental units, the code needs to be open source so as to gain absolute control over the converter, however this comes at a very significant risk and responsibility.

most of which can be mediated if one takes the time to do so.

i probably would not mount an experimental piece of power electronics in the house, or any other building that if caught on fire would spread and cause serious loss of property or worse.

however is see no problem doing such in a standalone power shed, using some common sense, and with the idea of probably doing lots of worst case scenario testing before i left it to run unattended for any length of time... even then if the power shed is a cinder block structure, and the unit is properly enclosed as it should be, even if it went thermonuclear and had a major fire it would be self contained.

probably would not be sensible to just start cobbling together a breadboard mppt and leave it to run unattended in the back corner of the spare bedroom with stacks of old newspapers laying all about. or out in the garage with sawdust and a gas can for the mower under the bench it sits on.

no one here would do that i am sure   

bob g

[DamonHD]

Oh, you think I should stop doing that then?  B^>

Rgds

Damon

[bob g]

oh hell no!  flame on my friend!
lmao

bob g

[Yyrkoon]

Well, for the last month or so I have been playing with the idea of making a bench top power supply. A SMPS has always been in mind. So I was trying to talk my buddy here  into designing the power stage for me.

With the above in mind. The circuitry as I understand it could/would be the same as a MPPT DC/DC converter. The software on the other hand could be made to make the distinction between the two different types of operation.  Hell, I would imagine the same PCB could be used. With different component spots left populated/un-populated for the two different uses. Depending. Plus perhaps a separate output filter "card".

Just thinking out loud here . . .  but . . . yeah interesting.

Oh and yeah bob g -----> flame on ! lol