Open Source MPPT controller

[thingamajigger]

The Open Source MPPT Project:

I thought I would start this thread and get some input on what functions people would like to see as well as technical input for simplifying the controller. For those that do not know what MPPT is, it is simply electronics which allow a renewable energy generator to run at peak power aka "the sweet spot". I understand this is the "new thing" however I was never one to buy into what I refer to as Hype which is not practical and this issue will be addressed later.

To understand what is required I think it is easier to catagorize the sub-systems, 1)Mechanical -- the turbine and generator dynamics, 2)Electrical -- the dynamics of electrical energy and 3)Control -- how we control 1 and 2 to produce peak power as well as other control functions.

1)Mechanical:
This includes the wind turbine blades which will operate at peak efficiency at any given RPM and wind speed as well as peak power at any given RPM and wind speed, peak efficiency and power may not coincide. This also incudes the generator which also has a point of peak efficiency and power which may not coincide with themselves or the rotor characteristics. The trick here is to find some common ground where all these points are within acceptable values, for instance we could be running at peak efficiency in one area however overall we are not generating the maximum amount of power that we could be. Power matters however Energy matters more because Energy is the amount of power generated over a period of time.

2)Electrical:
This area gets a little tricky and efficiency can drop like a rock when things are not exactly as they should be. For instance some people think nothing of dropping $800 for a MPPT charge controller but then run long distances of small guage conductors at low voltage/ high amps so half the damn energy never actually gets to their fancy controller. Others fix this defect by running very large and expensive conductors the distance which is anything but cost effective because the most cost effective system must use the smallest cheapest conductors operating at the highest voltage and lowest amperage to reduce line losses. The MPPT is not a toy and we cannot expect good results from a system which is inherently flawed from the start, an MPPT charge controller should make a good system better.

3)Control:
This is an area which is relatively new to homebrew windpower where in the past people just threw some diodes on the output of their generator and tied this to some batteries which were usually far away, when the batteries were charged the excess power was dissipated in a dummy load --- this is the poorest way of doing things.
First if we are not charging our batteries then why is the damn turbine still spinning? and why would we generate the lowest voltage (12v) at the highest amperage which we know will produce the highest line losses?. Realistically the controller is the brains of the system and it should do the following, first allow the generator to operate at peak efficiency and power then shut down the turbine when it has done it's job. It should allow the generator to operate at the highest voltage to minimize line losses and charge the source at the proper electrical conditions. Batteries like low voltage/high amps while grid tie units should operate at the highest allowable voltage and the lowest amps for peak efficiency.

I'm going to throw another catagory into the mix which we will call practicality, it is completely impractical to buy a $800 MPPT controller for a 1000w wind turbine when we could simply build another turbine or two and easily show a 100%+ increase in power instead of the 10% or so with an MPPT controller. There is also the issue of the payback period and simply adding expensive options can easily push the payback period into the "not in my freaking lifetime" range. As far as practicality is concerned the best system will cost the least, last the longest lowering the cost of maintenance and generate the most power for the longest period of time.

This is the basic outline, looking forward to your feedback
Regards
Thing

[Watt]

Reading your posts in other threads, I had been under the impression you were already building MPPT controlers! 

[thingamajigger]

@Watt

Reading your posts in other threads, I had been under the impression you were already building MPPT controlers!

I have been building MPPT controllers but not for you, they are for me. An open source controller must be super simple, easy for anyone to build and have a few basic parameters to adjust. My controllers use some pretty freaky code and are way to complicated for open source, they are experimental and use new techniques so I can push the boundaries of every aspect of MPPT and every other control function. This is not what people need, people need something cheap and simple that works, that is why I laid out the framework for a new open source controller the way I did. This is not about me it's about you and everyone else, open source.

What would you like to see in a controller, what would you like to do?.

Regards
Thing

[OperaHouse]

I'll throw in my two cents.
1) I would like to see it on something like an Arduino UNO which is simple enough for people here to use and fairly complete.
2) It should have a low PWM rate like the UNO 490hz this makes transition time less critical and design easy.
3) Output stage shuld be multiple parallel units that are hot swapable.  Any single failure does not bring the complete system down.  This adds to the reliability and makes sizing the system easy.  Each module should be about 5A

[Watt]

@Watt

Reading your posts in other threads, I had been under the impression you were already building MPPT controlers!

I have been building MPPT controllers but not for you, they are for me. An open source controller must be super simple, easy for anyone to build and have a few basic parameters to adjust. My controllers use some pretty freaky code and are way to complicated for open source, they are experimental and use new techniques so I can push the boundaries of every aspect of MPPT and every other control function. This is not what people need, people need something cheap and simple that works, that is why I laid out the framework for a new open source controller the way I did. This is not about me it's about you and everyone else, open source.

What would you like to see in a controller, what would you like to do?.

Regards
Thing

" What would you like to see in a controller, what would you like to do? "

I am not looking for any further options than what I already have.

I'll stay with my Midnite Solar Classics.  Midnite Solar has ' perfect ' support, open to feedback and are very willing and able to build great products.


As far as my comment to your post as stated above, reading other comments you have made in previous threads I was under the impression you already had been working on an open source MPPT controller which you have been hooping and hollering about.  You commented on other options, such as Midnite Solars, being too costly.  The first post to this thread sounds as if it is the beginning stages of research yet you have promoted open source as you have already a solution previously.  You say what you have is not for me, you are probably right.  I've already bought what I feel is the best controller on the market. You could not have any idea if I could afford or would purchase what you have already built if I have no idea what your product can do or if it is reliable.  From you description above, you may too be lacking in the true understanding of MPPT technology, 10% would not be worth the effort if that is really what your product gains.  I'm pretty sure I've commented to you before on Midnites Classic but I'll say it again -  " Buy one, check it out, play with it and understand it " - you may find as I have, they are cheap for what they offer and can do.  It would take a group effort and lots of time to duplicate even part of that controllers " brain ".  To me, so far, I have only seen propaganda from you and, this thread is nothing more than fishing.

I say post what you have already built and maybe, just maybe, myself as a user/customer and others with technical experience will chime in being able take you serious.  Take the post as you want, I'll leave this one alone.
 

[bob g]

its my understanding, a true mppt besides having a rather complex micro
controller/ and code, has to run at very high hz in order to keep the converter components small and inexpensive.

if you wanna run at 500 hz or so, the inductors required are rather large and expensive compared to those that run at 50khz

the complexity goes way up when you run at higher hz, design becomes supercritical, ground planes and connections must be carefully designed, unless you want the FCC to fine the living crap out of you.

then we get to coding this thing to get it to be a true mppt, and not just a very good buck converter. this is my opinion is likely not going to be a product of a bunch of guys on a forum, unless the forum is populated by design programmers that likely work for someone already in the industry and because of employment contracts cannot be involved with the design of such a product, that is if they not only want to keep their job but also not end up in a lawsuit.

on the comment regarding a 10% increase, perhaps that is attainable via a group project, however quantifying and proving a 10% gain is problematic to start with and likely not worth the effort in any case.

for me i would not be interested in any controller that purports itself as being a true mppt unless it had a proven record supported by outside testing facilities in the range of at least 25% improvement over not using it.

there are also liability issues at stake, what happens if someone gets electrocuted, or it causes a fire?  something to think about in this age of litigation? i think so!

800 bucks might seem like a kings ransom for an mppt controller, however once you consider everything that went into it, the warrantee, the support, the liability insurance, the FCC approvals, UL listing, etc. and the fact that it is proven to work reliably and in a user friendly manner, in a market that is small compared to most consumer products, its a wonder they can sell it for 800 bucks retail and make a profit.

considering that for  product to sell for 800 bucks, it must be manufactured for a fraction of that price, to support the wholesaler and his overhead, shipping, the retailer and his overhead, and on and on.

so they have to build them for about 400 bucks a copy, which pays their overhead, R&D, support, advertizing, management, cost of funding the project, and the business has a right to a profit, so...

maybe they have 200 bucks a copy in the thing in basic parts, how the hell is one to build something that can come close to competing with this product for anywhere near 200 bucks?

when it is all said and done, my bet is the controller will not function as well as the corporate built unit, and will cost just as much as one at retail.

i think anyone that thinks he can build a competitive unit, and do it for less money, should be thinking of finding investors and not even consider an open source project. most especially when it comes to this technology.

usually when someone steps up talking about an open source project, especially when we are talking about a complex product, it generally ends up he doesn't have either the wherewithall to do the project, the experience, or a clear understanding of the product, or all three.

maybe i am wrong on this one, but somehow i don't think so.

prove me wrong, i will be the first to admit it, and apologize, and will promote you and the project on the two forums i manage. both of which are populated with members who would be very interested in a "true" mppt controller.

bob g

[dnix71]

This high tech push leaves me wondering. I grew up before even pocket calculators were available, and yet we put men on the moon and could have just as easily wiped all life from the earth with nuclear weapons guided by analog computers and mechanical gyroscopes.

I even had a course in college were we programmed analog computers because they were still extremely cheap and fast at certain tasks, like unfortunately, ballistic missile trajectory.

After 2 significant figures, most calculations don't matter much anyway. I knew guys with slip-sticks that could outrun anyone else with those new fancy Texas Instrument hand-held calc that had just come out.

I still keep a couple of model 6 CDV-700's around for the possible Apocalypse, because those things will still work even if the radiation kills me. The fancy digital chip rad detectors will probably fail from the EMP after the first air-burst.

Why not an MPPT design that is a simple analog impedance matching circuit fed by high frequency dc made from panel power?

[Watt]

Very well said!  I wish I could have answered in such a manner as you have.   

its my understanding, a true mppt besides having a rather complex micro
controller/ and code, has to run at very high hz in order to keep the converter components small and inexpensive.

if you wanna run at 500 hz or so, the inductors required are rather large and expensive compared to those that run at 50khz

the complexity goes way up when you run at higher hz, design becomes supercritical, ground planes and connections must be carefully designed, unless you want the FCC to fine the living crap out of you.

then we get to coding this thing to get it to be a true mppt, and not just a very good buck converter. this is my opinion is likely not going to be a product of a bunch of guys on a forum, unless the forum is populated by design programmers that likely work for someone already in the industry and because of employment contracts cannot be involved with the design of such a product, that is if they not only want to keep their job but also not end up in a lawsuit.

on the comment regarding a 10% increase, perhaps that is attainable via a group project, however quantifying and proving a 10% gain is problematic to start with and likely not worth the effort in any case.

for me i would not be interested in any controller that purports itself as being a true mppt unless it had a proven record supported by outside testing facilities in the range of at least 25% improvement over not using it.

there are also liability issues at stake, what happens if someone gets electrocuted, or it causes a fire?  something to think about in this age of litigation? i think so!

800 bucks might seem like a kings ransom for an mppt controller, however once you consider everything that went into it, the warrantee, the support, the liability insurance, the FCC approvals, UL listing, etc. and the fact that it is proven to work reliably and in a user friendly manner, in a market that is small compared to most consumer products, its a wonder they can sell it for 800 bucks retail and make a profit.

considering that for  product to sell for 800 bucks, it must be manufactured for a fraction of that price, to support the wholesaler and his overhead, shipping, the retailer and his overhead, and on and on.

so they have to build them for about 400 bucks a copy, which pays their overhead, R&D, support, advertizing, management, cost of funding the project, and the business has a right to a profit, so...

maybe they have 200 bucks a copy in the thing in basic parts, how the hell is one to build something that can come close to competing with this product for anywhere near 200 bucks?

when it is all said and done, my bet is the controller will not function as well as the corporate built unit, and will cost just as much as one at retail.

i think anyone that thinks he can build a competitive unit, and do it for less money, should be thinking of finding investors and not even consider an open source project. most especially when it comes to this technology.

usually when someone steps up talking about an open source project, especially when we are talking about a complex product, it generally ends up he doesn't have either the wherewithall to do the project, the experience, or a clear understanding of the product, or all three.

maybe i am wrong on this one, but somehow i don't think so.

prove me wrong, i will be the first to admit it, and apologize, and will promote you and the project on the two forums i manage. both of which are populated with members who would be very interested in a "true" mppt controller.

bob g

[thingamajigger]

@Opearhouse

I'll throw in my two cents.
1) I would like to see it on something like an Arduino UNO which is simple enough for people here to use and fairly complete.
2) It should have a low PWM rate like the UNO 490hz this makes transition time less critical and design easy.
3) Output stage shuld be multiple parallel units that are hot swapable.  Any single failure does not bring the complete system down.  This adds to the reliability and makes sizing the system easy.  Each module should be about 5A

I had exactly the same thing in mind, I have three Uno's and a Mega in the tickle trunk as well as a few nano's on the way which may to be used for this project. I like the programming because it's super simple and anybody can download upgraded software as it becomes available. Were not talking rocket science here and I think we should stick with something simple. As well the Arduino is all open source which fits perfectly with the spirit of this project.

It should have a low PWM rate like the UNO 490hz this makes transition time less critical and design easy.

Regarding the PWM, I already have an Uno as a buck/boost controller and the fixed frequency PWM can wreak havoc with feedback loops. As well using any interrupt schemes will cause serious timing issues. I let the Uno freewheel at max speed with minimal code using external analog components for signal conditioning. I will have to think on this one for a while, too many choices

Output stage should be multiple parallel units that are hot swapable.  Any single failure does not bring the complete system down.  This adds to the reliability and makes sizing the system easy.  Each module should be about 5A

I agree 100% , plug and play is the way to go.

I made a killer Wind Power GUI that will work with the Uno however I like where your going with this as far as simplicity is concerned and I had the same thoughts as far as the core functions go. It's hard not to as the Uno was basically made for this type of application and it's just a matter of how far we want to take it. I think the nano or a stripped Uno may be the way to go because those solderless terminals are a disaster waiting to happen.
Regards
Thing

[OperaHouse]

It is all about the inductor.  If I could find a couple thousand  really cheap I would have designed and sold one by now.  There is an appeal here for anything that looks like it was sourced  from Sanford & Son so my pet design would use a couple old APC computer inverters and maybe a microwave. Fortunantly this is only charging a battery so who cares avout sontrol loop issues.  Everyone says go at fast speed and commercially that is the only way to go.  I am trying to remind people that there is a beauty to the simplicity of low speed even if it is a bit clunky.  Wish all you wnat but a finished design with all the work done for you is not going to just show up.  In the next couple months I will be posting some trash can regulator designs with the UNO for educational purposes only.

[oztules]

Well, I see it a bit differently.

MPPT is just a thing you do when you have nothing else left to do.

More power lies with more swept area.

If you must MPPT, then Operahouse has the answer that best suits him, Thing the same.

When it's all over, it is just a buck with a jump table if you go digital, or a buck with an analogue simile.

I like the Opera idea of modules in parallel operation for the buck, I don't like the software angle though. Flux seems to have done it without software successfully a long time ago, and Gwatpe has done it some 7 years ago I think it was.... digital and analogue....long before midnight ever got near it.

The control of the buck is simple stuff ( I believe the midnights get the users to do that for each installation, which is fair, they never got an auto black box work as I understand it), the power buck circuit would be slightly more difficult for me. Although  I have lots of experience with off line converters, I have never had the need for a buck and so never bothered to make one.... I always needed mains insulation for my stuff. (.... pushpull or half bridge )

Keep at it Thing. I wont be making one anytime soon (ever) as I don't feel the need. Simple transformer matching leaves us with easy an 2kw 24/7 a lot of the time on the AWP, and we can't use that anyway.... with all the mod cons inc elec. hot water ( another property)

The axials biggest problem is their furling, and doing a Mr Olsen with a bunch of those midnight devices made him very happy, and allows him to avoid most of it... and he does it like everything he does... very very well (albeit expensively).... but as he said to me... money is no object.

I happen to think it's poor use of resources, as that money would be better spent on cheap solar. Now solar is so cheap (dollar a watt even here), mills for anything other than playthings is not on my radar..... and so I'm stuck with hundreds of neos I'll never use....... but that happens. My two 4 meter units are now novelty items.... even though 5kw each is easy for them with load matching (article somewhere here).

I would never buy a manufactured product for this application, as they are too far away to help, I'm too isolated to rely on them, and they are too expensive for what they are...........and I could do the same thing myself if I felt it was a sensible need.....and I don't... using a  better designed mill, with and a boost circuit will have improvements aplenty if folks find high powered bucks a problem.

Somewhere here in the past I did an article on my little 600w boost for amusement.... up to 60A @ 14v Tl494 based (used to charge a 12v batt from 6v..... coz someone asked about it) I had later intended to use an op amp controller on it for a booster, but the mills did not need one here.... wind is too good.

For me, if you have a mill and need an mppt to try make it useful, your in the wrong place to have a mill....... or you built it too small. That said, as a project it is a good thing to try. If others can benefit, then better still. Will follow and help if I can add something useful, but Opera seems to know more about buck  pwm than me.

Keep your chin up, and don't listen to the knockers with no experience anyway.

Now..... a Big ..up my nose rant:

Bob g  I don't need you to tell me how you think they should/do  manage their business..... thats for them to represent..... and if that is you..... then your a selfish person to boot.

I  understood this forum was for helping folks achieve their  aims/dreams , not to support world class manufacturers ( you think they are ... I don't know first hand, so benefit of the doubt goes to them)

If people want to dabble in this sort of project with their spare time or cash...... then bug off, and rant somewhere else where you want to curb others liberties to play, and tell folks what they should or should not do..... or become a politician, so you can curb everyone's talents  all at once....I have no time for armchair knockers, and less for those knocking for the sake of their favorite manufacturer......

And this:

"usually when someone steps up talking about an open source project, especially when we are talking about a complex product, it generally ends up he doesn't have either the wherewithall to do the project, the experience, or a clear understanding of the product, or all three.

maybe i am wrong on this one, but somehow i don't think so."

What the hell do you think the thread was started for and what do you think a forum is for...... just to tell us how stupid we are for playing with things, and how smart you are for buying instead..... try Ebay and get your jollies there.

And this:
" i think anyone that thinks he can build a competitive unit, and do it for less money, should be thinking of finding investors and not even consider an open source project. most especially when it comes to this technology."

Your mind may work like that, mine  does not. Fortunately most here like to help other folks achieve their goals, not screw them over.

Open source is by decent people for all people.... and I subscribe to it, and I do hundreds of articles to help people where ever and when ever I find I can. I'm not selective. If I can help I will.

I do it on all kinds of projects for all kinds of things.... and if I find a better way to do things I tell people so that they may benefit from my experience.

I don't see this in your behavior.

In short, if you can't help usefully..... then keep your opinions to your self. You don't seem to understand anything to do with high power buck, or code to help anyone with.....except what you may have heard from other slobs...... so  why are you commenting?

Do you have any circuits, pcb layouts, or any completed project using even simple pwm. I have.... they are on the net... where are yours?

If you can add useful material, then I will stand corrected.... but no more pious piffle please.

More crap about how hard it is or how silly folks would be to TRY,or why they should buy your favorite brand of gear is not useful in any way.

They may try it because they think it's hard.... ever thought of that? It's called a challenge, that's what makes life worth living.

And yes,I have a changfa 195 as well...... I rebuilt it from the ground up.
I found it in a heap on a farm, it was seized in the bore, and bits missing..........I did it for a CHALLENGE.
I had no parts, not even gaskets, but I had a lathe and press and milling machine, and I made my own parts.... how silly of me when you can just buy them off the shelf where you are.

Well you can't here. It was a challenge......... but after reading that crud, I think I just might dump it in the tip on principal. I don't use it as I prefer the Lister, and don't wish to be reminded of the likes of you.

End useless rant.


Forums used to be places to find help, share ideas, solve problems together..... now it attracts sanctimonious wannabe smarties  to howl those down who do ask for help..... read the old board.... we used to be different.

Damn disappointed.




................oztules

[OperaHouse]

I think you are right.  The greatest benifit is for mills under a couple hundred watts in iffy speed conditions.  When you build your own generator and blades it is never known what exactly you will come up with. A micro can help with that initial cut in.    What bothers me most is I see a lot of power generated and just wasted.  There is little discussion of integrated smart systems diverting loads.  MPPT with a UNO would only be a side issue.

Back to technical. An external watchdog timer should always be included just in case.  It seems a sizeable wind generator isn't the best thing to practice writing code with. I am not going to attach my name with anything I have not actually built and tested.  I built a small wind generator only trying to get a few watts average.  That was a disaster, the site was too restricted.

[oztules]

"An external watchdog timer should always be included just in case. "

Thats an interesting term. I guess that is what I had done with mine. I used one of the mills to recharge the electric car I built, and instead of using dump loads and things to control it, I used a set of op amps that monitored the battery voltage. If the battery voltage went over "float" by about 6 volts for more than 4 CONSECUTIVE minutes, it would shut the turbine down. If it fell back within speck within that 4 mins, it would reset and wait for over voltage to occur again and start counting again. This stopped gusts or a sucession of gusts tripping it. It needed to be for real to trigger stop. Voltage was measured at the turbine tower.... so line loss was part of the deal. Timing was by a 14 bit shift register with osc built in....a 4068 I think from memory.

Because it was wired for a booster that it did not need, it was a very low resistance stator, and could handle multi kw for extended periods.... but if the wind was high, the volts would rise too far too fast, and if it went for a 4 solid mins or so, it was a fair bet the wind was excessive,  or the batts were fully charged, and were not holding the mill back. Either way, shut down would occur. The stator had nine half inch bolts holding it..... I would not even begin to try this with the usual three I see here. If it gets to runaway condition (lost batteries) for 4 mins here, it is a safe bet it will be breaking from over 5kw or more..... lots of torque. But no problem for it.

I found that for three years or so, this worked very well, as any high winds would see the mill stopped, and batts charged. In normal winds, the batts would be charged and partially equalised..... and if it got away from poor batt connections or any other reason, it would shut it down.

The mill never failed or burnt up in the whole time it was up in that format...... but here the magnets got a bit rusty..... so it is down at the moment for a rethink of how to best protect them..... and no the crappy chainsawn blades never failed in a runaway condition when it was tested for it..... but the heart nearly did waiting to see what would happen.

I now will call it a "An external watchdog timer" I think...... a very good safety device I can attest to.



...................oztules

[bob g]

Oz

so sorry to have blown your dress up my friend, it was not my intention to piss you or anyone here off.

however i stand with everything i posted, and was careful to also state that if i were proven to be wrong, then i would be the first to admit it, and not only apologize but also promote the end product.

having been on this forum for many years, albeit not as actively as some in recent years, "if" you go back to the early years of this particular forum  you will find i was one of the first if not the first to posit the question about load (impedance) matching via a converter. it would have been open source and basically exactly what the OP has in mind.

this is not my first rodeo either.  again i never meant to piss off anyone, however if i have done so perhaps it should be used a motivation to actually do the project and see it through rather than take it as an opportunity to attack me for being less than optimistic about the prospects.

as for your changfa, good on you for saving the old girl. i too resurrect old engine's sometimes without anything more than my hand tools. i have hand knurled pistons, reground piston rings with a hand stone reworked babbit brgs, weld repaired all manner of castings, welded up pistons, and replaced gear teeth...on and on and on.  i will match my wrench with yours any day of any week. i don't think you want to go there my friend.

as for microcontrollers i have reverse engineered a plc and its backplane, along with its control modules, something i was told was not going to happen, but something nevertheless i did and got an email from the oem asking for first rights to whatever product i came up with.

i have spent thousands of dollars on books, parts etc learning about controllers in particular, and as i stated earlier i stand by what i wrote.
i can program a basic stamp just fine and find it capable of doing everything i need, it is well supported and i like it. probably not the cheapest or best for this project but an option nevertheless.

i guess i would like to end this with the following.

my purpose is never to discourage anyone from doing anything they like, however just sitting back and not providing some of the realities is not being responsible as a member of this or any forum in my opinion.  there are too many forum's that support all manner of idea's where the members get all excited, spend all sorts of time and money, and end up with nothing to show for it but aggravation. i am not saying this is the case here!

and finally, some thoughts fwiw as it relates to a mppt controller for a wind generator. the FCC concerns are very real, the liability issues are equally real, and the complexity if you are going to go to the trouble to build it and make it right are also very real.

now to add constructive input
in my opinion you need to have the following control handles, at least...
1. a rpm sensor to read the rotor speed, that can be hz from the generator fairly easily.
2. a windspeed sensor
3. battery voltage sensor
4. generator amp sensor
5. battery temp sensor,
6. maybe an embedded temp sensor in the generator stator core? (if you going to design it, then maybe you provide for protections too?)
then...
7. the converter should be a buck converter, probably not worth the effort to use a boost version as there is very little power down low anyway so why spend a lot of effort trying to boost something that is very low to start with and spend the resources and grief to capture it.
going the boost route allows the builder to wind for as high a voltage as he can get, this is not only useful for the buck converter, but lowers transmission losses as you already know.

then the microcontroller along with is program code, a couple look up tables and a lot of testing then you will probably get something useful.

the problems as i see them aside from those previously stated,

wind speed is hugely variable, which makes meaningful real time testing and results that you  can interpret to make subtle changes in the coding is going to be very difficult.

battery charging is also variable and non linear, especially over the life of the batteries, and over the temperature ranges during not only charge/discharging cycles, but often across the seasons unless the batteries are in a climate control room?

so you end up with a problem with many factors that have to be accounted for, which would not be as difficult if they all held still rather than move all over the map!

i know you already know all of the Oz, i only state this to illustrate that i have "some" basic understanding of the project, the procedures, and problems likely to present themselves all along the way before a useful product emerges.

again in my opinion

this is going to be an ambitious project, one that is compounded by the distances between all those involved and the various levels of experience.

now if you could get together a group of perhaps a dozen well motivated people, that have a decent grasp of the technology, "and" have a leader that can keep the group focused, "and" could get everyone to agree as to what needs to be done, how it will be done, and with what controller and what converter, then you likely will succeed... otherwise it is much more likely to happen in one guys shop in my opinion, where he has control over all aspects of design, build and programming.

its been my experience that most folks that get involved with this sort of project (read that as complex) and are members of this sort of "serious" forum, are generally type A folks.  i don't know about you, but getting a room full of type A folks to agree on anything is like hurding cats! 

i can just here it now, "i want aurdino" , "i want a stamp" , "i want a cubloc", "i want a motorola" , "i want a plc"... "i want buck" "i want boost" " i want both"

i have always thought of this as a worthwhile project, and have been a strong proponent for many many years, just not overly optimistic about the prospects of success at this point.

now that is my opinion, take it for what you paid for it!

group hug?

bob g
ps. we can't get changfa's anymore here in the states, the epa has ruined that, and in doing so has made it also a necessity to import the parts direct from china. so we are in the same boat as you guys down under.

[bob g]

Oz:

respectfully i would like to address your concerns

i went back and reread my original response to the topic, and while i might have stated things a bit more politely (perhaps), i don't see where i stated anything that was false or misleading. i was also careful to add that if i were proven to be wrong i would be the first to apologize, etc.

then i went to your "rant" (your term not mine) and for the life of me i don't get where you are coming from at all!  you derive much from about me from my reply that just ain't so my friend, perhaps that is my fault, i have been wrong before just ask my exwife!

for the record and for your information, which might help you understand where i am coming from.

1. i am not an oem of any product
2. i am not a retailer or wholesaler of any product
3. i own no stock in any alternative energy company (unless buried in a mutual fund)
4. i have been active in alternative energy research for a long long time, on my own and not working for someone else.
5. i own two forums, one listed here and the other relating to your beloved lister's and their clones the listeriods.
6. i am very supportive of anyone that has a good plan or a willingness to learn. (however for me to be simply a cheerleader and not relate some of the challenges likely to present themselves is just not responsible in my opinion.)
7. yes i like the changfa, is it perfect? no! is it more efficient than a lister? yes!
and...
8. i have long been a strong critic of certain big name mppt controllers, which in my opinion are long on marketing hype and short on performance.

i really think if you take the time to get to know me, or do a little background on me, you will find that contrary to your belief i am not a corporate guy. i am all about helping the little guy and his projects, after all i am a "little" guy as well.

as for your threat to pitch your changfa in the tip (dumpster here in the states), perhaps you might want to wait till cooler heads prevail, or donate it to someone that can put it to good use. either way is fine by me, i only own about a dozen of the changfa's (among many others).

bottom line is i have followed your posts over the years and have a certain respect for you and your postings, i can only assume you have had a bad day (i have many myself). i also know you possess  a higher intellect than what you illustrated in that rant directed toward me.

no worries here my friend, i have a very thick skin!



bob g

[JeffD]

In 2010 I used an Arduino based board to control a buck converter for one of my small wind turbines.  It was a steep learning curve with many hours of frustration and burned components but it was all worth it in the end.  I posted a little bit about it on this forum back in 2010.

http://www.fieldlines.com/index.php/topic,143825.msg968747.html#msg968747

http://www.fieldlines.com/index.php/topic,143829.msg968953.html#msg968953

http://www.fieldlines.com/index.php/topic,143829.msg968973.html#msg968973

The two biggest hurdles I had were board layout and the inductor.  For board layout I read a lot of the application notes put out by mosfet manufactures and looked at the boards for commercial buck converters.  I still ended up making seven boards before I got a layout that performed well.  The inductor proved to be a big hurdle for me but only because of my lack of knowledge at the start of the project.  Using the online calculators to determine the inductance I needed was not enough.  Size and core material were critical too.  After burning up a few mosfets, I learned real fast about inductor saturation when  DC bias current in the inductor got above a few amps.

This kind of project can be done by those that are willing to do the research/learning and can deal with failures.  The way I approached the project was to do it in small steps.  I built a small buck converter based on a TL494 controller that could handle about 25 watts going into a 12 volt battery.  The next step was adding the arduino to take the place of the TL494.  The next few steps were all about controlling the load from a very small turbine (0.5meter).  Then I switched over to a synchronous buck converter scheme which proved to be a challenge.   Then I stepped up to the 1meter turbine along with increasing the power output of the buck converter.  This is when the flaws in the design started showing up.  But again it was due to my lack of experience with power electronics and not doing my homework.

It took me about 8 months before I had a board that could handle 300 watts.  I built a total of seven boards starting with a board that could handle 25 watts.  It was at the next power step of 100 watts where I started having a lot of problems and most were due to board layout.  It was at this point I also found out why its critical to have low ESR caps on the buck input.  I had opted for the cheap caps figuring it would be ok at the low power level I was working at, boy was I wrong.  I spent about $300 in parts and board manufacture during that time period,  but I easily spent over $800 in my labour and research time.  For me it wasn't about the money but the learning experience and challenge.

[OperaHouse]

I have been doing some experimenting with extremely simple high side drivers for N chanel FETs using plain vanilla LED/transistor opto isolators or small audio transformers similar to those found on old modem cards.  Transformers as you can imagine won't work at 0 or 100% PWM duty.  Don't have the numbers handy but it is something between 40 and 70% PWM for a real nice drive with a little waveshaping.  That isn't such a hard work around when you have a micro to drive them and have two or more parallel converters.   The same is true for the opto isolatord which get the gate drive voltage.  When using these with a solar panel the input voltage is pretty much fixed as is the output voltage.  The PWM works well from about 15% to 85%.  Spec Jockeys would find this intolerable but it is a prety easy work around in code.  The very small number of common parts make for interesting possibilities.

I've mentioned this before but PWM is prety much yesterday.  I have played around wit some of the TNY series switching regulators.  These are either on or off with only a control pin.  Given that they switch at over 100K there is plenty of opportunities to turn them on and off.  These little switchers can easily drive small common mode chokes which peak in that vicinity.  A little wave shaping and it makes a good high side fet drive.  A simple solution right out of a junkbox.

[oztules]

I think you can do a bit better than that with transformers.
Here is a computer supply that I have butchered to be variable voltage, variable current low impedance drive, note the waveform and duty cycle. Thats only at 35W, but at least it gives some idea of waveform through the tranny and switches. This is from a 1 turn winding around the transformer core. So is a accurate reflection of the magnetics in the transformer core itself.

So at 7v 5A we get this:



and this is squeezing it to max output for a 250 watt supply ( over 300w here) 30v@12A)



So transformer coupling should be enough to get a 5:1 stepdown, and this should be enough hopefully. A transistor and diodes on the gate will help shape it as well. Even at full pulse rate (90% here) it is not such a problem as this will be at lower power levels. The transformer seems to still look fair to me and solves myriads of problems. Crank it up and stick a zener on the gate perhaps for better shapes.




..................oztules

[thingamajigger]

@JeffD
I feel your pain brother, but I have not had any problems using the Arduino to drive any circuits and I jerry rigged a circuit to buck a 1Kw input in probably 20 minutes. Mosfets can be tricky but there are a few simple rules, first drive them hard and fast with a mosfet driver, opto-isolate wherever you can, use big freaking heat sinks, when we switch fast and hard then transients are an issue so power isolation is required **** Do not power the Arduino from the source it is controlling**** without a major amount of power conditioning ie. filters and regulators and an isolation transformer.

@Bob g

7. the converter should be a buck converter, probably not worth the effort to use a boost version as there is very little power down low anyway so why spend a lot of effort trying to boost something that is very low to start with and spend the resources and grief to capture it.

I would disagree, if you go to this site ----http://www.thebackshed.com/forum/forum_posts.asp?TID=2164,   
Windmill DIY Analog MPPT Circuit Design and Construction

The first thing we see is a real world graph clearly showing why a low cut-in and harnessing low wind is so important. Now one could say so what we only gain  less than 250w by changing the cut-in from 7mph to 3mph however if the average windspeed is 5mph and it occurs 90% of the time and we considered the massive amount of difference it would make over the course of say 10 or 20 years then we realize what should have been obvious. Power does not mean diddly squat and it is ENERGY which matters the most which is power over a span of time, our world is not run by Power it is run by Energy. Any turbine with a low cut-in and MPPT will obliterate any equivalent turbine without it in regards to Energy production and Energy is all that matters.

Regards
Thing

[JeffD]

Hi thingamajigger,

I am curious about your 1KW buck converter,  it sounds like it would fit the bill for this project.  Your expertise would greatly help in solidifying this open source hardware initiative.

Can you post a schematic and parts list?  Maybe it could be the first version.

[bob g]

thing:

hmmmm, 250 watts using a mmpt and dropping from 7mph to a 3mph cut in speed?

not to beat a dead horse, but

according to the "cube law" as wind speeds double you get 8 times the power, and according to "betz" you can only get something like 57% of the available wind power to the blades, and give the typical wind generator is ~50% efficient to start with...

how big of a machine would it take to be able to net anywhere close to 250 watts with an mppt within this envelope?

the machine likely would be quite large, something capable of making 1500-2000 watts at 7mph, might well make 250 watts at 3mph unassisted. now to make a gain on that?

btw, i understand the difference between power (watts) and energy (w/hours) and yes it is a noble thing to gain all you can from a machine.

if you go back and take a look at my statement i stated "probably"

this is not to say that there are not exceptions to ever rule, just that the exceptions will probably not be the rule.

anyone building a machine where his windspeed is on average 5mph, probably should look into something other than windpower? don't you think?  or... maybe he ought to look into much larger swept area to start with? or multiple machines?  or living a spartan lifestyle?

back to buck vs boost converters

if you are going to the trouble to engineer and develop a converter, there certainly is no harm in doing both a buck and a boost, or a buck/boost hybrid.

in closing don't ever let me or anyone else talk you out of something you are passionate about, rather draw if not inspiration then motivation from those that would tell you otherwise.  (like me   )

i can't tell you how many folks, including EE that have told me, "bob that can't work, or won't work as expected", (or some other negative stuff) . As much as i don't like hearing the negativity, there is always something to learn from the interchange, sort of like sorting the wheat from the chaff.

i realize i put a lot of wet blanket on your plan, however if there is only one thing that you hadn't thought of, and can use to your benefit, then perhaps you will at some point look back favorably and take back half of the ugly things you think about me?   

(ok, i will take back half the ugly things i was thinking about you) is a long running joke in my circle, so don't take that statement with "any" seriousness!

now that we got that out of the way...

i can think of one possible benefit to the boost converter, but i want to further develop that thought on another thread as it relates to the use of a transformer pack. 

good luck in your endeavor

bob g

[thingamajigger]

@JeffD

I am curious about your 1KW buck converter,  it sounds like it would fit the bill for this project.  Your expertise would greatly help in solidifying this open source hardware initiative.
Can you post a schematic and parts list?  Maybe it could be the first version.

I'm working on it when I can find the time which is an issue, I'm in the middle of rebuilding my 5' 3pt pto rototiller which tried to destroy itself, my trees need cultivating, the grass is growing like wildfire, the hay is almost ready to cut and I have about a mile of new fence to run, life is getting in the way of the all jobs which ain't really my real job.
Right now I'm rethinking the project for plug and play operation like sensor shields then I have a GUI which needs to be trimmed down so people can understand what in the heck their adjusting. I will post my old GUI tonight so you can understand what I'm talking about, it's a gong show and is overkill to the extreme but gives a great deal of flexibility for testing. The Buck/Boost or Cuk is pretty straighforward and shouldn't take long, one is easy, a programmable phase differential between more may be an issue. In any case it ain't happening overnight but I'm pumped about how it's shaping up and I think the open source end product will look very professional, maybe moreso than most retail considering the GUI.

Regards
Thing

[thingamajigger]

@Bobg

according to the "cube law" as wind speeds double you get 8 times the power, and according to "betz" you can only get something like 57% of the available wind power to the blades, and give the typical wind generator is ~50% efficient to start with...

how big of a machine would it take to be able to net anywhere close to 250 watts with an mppt within this envelope?

the machine likely would be quite large, something capable of making 1500-2000 watts at 7mph, might well make 250 watts at 3mph unassisted. now to make a gain on that?

I understand what your getting at and agree, I believe the link I posted was using a 20' Diameter rotor which is fairly big however I tend to think forward incrementally. Now let's say the cut-in is normally 7mph now what happens at 6.9mph?, well probably nothing so we could imagine there is an infinitely small division between sitting there doing nothing and the cut-in. So if we reduced the cut-in to 6.9mph then we could generate maybe 250w or more which over 20 years in marginal wind could produce some very big numbers. If we are now at 6.9 from 7mph then why not 6.8?, you see where this is going it's not 7mph versus 3mph it is every single increment inbetween them.
Now the kicker, why not low windpower?, it's not as if there isn't energy there but the problem isn't the wind it's the wind generator and the maintenance when it's working. The fact that nobody knows how to build a zero maintenance turbine that will last for 20 years capturing the energy at all wind speeds in which case energy production could easily double or triple. If it didn't cost anything more to have the turbine spinning all the time including low winds in a high efficiency system then why not?.
It is kind of like saying there is no power in solar after 3pm so I will just cover my panel with a blanket, I just don't get it.
Regards
Thing

[bob g]

good luck with the project, not sure how you figure on getting the power in such low winds, but go for it!

might well be easier to build two turbines, one to harness the prevalent low speed winds which would likely be a very large machine, and one to harvest the higher speed winds which would likely be a much smaller machine than the former.

if you can accomplish your goals, you will have done something no one else has been able to do, that i am aware of.

the part that bothers me is this

a buck, boost or buck/boost converter cannot make more power than what is available to it to start with.

you mention getting 250 watts additional dropping from 7mph down to 6.9mph windspeeds, with the use of a converter.

again unless you are planning on building a rather large (ok like the 20'example) there just isn't that kind of power available in these low wind speeds. (call it power or energy, i don't really see the difference for the sake of illustration)
there has to be more than an extra 250watts on available input power/energy in order for your converter to capture it?  at least as i understand it, or rather i think i understand it.

bob g

[bob g]

giving it more thought, how about this option

a buck/boost hybrid, coupled to a decent size machine of say 17'

design the machine for perhaps a 8mph windspeed cut in under normal operation,  now then anything under this speed will not reach cut in without some sort of boost controller.

center the buck/boost design around maybe 10mph windspeed, so that as the windspeed drops and crosses through the 8mph threshold the converter continues to boost down to maybe 5mph. as the windspeed passes through the 12 mph threshold the boost converter takes over.

something like that might well over time make more power than simply building a windgen to directly charge a battery bank.

is this about what you have in mind?  maybe with somewhat different parameters of windspeeds?

bob g

[fabricator]

I don't know about the rest of you guys buy I'd vote for Oz for world emperor any day.

[GoVertical]

Hi, if the 5 amp power module concept was used, would you have to increase the switching frequency as more modules are added in parallel?

[oztules]

"Hi, if the 5 amp power module concept was used, would you have to increase the switching frequency as more modules are added in parallel?"

No...... you can stick what your comfortable with.



.............oztules

[tecker]

 If this is a info share of Maximum power point . I like the circuits in with the device that's input power . For instance the circuit from Go V  is of a three phase genset .
  The Fet even though it's set to buck will be more efficient if on the battery side of the coil and not shunting the coil back emf to circuit ground . On the down stream side the diode after the coil is not necessary The battery will do the buck and the fet can be more of a PWM .
  The power track can be on the Turbine somewhere in the form of a pickup coil in the stator . References RPM and be anded with the 555  pin 4  frequency side or a cutin Program reference to mirco controller  .
 The real power tracking is in that ever changing Battery impedance . The best way to since battery impedance is with a Running capacitor. I'll punch in the changes to the circuit from Go v and repost see what you think   

[joestue]

"Hi, if the 5 amp power module concept was used, would you have to increase the switching frequency as more modules are added in parallel?"

when you get up there around >20 amps you have to phase shift all the individual buck regulators, to save money on the capacitors you don't need anymore for the input filter.

[thingamajigger]

@All
It seems I'm behind the times again ---- http://www.timnolan.com/index.php?page=arduino-ppt-solar-charger.
The schematic, parts list and software are at this link and it looks like he has done an excellent job on documentation. Oh well I will keep plugging away .
Note the date on the schematic bottom right hand corner --- 2009.

Thing

[GoVertical]

Hi, after fabricating a few different Buck converters and testing them with different PMA's, dealing with the max input voltage value is a big problem. If the input voltage is above 100 volts DC the parts for the circuit become very expensive.  Is there away to limit the input voltage??? Setting a limit to the max input voltage should be established for the project. 

[Flux]

As you say, the converter alone is not all the issues with a reliable mppt converter. Protection is vital and most likely will be more complex than the converter alone.

Unless you can run your alternator within the ratings of the converter you will need something to limit the input voltage to a safe figure. Converters are not too difficult up to about 100v if you can solve the driver problem. Over 100v design and layout becomes very critical. Protection for loss of battery and excess input volts and various other things will be essential if you are to have something completely reliable.

You can simplify things a lot if you rule out the loss of battery condition but even so you will nearly always need some form of limit to the input voltage.

This is where I believe the Classic has been so much of a success and I strrongly suspect an open source design is likely to fail.

It is good to see more interest in this side of things and electronics experimenters should have little difficulty building a working converter but producing something bomb proof for the non electronic people is something of a challenge and one that the designers of the Classic seem to have mastered.

Flux