Open Source MPPT controller

[GoVertical]

Hi, they limit the VOC of the input. So you can not use any PMA that produces a VOC above their circuit input limit. I found using a isolated DC to DC supply with a FOD3180 gate driver works will for both charging batteries and resistive loads. The ground of the isolated power supply has to be connected to the Source of the N channel mosfet.

[GoVertical]

Hi, is it possible to use high voltage capacitors in series to divide the input voltage to a more useable level?

[OperaHouse]

Capacitors divide AC voltage.  That is commonly done to eliminate p problem with capacitors called corona.  Rapidly changing voltages in a capacitor field cause microscopic air bubbles in the wrap to light up like little fluorescent tubes and that action eventually causes the insulator to fail.  X2, across the line caqpacitors actually have a non connected internal metal foil that makes it two capacitors in series.  With half the voltage on each capacitor, these corona effects can be reduced. 

You are getting bits and pieces of theory and it hasn't quite all jelled together yet.  I did like your RC network on the FET, who taught you that.   While higher voltqages can cause some issues it makes a lot of things easier.  If I built a mill, I would want it to put out 200-400V and use PC power supply designs.  Be aware that flyback designs have a practical limit of only a couple hundred watts.  These are very simplistic but you really have to knjow what you are doing to get over 100W.  That is why I suggest parllel circuits if going for more power.  I am now at the camp and hopefully in a week I can start building some stuff after I get all the plumbing leaks fixed.  Off to ebay to order another uno.

[GoVertical]

  I did like your RC network on the FET, who taught you that.     That is why I suggest parllel circuits if going for more power. 

Hi, I seen the RC snubber that is across the mosfet in a paper on buck converter design, I have not tried it yet. They use them to reduce voltage spikes. I am going to set the input voltage limit to 100 volts and use a PMA that does not produce a VOC higher then 100 volts at max RPM.  I am proceeding with the parallel buck converter idea. One controller board with three buck converters. Each buck converter will have it owns input capacitor. I will post results when available. Thank you for the help. 

[joestue]

A boost-buck configuration, using three low side mosfets, 3 high side fast rectifiers, and three ferrite inductors on the line separated from the turbine by adequate capacitors is certainly an option. (followed up with a buck regulator if needed)

The problem with such a topology is that you've got a very significant ac voltage riding on the output. (or input, whichever is grounded)
this voltage is sufficient to require more than just a capacitor on the line side, you also need a common mode choke.
depending on the frequency such a common mode choke could be half a pound of copper and ferrite, or it could be 10 pounds of 4 mil silicon steel. ($$)

the reason why i don't like it is because if you can get away without a common mode choke then you could get away without  the inductors and capacitor at all.. just pwm the lower fets and use the inductance of the turbine, set the frequency high enough so the current ripple is less than 20% at full load.

however, it is probably the best option if you do not want to lose the additional voltage drop of two more diodes, and the conduction losses of the boost switch.

The inductor of a traditional boost converter does not even have to handle the full load current (beyond which you've long turned off the boost converter) if you bypass it with a diode, and ensure that the diode's voltage drop does not cause too much current to flow through the inductor.

[OperaHouse]

A couple of days ago I was thinking about building one of those "store bought" sync converter designs.  You know, using one of those single chip hi/low drivers that saves you all the work.  So I looked on ebay to see what I could get in quantity cheap if I changed the design around, IR2101, 2104, 2110, 2111, etc.  Also did some searching to see what people were building.

I found this site where someone couldn't get one to work.  The circuit was familiar having just seen it in a data sheet.  It would work for a couple of seconds and just stop.  Everyone was trying to be polite and explain why the bootstrap wasn't working.  The OP was convinced it was some phase shift and stopped posting.  Someone should have told him the circuit would never work because it had a fundimental problem with how the output was connected.  Special function chips are great.  They can turn almost anyone into a successful designer.  The problem is we can connect a number of these building blocks together and loose sight of what the overall circuit is doing.

I was looking at that Arduino circuit and thought the top FET that replaced the normal diode was pretty trick.  It used the high side driver to get the voltage over the supply rail and the gate capacitance of the FET to keep it on.  I always like to save a part.  But the circuit is a little suspect and care must be taken in how it is applied.  Even the slightest leakage in the high side driver can turn the FET on.  I don't doubt that it works on 12V or even 24V but at higher voltages the driver could break down.  A lot of people don't like leaving the gate of a FET just floating.  Putting even a several hundred thousand ohm resistor as a gate drain is enough to turn on the other FET switch.   For wind that extra FET is not needed when you have the diodes in the mill.  Funny thing is that early power FET did not have the internal protection diode, Just try and find one without the diode today.   That internal FET diode is the only reason that additional FET switch is needed.

Back to chip selection.  Some of those have seperate HI and LO inputs.  70 cents more seems like a deal to not have to add dead time and invert the signal.  And who needs enable.  I'm going to be running all the time, don't need that.   There is something to remember, run at 100% duty cycle and it will be on till the bootstrap capacitor drains down.  Generally you need at leat 1% off time to charge the capacitor.  Sorry GoVert but hat is the reason you need an isolated supply with your just high side driver.  If you could get it to start oscilating it would run!  There are 555 designs to run the bootstrap.

Now here is the real fun part of sync converters.  The low side FET replaces the diode and the low on resistance saves you all that diode loss.  Remember those HI/LO side drivers without the enable input.  Either the hi or low side is on.  Having the low side on is great as long as you are still dumping the inductors energy into the battery.  At some point that current will go to zero and then current starts flowing FROM THE BATTERY.  Something to think about when writing that code.  Those old diode designs were pretty foolproof!

[GoVertical]

Hi, I completed new concept drawing for the buck converter. Basically it the same circuit as before. Each parallel buck converter will use a IRF840 n channel and bridge rectifiers. Again I am sure if will work and comments welcome. 


I had a high side boot strap circuit working with a resistive load but I could never get it to work with a 12 volt deep cell as the load.

[joestue]

The reason the bootstrap doesn't work with a voltage source for a load is because there is nothing to pull the source of the high side mosfet to ground.

we've been over this before

[tecker]

I have been distributing the power directly to lights and motors (Mostly fans) and heating elements since I started . There's not really been a plan but these loads  I have are configured in series in the charging circuits . That's really the best plan overall . when not charging the chargers are switched out . Kinda like at night and mornings mostly . Pulse motor circuits on the motors and regulators and ,force resonant coils on lights . Working in the 50 volt supply range but it varies .
 I can see an advantage to having a big bank on charge but it's not really necessary with loads in the charging circuits.

[GoVertical]

Hi, I am making progress on the hardware. Photos show CNC milled PCB's, buck converter and I still have to populate the controller. One controller can drive 4 bucks converters. A micro can be interfaced using 1 I/O pin.  The project really needs away to limit the input voltage. I am using a fuse until a better solution is found.

[definitionofis]

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
...
Regards
Thing

I looked at that interesting circuit. I think it can be more precise if it uses input current and voltage rather than just voltage
and then pegs a fixed volts/current ratio.

The reason I think that strategy is slightly better is because the output battery voltage varies 20% depending on what part of the charge cycle and the duty width has to change to compensate for that. Plus the impedance difference between the boost input and the generator can move the input voltage depending on how much current is drawn (Bob S mentioned that indirectly, too).  Using current and voltage, instead of voltage alone, can compensate for both of those problems.

I have not tested my boost converter sufficiently to know if it does a good analogue version of MPPT yet, and I only went up to 110 watts, so far.
I do prove that I can fix a V/I ratio using an analogue computer instead of a microcontroller. I think it will approximate a good maximum power point track. I will post another video when I have data.

I posted this video in the "notices/diary" section under the LTC3703 thread: http://www.youtube.com/watch?v=crSkmrQavXg
There are seven other videos in my youtube channel, if you are interested.

[GoVertical]

Greetings, I found a circuit that help limit the input voltage.
 http://www.pssurvival.com/PS/Lightning/Surge_Protections_Circuits_2009.pdf

Comments welcome.


I spent the day increasing the size of jumper pads. I was making to many solder shorts during assembly.
Big pads rock!

[definitionofis]

I did a similar crowbar circuit. I made it separate from my boost converter so I can locate it out in the turbine box instead of in my house.

So mine looks more complex because it is on the 3-phase side instead of after the rectifiers, like yours will be. It is the same idea, though.

Keep in mind your SCR has to handle all the current. It is a dead short of the turbine.

I did not test my circuit yet. You can tell I had a bunch of old SCRs in my junk drawer. I used a lot.

[DamonHD]

Hi,

I use this crowbar with my small 12V-nominal turbine:

http://www.earth.org.uk/img/turbine-crowbar-emergency-stop.ps

http://www.earth.org.uk/wind-power-pilot-autumn-2007-MotorWind.html

It fires (a little more often more often than I'd like still) when the LiPO4 battery it charges is full and balanced and the dump load is not keeping up.

Rgds

Damon

[definitionofis]

http://www.opencircuits.com/Projects

I found these two MPPT open source projects from that link:
http://www.freechargecontroller.org/  with a forum: http://groups.yahoo.com/group/charge-controller/

and

http://thesolarpowerexpert.com/a-diy-dual-voltage-converter-turning-the-solar-charge-controller-circuit-on-its-head/

[GoVertical]

Hi, project update. I installed IC's into controller board and applied power. The board functioned ok. Manual adjust of duty cycle was measured at the output of gate drivers. I have to fabricate more buck power modules. I should be able test with PMA soon. Cheers

[definitionofis]

How much will parts and PCboard cost to make one?
How many watts can one board handle?

Mine will cost $95 for parts for one and around $50, if I bought parts and boards for ten or more, because of the volume discount of cost. I included the six input rectifiers. Mine is controller and power module all on one board. I do not know my upper limit for watts for one board yet.

[GoVertical]

Hi, I am using 1 controller board that can control 4 Power Buck Modules. Each Power Buck Modules can safely handle 5 amps. The limiting factor is the inductor, it has a IDC of 5.8 amps. I have been searching for a inductor that can pass a high current at a low switching frequency. I have not concerned myself with cost at this time. I have been focusing on finding a circuit that will work and can be easily interface to a micro controller.  Once the circuit is working the real saving will be the cost of the PMA that can be used. You will be able to use smaller wire and magnets to fabricate the PMA.

[definitionofis]

I see the inductor problem. I used a high switching frequency, 100Khz, so I could wind my own air core inductor with 30 feet (10m) of #14 wire to handle 20 amps.  A low frequency would need a big inductor.

[GoVertical]

Hi, I tried making one but it is very impractical. It is the size of softball and is very heavy. Are you have any noticeable trouble operating at higher frequencies? 

[definitionofis]

Yes, I destroyed and swapped $250 of parts and PCBs, before I worked it out. The circuit board layout mattered a lot. I also had some feedback issue that changed with frequency and part values (which is typical of boost, not buck). I saw it in LTspice. It was very frustrating and I thought it would be, before I started. I learned a lot.

I revised the board and now it works above 60 watts.  I tested it up to 110 watts. I decided to wait for wind.  If it works with wind, I think I'll just enjoy it for a while, rather than push it. I want to see what the blade and generator can do, now. I still think the board can do 500 watts or 1000 watts or more, because as input voltage rises, not much boost, and eventually no boost, is needed at the high power end.

I don't have a power supply to do a proper test, over a broad range of voltages and currents. I need a vacation from it.

You can see my coil in any of my nine videos.  It does look like less wire weight than yours.

[definitionofis]

Correction: The current board can't do 1000 watts, or more. However, the design might be able to be grossed up, with different mosfets and diodes and not thin copper traces. I think this board can do 500-700 watts, and at most 1000 watts, if I add some copper wire.

[ChrisOlson]

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.

This is sort of an old thread, but Oz mentioned my setup with MPPT so I thought I'd chime in on it.

The reason the Classic makes me happy is because it does a lot more than just MPPT.  It's also a charge controller that does proper bulk/absorb/float of your battery bank with wind power - without using dump loads.  It's a data logger too.  It stores output data from your wind turbine for 380 days.  That alone, if you're into data logging, is worth $800 because you can spend way more than that on something like an APRS logger.  It comes ethernet ready and plugs right into your home computer network so you can interface with the controller with your PC and get real time monitoring of your turbine, as well as detailed desktop data logging.  It controls my water heating system with greater precision than anything else I've ever tried - properly charging the batteries as first priority.  Then when there's excess power available it uses it to heat water without affecting the proper care and health of the battery bank.  That again makes it worth $800 to me because my batteries are happy and they last.

So it's a lot more than just a simple MPPT controller.  Thing asked about what folks would like to see in an open source version - all the above for me.

The Open Source guys have for years designed and built stuff like this that the homebrew people can do themselves without spending the $800 on an expensive controller.  Take a look at Linux as an example in the computing world.  Linux isn't for everybody because it involves the nuts of bolts of a computer operating system, and so some people will still use Windows, or whatever.  But lots of folks use Linux (and it pretty much powers the server world), and it's free with the source code openly available for other folks to modify and contribute to the project.

So I think an Open Source MPPT controller would be really cool.  It doesn't have to be for everybody.  But it could be there for folks who like to "roll their own".
--
Chris

[GoVertical]

Hi, I am just working on the core circuit of the MPPT to start with. Once the Arduino is added it maybe possible to add the other features you mentioned.
My real reason for working on this project is to learn about the technology, it can be applied to many of systems like hydro, solar trackers, and steam.

Progress is slow. I still have to route connections. I included standoffs for plexy-glass safety shields.  Once a suitable inductor is found the foot print can be reduced.   

[definitionofis]

I searched this board and saw other buck/boost efforts over the years. I guess this is old hat.
So now goVertical has the buck version going on and I have the boost version going on. Maybe he will keep me interested.

I was going to add all the bulk/absorb/float features too. Progress is slow for me, because I have a very poor suburban location with no hope of real purpose. I got 4 watts for one or two minutes, a couple days ago. I am charging a 48v battery with an ECM 1HP furnace motor on an 8 foot diameter and 5 foot diameter, 4-blade turbine. Cut-in is 9.3 volts. I might sputter out and quit, now that I proved it works. I am just doing this as a learning experience, too.  I have a microprocessor option that could do a charging algorithm and data logging and computer interface. I have this: https://www.adafruit.com/products/296 to jumper over to my circuit.

I already have an Outback MPPT controller on my 10KwH of solar. Wind will be tiny compared to my battery and so I don't need the proper charging feature.

Regarding Linux, yes it is great. I dropped Windows7 and replaced it with http://www.ubuntu.com/ because windows7 update mangled itself three times, after two factory fresh restores. I got worried about the lack of security-updates. Linux has been more reliable with the auto-updates.

Movie editing and camera software was the only thing that had me stuck to Windows7 for a while. I did some crude video experiments about my solution switching to Linux movie editing and using a LogitechPro9000 camera. The description notes in these links are useful, if someone wants to switch:
logitech and guvcview recorder: http://www.youtube.com/watch?v=pHaZE45Rnlc
kdenlive editor: http://www.youtube.com/watch?v=4cK1sGvQaCo

I do everything in Linux now.

[Bruce S]

Regarding Linux, yes it is great. I dropped Windows7 and replaced it with http://www.ubuntu.com/ because windows7 update mangled itself three times, after two factory fresh restores. I got worried about the lack of security-updates. Linux has been more reliable with the auto-updates.

Movie editing and camera software was the only thing that had me stuck to Windows7 for a while. I did some crude video experiments about my solution switching to Linux movie editing and using a LogitechPro9000 camera. The description notes in these links are useful, if someone wants to switch:
logitech and guvcview recorder: http://www.youtube.com/watch?v=pHaZE45Rnlc
kdenlive editor: http://www.youtube.com/watch?v=4cK1sGvQaCo

I do everything in Linux now.

Welcome to the dark side  we have better cookies too  .
I'm a Slackware user myself. Redhat was what I learned on, but when they went public it went more main stream and got bloated.
Bruce S

[GoVertical]

Hi, I completed the connections and applied power to the system. I tested the switching signal at each power module. I should be able to test  the system with a PMA tomorrow. Cheers

[definitionofis]

Welcome to the dark side  we have better cookies too  .
I'm a Slackware user myself. Redhat was what I learned on, but when they went public it went more main stream and got bloated.
Bruce S

I started with Slackware in 1992 on 20+ floppy disks. I don't like ubuntu's clunky icon-disease. I put in a text menu ... lightboxxxxx or some name.
Should I switch back to slackware?

Windows7 came on a refurbished laptop and I got hooked for a while. I thought I would turn to stone. I survived.

goVertical
Do tell us how many amps you handled with multiple modules on the PMA tomorrow. I am following along.

[GoVertical]

Hi, tested system using hand crank on the stock F&P. Three out of the four Buck converters are working. With the duty cycle between 30 and 50 % the input voltage stayed at approximately 40 volts DC. The max output current was about 5 amps at about 12 volts. If the duty cycle was below 30% the input voltage quickly increased above the 400 volt input limit. I stopped spinning the hand crank and avoided damage to the circuit. As I added more buck converter the input voltage remained constant for the same RPM. 12 volt deep cell as test load.

I have to trace down the problem with the bad buck converter , fix the F&P to be driven by the drill press and include input voltage limiting circuit. At this time it appears the circuit will work. 

[GoVertical]

Hi, replaced the multi meter measuring the input voltage and the measured values are more realistic. At 20% duty cycle the max input voltage is about 100 volts DC under manual operation of the PMA.

[definitionofis]

Do you have a way to automatically set the duty cycle width, like that backshed link, which is earlier in the thread, is doing?

[GoVertical]

Hi, not at this time. After some tests, the next step is to add a micro with some current sensing feedback circuits. Before I can do that I have to have a better understand of the MPPT algorithm. This the reason I started with this circuit, I have to observe how the turbine will react to different wind conditions and how the circuit will react to different power at the input with different duty cycles. Hopefully others will offer a better circuit but until a better root circuit is found this will get me started.   

[boB]

What is your maximum input voltage now ?

Are you up to around or passed 50 volts yet, or did I missed it ?

boB