Shop OBEX P1 Docs P2 Docs Learn Events
PWM Solar Charge controller (Restarting old project) - Page 2 — Parallax Forums

PWM Solar Charge controller (Restarting old project)

2»

Comments

  • kwinnkwinn Posts: 8,697
    edited 2014-08-20 07:30
    Since I can't afford to buy a PCB, I am hoping to stay with through hole components. And.. Since I have over 200 Watts of 12V (22 to 24V output) solar panels coming in, I need to have a charge controller that can handle this and plenty more room for double or triple the input. Should I just bite the bullet and buy a ready made one or should I buy parts to build my own?

    Since it appears my mosfets will not be able to be driven with my current inventory, I will need to buy parts. If I buy parts, I might as well buy the right stuff the first time so I don't have to keep upgrading every time I get more panels. Is there an all around good mosfet and driver that I should look into that can handle the very minimum of 20A and can handle the 20 to 50 khz?

    I found this FET which looks like it can handle the voltages and current, but I am not sure if it can handle the frequency or what driver to use with it. Are all drivers good for any FET (Mosfets)?

    @jmg, the MCP14628 only handles 2A. The system I am working to set up would fry that instantly correct? (Hmmm. I think I see what this part is now. For some reason I was thinking this was the mosfet. This is a driver....)

    I think you would be better off biting the bullet and buying a solar charge controller. By the time you add up the parts cost, design/build costs, and the time involved you would not save any money doing it yourself. I suspect even a part time job at McD,s working the number of hours you would spend on the project would net you more than you would save by building one.
  • eagletalontimeagletalontim Posts: 1,399
    edited 2014-08-20 08:03
    I am not to worried about time as much as I am buying an entire controller already made and having to upgrade it later because I add several more panels and it can't handle it anymore. If I build my own, I get a great learning experience with parts I have never worked with before and can upgrade pieces as needed, not the whole thing. Sometimes I spend the extra time and money on things just to teach myself how it works so I can change or build another one later. I learn by hands on tinkering and in this case, I don't understand quite a few components so I reach out and ask for help. Researching usually leaves me confused because of all these math formulas and there are so many different controller schematics that use different components. I like to keep it simple. The least amount of components, the better I can follow what I am working with.
  • kwinnkwinn Posts: 8,697
    edited 2014-08-20 13:29
    If you want to be able to upgrade as you buy panels you will need to build a modular system. No single charge controller will work over a very wide range. There is no single controller that will handle a 20W, 200W, and 2000W panel even if the input voltage from the solar panel and output voltage to the battery was the same for all three. A more typical range might be from 150W to 250W.

    You also need to match the outputs of the solar panels and charge controllers to the batteries, and take into account the purpose of the entire system.

    Take a look at http://www.freechargecontroller.org/images/a/ab/Charge_controller_4_04b.pdf

    This is about the simplest charge controller I have found, and it could be modified to work with the propeller and most of the parts you currently have. Each module you add to your system would then consist of one 200W solar panel, one charge controller, and one battery.

    A single prop and two MCP3208's (or similar) could handle up to five such modules.
  • JonathanJonathan Posts: 1,023
    edited 2014-08-20 17:57
    I live off the grid. I spent a bit of time playing with building a MPPT controller. I got it all working until I started pushing 100 watts or so. Then stuff like gate traces and FET drivers started really mattering. Perfboard wasn't cutting it for me. I burnt up quite a bit of stuff, not to mention my fingers. My conclusion is you might get as far as 200 watts if you do it perfectly on a perf board but past that you need PCB's. Also, the best FET drivers seem to be coming out all in SMD so PCB's.

    I bought a Midnite Classic 150. Amazing bit of equipment. Really squeezes more on overcast and foggy days then my last one. I'm pumping 4000 watts through it. When you look through the specs noting can process as much power and few can accept as high an input voltages. I run mine at 100VDC input, keeps the wiring nice and cheap.

    Jonathan
  • eagletalontimeagletalontim Posts: 1,399
    edited 2014-08-20 20:16
    Ok, so let me get this straight... If I wanted to start at a lower voltage / amperage circuit, I can't use High rated components? If I found a FET that can handle 100A, but I am only getting 10A, that FET cannot be used?

    I found this : http://www.digikey.com/product-detail/en/IXFN180N25T/IXFN180N25T-ND/2126318 and that could handle pretty much anything I throw at it. The price is reasonable and has plenty of room for calculation errors on my part. Could I use this and when I get to upgrade my system, just swap out the inductor and change the PWM in the PROP to handle the additional amperage?
    (Edit : This is an N-Channel FET... Don't I need a P-Channel?)

    Also, If I get a 100A current sensor, would that cause my resolution to not be accurate enough for a lower amperage setup? Would I need a smaller one then changed it out when adding more panels?

    I know that is a bunch of questions, but please bear with me as I learn!
  • kwinnkwinn Posts: 8,697
    edited 2014-08-20 21:41
    ***WARNING*** At high currents and voltages there is a risk of serious injury or death, particularly with DC since it tends to make the muscles contract and lock in place. Perf board construction and point to point wiring is not adequate.

    If you designed and build it to operate at a higher voltage and/or 100A it would work at lower currents and voltages, but all the components used would have to be selected for the higher voltage and 100A operation. A 100A full scale would provide resolutions of 0.39A/bit with an 8 bit adc, 0.098A/bit with a 10 bit adc, and 0.024A/bit with a 12 bit adc, so resolution is not a problem.
  • jmgjmg Posts: 15,173
    edited 2014-08-20 22:17
    I found this : http://www.digikey.com/product-detail/en/IXFN180N25T/IXFN180N25T-ND/2126318 and that could handle pretty much anything I throw at it. The price is reasonable and has plenty of room for calculation errors on my part. Could I use this and when I get to upgrade my system, just swap out the inductor and change the PWM in the PROP to handle the additional amperage?

    You should read app notes on Solar battery chargers, like
    http://www.ti.com/lit/df/tidr393/tidr393.pdf
    and if you find one that matches your needs, build that.

    There are single-chip controller solutions like the LT8490, that have MPPT Battery charging.

    The FET you link to is just one (you would need at least 2), and the Rds ON is quite high, so you can expect 5.16W of loss at 20A just from i^2.R effects.
    As mentioned already 20A is into PCB design ( and thick copper) area.
  • eagletalontimeagletalontim Posts: 1,399
    edited 2014-08-21 17:34
    Having control over the program is kind of what I am shooting for so maybe I can add data logging on there so any changes I make can show me if there is any over all improvement or loss. This is why I would like to use the Prop as the controller. I just got in my first solar panel and it is much bigger than I expected! The next panel I have to build and I am still waiting on the cells to come in.

    Here is the specs of the first panel :
    Temp (C)                       25º      |        47º
    Pmax                         120W     |      87W
    Vpmax                         16.9V   |      15.2V
    Ipmax                          7.1A     |      5.74A
    voc                              21.5      |
    Isc                              7.45A    |
    

    I do understand about high currents on PCB's and was wondering how a small circuit board can handle 100A +. I have seen some that use a large "ground plane" as the power and leave the solder mask off then add more solder to the plane to handle the higher amperage.

    https://www.youtube.com/watch?v=D_8L7QbJUpY&list=PLjzGSu1yGFjWv4KeN-7TSYeQIcicM9Ghl&index=7

    I have watched all the video's of this guy's Arduino run MPPT controller and it is SIMPLE! Very few components and it appears to work but needs a little improvement for controlling battery voltage once charged. This is what I am interested in building to get started. The problem is, there are THOUSANDS of FET's and I have no idea what is a good one to get started with. I don't want to go too small and have to replace it instantly as soon as I add another panel or 2. Based on all the gathered data on here and from researching, the Rds ON is a big thing and so is the frequency the FET is driven at. Also a way to drive the FET is crucial at the speed I am wanting to run it. What just as not clicked for me is how to find a FET that will handle exactly what I am looking for and the proper driver that will connect to the prop via a pin or 2 and maybe a resistor. I have never build a circuit to drive a FET other than ON or OFF for long periods of time. Saturation or power dissipation was never a concern with those circuits. This is new to me and I want to learn, but just can't seem to comprehend :(
  • kwinnkwinn Posts: 8,697
    edited 2014-08-21 22:17
    For 100A you would need 6 gauge wire so the transistor would not be mounted on the pcb. Note the big lugs on the 100A mosfets you posted a link to. Those kinds of currents generally require copper bus bars and transistors on heat sinks.

    Yes, the arduino guy's setup is simple but it's only 20W and he is not trying to build an mppt controller to cover a wide range of power. He also has a small pcb that monitors the panel voltage and current so the arduino can control it.

    You may be able to cover a 4x range with a single circuit by starting with a single panel, then putting 2 panels in series, and finally adding a second pair of panels. For selecting a mosfet I look at the voltage, current, frequency, and power dissipation I expect for normal operation and select a part that has specs 1.5 to 2 times better. Then select a driver chip the same way.
  • MJBMJB Posts: 1,235
    edited 2014-08-22 04:21
    There are some VERY cheap buck modules on ebay that take up to 40V input and adjustable output 3 - 35V - they state in the 100W region.
    If you have a modular system you might use just several of them for each individual battery.
  • ErNaErNa Posts: 1,752
    edited 2014-08-22 05:40
    Ok, here is the schematic of the circuit I currently have. I think I found one design flaw just by drawing it out... The mosfet gate gets pulled up to whatever the input voltage is and I think that is an issue. I will have to pull it up to the 12V battery instead of the 24 + volts I plan on running into the circuit.

    I think you will find more ;-(, latest when testing. As it looks, you are not very experienced in power electronics, it is the best way to buy something that is functional and re-engineer it, so learning by observing, what happens. Then, make a copy and bring it to run the same way. That will give you a good starting point. Linear technologies, Fairchild, TI, they all have specific ICs to do the job and a lot of app-notes. What you try to reach is definitely off-limits for beginners. Uncontrolled current flow could be the source of the next wildfire ;-) Take care! Erna
  • kwinnkwinn Posts: 8,697
    edited 2014-08-22 07:15
    I agree 100% with what MJB and ErNa have posted. Having a panel charge a single battery avoids a lot of the difficulties involved in making sure that all the batteries in series/parallel arrays are charged equally.
  • eagletalontimeagletalontim Posts: 1,399
    edited 2014-08-22 17:23
    My question is... If I buy a premade controller, lets say from Ebay (Yes, I am that cheap), that is able to handle 200W but my solar panels output 600W. Can I use 3 batteries and 3 charge controllers for the solar panels that would be wired in series?

    I do admit I am not highly familiar with anything 12V and over 3A since it begins to get in the range of blowing many components with one fault. Since I can't afford to buy 3 charge controllers, I was hoping one would do the trick if I can get enough information to build my own even if the efficiency is not 90+%. At least I have it started and somewhat working. As I learn, I can improve. That is my goal. Buying parts that are HIGHLY over the rating of what I am working with give me a little room for error I would think... maybe I am wrong. If I buy the right components at first, I will save money in the long run. There are THOUSANDS of different parts that are all quite close to the same specs, but I just don't know what would be best for me.

    What I am searching for is a mosfet and gate driver setup that can be driven by the Prop and handle a minimum of 30A to start with. Since my panel can output 7+ Amps, and I plan on adding another one soon since I have already ordered the parts, I need to build this charge controller to at very minimum handle 14A and 40V (Highest I think the panels would output in series). I know running electronic components at their max causes heat and early failure. I want wiggle room.....

    If I purchase the recommended parts by whoever is willing to help me with this, I can at least have my hands on training started. Reading information about how they work is like bench racing.....One person may have the exact same setup but has different actual results when put the the test in the real world. In this case, the FET and driver are the biggest components that I need help with. Later comes the inductor and then comes the Amperage and Voltage readings through the processor instead of an external meter.

    Right now, I plan on using 1 car battery since that is what I have. I will be purchasing some deep cycle batteries soon. Probably one to start off with since all I want to run is my shop lights for about 1 to 2 hours per day. My hacked UPS can handle the draw with no problem. I have already test this.
  • eagletalontimeagletalontim Posts: 1,399
    edited 2014-08-22 22:27
    Ok, I have a feeling this thread is going nowhere and I have not figured out anything except circuits made to handle less than 20A and are not Prop driven which is not my final goal.

    To keep this simple, here are a few simple questions...

    What FET / IGBT can I use that can handle no less than 30A and can handle the incoming voltage of several series solar panels that can switch at no less than 10khz?
    - I like the Voltage and Amperage abilities of these, but have no idea if they would work or how to drive them at the desired speed : http://www.digikey.com/product-search/en/discrete-semiconductor-products/igbts-modules/1377479?k=IGBT
    - This one is the least inexpensive from the search and it looks like it can handle quite a bit (Less likely to have to upgrade later) : http://www.digikey.com/product-detail/en/IXDN55N120D1/IXDN55N120D1-ND/1651833

    What driver would be best for the above IGBT that can be controlled by the Prop?
    - From researching, optical couplers seem to do the trick, but I am worried about frequency abilities. : http://www.digikey.com/product-search/en/isolators/isolators-gate-drivers/3736886?k=IGBT
    - Is there a better driver for this?

    There is a debate about using a Diode from ground to the Inductor VS a Mosfet which I am unsure the purpose of using a Mosfet. Is it controlled via a driver with 2 outputs that knows to turn one mosfet / IGBT on and the other off at a specific time?

    I will ask a few more questions once I understand the above questions.
  • ErNaErNa Posts: 1,752
    edited 2014-08-23 01:19
    ok, to go to somewhere: the questions you ask are mainly not related to the power you want to convert, so it is a good idea not to go straight forward, if you are at the foot of a cliff. From experience I can tell, most circuits you will find in the hobby area are not well engineered and even the app-notes from the manufacturers often only show the directions or show advantages over the competition, but not the obstacles all face. The problem not reaching 99% efficiency from 98% is not the 1 % energy you are missing, but the 1 or two watt of dissipated power heat up the circuit, so reduce reliability, what is an issue. ... In a power fets data sheet you often see test circuits and it is a good idea to set up this, run the tests with different drivers and look to actual switching behavior. But to do this you need equipment that is much more worth than the whole thing you want to do. Thats my advice for free ;-)
  • eagletalontimeagletalontim Posts: 1,399
    edited 2014-08-23 07:10
    Hmmm. So basically, I don't own the right equipment to build a MPPT charge controller? I know I don't have the knowledge to jump right in and build one or I would not be here asking questions. In a previous post, I was told to look into IGBT's which I did and I am curious if that would be the best route for me to go and if so, what would be a good one to choose. Based on all the circuits I see online, they all handle < 20A and have a max voltage of 30 to 40V. The IGBT's handle the voltage and amperage I am hoping to eventually get to, but now I am told to just buy random parts and test each one's switching behavior until I either run out of money or find one that happens to work? This is what I was trying to prevent by asking questions.

    Right now, I have a useless solar panel I just bought since I can't simply hook it directly to my battery. I have a bunch more solar cells coming in that are also going to be useless because I can't afford a $1,000 charge controller and building a basic MPPT controller is impossible without having gone to college. I am not looking to hit 90% + efficiency on my first try. I know this is not going to happen. I just want to get the basic circuit built that can handle the MPPT function and improve from there. I am sure that if I get pointed in the right direction with the correct parts, I will at least succeed in building the basic circuit even if it is only 60% efficient.... at least it works and I don't have useless panels laying around.
  • JonathanJonathan Posts: 1,023
    edited 2014-08-23 08:15
    Tim,

    I hear ya! I'm in about the same spot. Or was, anyway. I wanted to build my own but ended up knuckling under. So where to go from here?

    1) Someone mentioned 100 watt buck converters pre-made. Gang up a couple.
    2) Get a non-mppt controller for less than US$100. It will make a little less than an MPPT controller, but not THAT much and it gets you up and running. Cheap and easy.
    3) If you really want to build yer own, there was a link earlier that had a schematic for the driver portion of an MPPT. With that, get a CAD program and get some PCBs made. Once it works, I would see about upping your panel voltage. That way you can keep the amps low and process more power. Yeah, higher voltage is more dangerous etc.

    Jonathan

    <EDIT> One last option, one that I mentioned before. Save yer pennies and get a Midnite charge controller. Expensive, yes. But more than just a superb charge controller, it can be used with windmills, run extra loads when yer batts are full and so on. I love mine.
  • eagletalontimeagletalontim Posts: 1,399
    edited 2014-08-23 08:50
    The problem is, anything over a $100 expense in one purchase is a HUGH EXPENSE! I am trying to piece together parts in < $100 increments between paychecks (and not try to save for 8 or 9 paychecks or 4 months to save for one premade controller) so I can learn and experiment with my own controller. I already use Diptrace as my CAD and use either Itead or OSHPark as my PCB manufacture. If I have to build a PCB, so be it, but it still turns out that I should just give up the entire idea of learning while building my own MPPT charger. Instead of me trying to be a DIY'er and learning on my own design with help of others (even though someone else has done it already), I am wasting my time here.

    I do understand that higher voltages can cause big issues including arcing at distances and many other things. I know that high voltage can be used to make low voltage, high current which is what I am attempting to accomplish. Knowing and doing are 2 completely different things which is where this thread has gone. I know it can be done, but all I have gotten is links of where others did it for "LOW VOLTAGE" and / or "LOW AMPERAGE".

    I give up.
  • JonathanJonathan Posts: 1,023
    edited 2014-08-23 14:07
    Tim,

    The reason that you got links for low voltage and amperage is that high power switching isn't easy. I'm sorry you are frustrated but if it was easy there would be DIY designs out there. Parallax once hosted an MPPT challenge that died due to lack of designs. And ganging up cheap smaller watt designs can have issues too, especially if you want to charge batteries optimally.

    If you really want to DIY, I would find a design, get some boards made, get them to work at low power and then see what you can do to jack it up. That won't be cheap in the end though. Entertaining however... :-)

    Jonathan
  • ErNaErNa Posts: 1,752
    edited 2014-08-24 07:09
    http://www.linear.com/product/LT8705 http://www.linear.com/purchase/LT8705
    This seems to be a product, that fit's your needs. The chip is less than 10$ and there also is a lot of documentation, even a demo board designed for 5 Amps, fully documented.
    One good thing about mankind is: we alway try to do things, we never did and that were never done. We also try to reach goals, that can not be reached. Ambition is coded in the gens. So if your goal is to get a feeling about renewables, what can be done and what not, you are on the right track. But I understood you intend to do something economically feasible, and so experienced people raised warning flags.
    Every $ spend in research of solar power is better invested than drilling for shale gas, that I believe. For once these problems are solved, this sustains. ;-)
  • eagletalontimeagletalontim Posts: 1,399
    edited 2014-08-24 09:37
    I have stopped trying to build my own and will just buy a pre-made one. There is too much room for error and I don't have the knowledge to figure this out on my own and.....I am extremely impatient if you can't tell already :)

    One thing I think I can do is a Solar tracker which I may be able to do on my own. Just got to get some metal to build the frame and some all thread to attach to the panels.
  • ErNaErNa Posts: 1,752
    edited 2014-08-24 12:01
    Perfect! You are on the right track now. They exist, but are not mass produced, so a lot of value is added, even in single pieces. And: it merges mechanical, servo, sensor, software and even data sampling, storage and evaluation. And the propeller can do these jobs in real time. Could be the start of another successful project.
  • MJBMJB Posts: 1,235
    edited 2014-08-29 15:43
    just found this -
    might be to late for this project
    but it is never to late to learn

    https://www.youtube.com/playlist?list=PLjzGSu1yGFjWv4KeN-7TSYeQIcicM9Ghl


Sign In or Register to comment.