Grid-tie inverter concepts may appear to be complex due to the many criticalities involved with them, however with some intelligent thinking it could be actually implemented using primitive technologies. One of the ideas has been explored here.
Introduction
The discussed idea of a simple grid-tie inverter circuit was suggested by one of the interested readers of this blog, Mr. RTO.
The images sent by him are shown below. In the first image we find the simple circuit diagram comprising a step down transformer for translating the grid data, a mosfet triggering circuit which accepts the grid data and a corresponding inverter transformer which is used to amplify the DC conversion of the grid data from the mosfet network.
A Smart Looking GTI Circuit
The idea looks pretty simple, and indeed very smart:
The left side step down transformer feeds the half wave rectified voltage to the corresponding mosfets which begin conducting in-sync with the grid input and convert the DC source into a corresponding AC across the inverter transformer at the right hand side. The output from the inverter transformer which is now a grid synchronized AC feeds the grid with the intended GTI results.
The idea has been tested by Mr RTO, but he complains about lower efficiency from the unit.
This could be because of one major issue in the design, that is the absence of a "neutral" wire across the output of the inverter transformer.
With the shown set-up, the output would respond with a push-pull action across the secondary of the right hand transformer, meaning both the ends would become "HOT" or "LIVE" alternately during the operations.
The grid will take this as a "short" for every inverted half cycle from the transformer, because the grid voltage always has one wire as the neutral which is never a "LIVE" terminal.
We don't want this to happen.
Using a Center Transformer
A simple solution is to use a center tap winding for the secondary of the inverter transformer. This would render the center as the "dead" or "neutral" wire relative to the outer taps of the trafo. The upper tap may be configured with the grid while the lower tap to a balancing load or more effectively fed back to the primary side for charging the battery or reinforcing the DC source itself.
Warning: The author cannot be held responsible for the results of the experiment. Please do it at your own risk!! The projects explained here are recommended only for the experts in the field of electronics.
The test set-up of the above design can be witnessed here:
Another issue which could remotely transpire is the conduction from the mosfet which wouldn't be exponential, rather an "awkward" and unrecognizable sinewave.
The mosfets could be replaced with SCRs, as shown below. This would allow a perfect sine wave to be induced across the inverter transformer and the grid.
Using SCRs for the GTI
A much improved grid-tie inverter circuit using the above concept and SCRs is shown below. The idea looks greatly simplified, and quite impressive.
The output of the right and transformer could be seen converted to a center tap topology, wherein one half winding is integrated with the grid, while the other half is subjected to a balancing load so that the center tap is appropriately conditioned to be the neutral for the system.
The balancing load could be replaced with a charger circuit for charging the inverter battery itself, this would reinforce the input with additional power and more backup time.
SCRs will not Latch
At first glance it appears that the SCRs would get latched since a DC is being used across its anode/cathode, however according to me it won't happen, because the gate of the SCR is subjected with an alternately reversing AC which would prevent the SCR from getting latched every time the gate AC feed changes its polarity
Questions & Answers
I think IGBT is much better on RTO design..i already seen that one other blog not from him some french guy made that one but he complaint became very hot in long run…suggestion made an igbt..which solve the issue..some also made using ir2113.., but this simple one i really like its work…actually but im wondering if ur using battery is not feeding while if your are using solar or dynamo it is feeding?
yes that's correct, IGBTs are much rugged than mosfets and better protected against avalanches and parasitic voltages.
Barbe, the DC source can be from any source, but the issue here is that the generated output from the inverter transformer is not tied-up with the grid voltage…..here the inverter voltage could rise or drop independently causing imbalance and heating up of the devices.
I'm using this circuit about 2 weeks. The temperature was fine using that tiny heat sink.
The other problem is when the grid fail the output voltage will increasing until the mosfet damage. To fix this problem, I use your Low/High voltage cut-off
1.bp.blogspot.com/-oZj7uR4dRu0/Txl54N6s01I/AAAAAAAAAvc/toKqbbIGZpw/s1600/OVER+VOLTAGE+UNDER+VOLTAGE.png
The relay cut the line from input transformer to output.
I dont know mate im generating 356vac/90hz..and tune it with 2uf and become clean square wave..and plug it to the circuit…nothing happen…its start feeding into the grid..i dont understand but the more i generate watts the more backward i get…for my experenmt…
RTO, that's strange how can the voltage increase, because if mains fails, the mosfets wouldn't receive any gate feed and would stop conducting instantly, there could some other issue in the circuit which needs to be checked.
Barbe, the voltage must be tied-up with the grid voltage level, otherwise the inverter may be subjected to unnecessary stress, because allowing more voltage into the grid would cause more load on the mosfets, while lower voltage wouldn't add anything into the grid, so the rise and fall must be synchronized with the grid voltage
Swag, I tried this circuit as normal inverter and the output voltage is 300vac. When this output transformer connected to the primary input transformer the secondary input transformer will also increased and feed to mosfet gate. The more voltage on gate the more voltage on output transformer. Then from output feed again the input transformer again and again until the secondary input transformer reached to 21volt the mosfet get damage.
Yeah mate, but its work…i made the wattmeter move backward and no issue….!all temperature is perfect…noise is perfect aswell…thats the work of thr control transfo…automatic when it tied up…its align with the frequency…i also have voltage on neutral…but it just because of the harmonic…i get it all and understand by scope…
If no grid the grid tie will auto switch off to prevent from damaging the igbt…i will try to put some video on reversing wattmeter mate
that's great Barbe! thanks for updating this info!
RTO, we have 12V zeners clamped at the gates of the mosfets, so how could the voltage increase at the gates??
Sorry, the 12v zeners were not yet clamped during that burst. I'm newbie of zener, I just read your conversation with barbe about damaging of 4017 and 555. Then I applied that on my circuit. I just add that low/high cut-off just for double protection and system will automatically shut off. Without it, the system still have output even without grid.
OK, I thought you were referring to the first circuit in the above article…anyway zeners are crucial in either of the designs, at their respective positions.
Regarding on the gate issue u need atleast 1 to 5 ohms 50watts resistor on your input from battery to limit the current..to save you from damaging the gate…!!! U can check it with the scope..as wellit will help alot
not for the gates, rather for the drain and source, because gates are voltage dependent not current dependent, so gates won't be affected by high currents.
to protect the gates you'll need make sure the voltage does not exceed 15V, therefore zeners are required for te same
I mean the igbt…and correct u need zener diode atleast 1 watt…i use 15 volts with one kilo ohm in seires.
What should be the amp of output transformer for 300 watts?
It needs to be matched with the DC source wattage,
more wattage would contribute more to the grid and vice versa.
Hi rto…i think atleast you need 20amps transfo 6 0 6 primary and feed aprox 18 to 24 volts and your zener diode make it 15v make it two 6 0 6 make it in parallel it will make it more wattage…more spin aswell but make sure you have good input aswell…also try to put SSR on your output 220 and scr on your input..it will help make your grid tie..always in grid.. But your design is seems perfect…no need pwm…the only thing i dont like in this design is the heavy transfo while on ic grid tie small transfo but good in generating watts.
Also put 100uf on you dc source it will help on pp voltage to much with your control transfo switch
He need a frequency control on the output..of his grid tie…i seen this one is actually working…just make sure that the polarity of the transfo is correct..! I already build one of that..its perfect…regarding the scr design mate it blows…:( i dont know why…i think the issue is with the gate not really sure..it seems short circuit or something not sure i use small transfo and small scr..
Hi Swagatam Majumdar
a good idea.i like all your design.thank's give me knowlege
thanks khang
Hi mate i attached the working gti…the wattmeter actually move backward but hence i think there is some kind of lock i dont know after one round it just stop and not moving…
s63.photobucket.com/user/barbe10/media/imagejpg1_zpsc965bfce.jpg.html?filters[user]=139832643&filters[recent]=1&sort=1&o=0
that's interesting mate, I'll research more about watt meters and let you know how these are specified to work.
I found the lock mate…actuallt is not the lock…i need to make sure that the meter is in 90degrees otherwise…it will lock or it will not move properly…
I attached my frequency monitoring circuit…the link mate
s63.photobucket.com/user/barbe10/media/imagejpg1_zpsef64bd1b.jpg.html?filters[user]=139832643&filters[recent]=1&sort=1&o=0
frequency monitoring? for what purpose mate?
It just optional…if you are in 110 110 output…bu can be also used in this one..o mak sure youre always 50hz…anyway no need..as i will buy just normal hz meter.that dis sign is one of my friend monitoring circuit for defferent purpos..anyway…
The most mportant is the auto switch of to prevent the igbt from damaging..
But so far all is perfect for 4days now..i will start feeding via solar now…but at the moment im using battery…so far both wattmeter my test wattmeter and our general kilowattmeter is spinning backward when no load..with load its to slow..actually
Hi barbe, can you provide the efficiency?
Mine, I only got 92watts (40-50% efficiency) from 200watts solar panel.
Im not using it through solar though..but using battry and some math..not really sure is running 80 to 90 percent rto.
WoW!! 80% efficiency is like GTI bridge mode mosfet with mppt 🙂
Can you attach the circuit diagram?
Yes correct bridge type..im using…!!..the only thing u may on this is that is to heavy becoz of too many transfo…but nterms of efficiency i thnk u will definitly happy.
Im just waiting for the meter from my friend to double check but it will not less than 80percent…
I thnk one issue on your scr design is the diode going to transfo..when you plug in it to the grid the reverse current will flow back to the battery result to damaged the batt..which just happen to me the water came out from it…and also the gate when activated not really sure but it create short circuit may be mate not sure but my scr just wasted…!
Sorry mate..do you have any suggestion…on this one…
No mate, the battery or any DC source will never be at risk from the reverse current, because you will see such diodes included in all types of circuits whenever inductors are used…in order to protect the switching devices.
However for scrs this may be not true, they are never at threat from back EMFs, so anyway the diodes can be removed.
The scrs must be rated to handle the transformer current, I think BT151 is not the correct option, I'll replace it with the appropriate ones soon.
Ah ok mate…it just blow..ok mate we will try that one aswell…by the way whre is our welding machine
mate, I am looking for a 12V 100amp smps circuit online, because it would be difficult for me to design, as soon as i get it will post it for you. it may be then applied for welding purpose
No prob..mate for the welding machine im starting collecting parts..i will try build inverter thing kind of welding machine but im goin to use round ferite nothng to found in dubai i buy online 2dollars so i have to wait for 2weeks
that's nice mate, if you succeed do provide me with the results and the schematic….
Hi Swagatam,
can u make a circuit witch after the battery is fully charge the solar panel will switch to the GTI for not wasting the solar panel power.
thanks in advance,
Sure mate…youre my mentor..and prof. Definitely i will…!!! Also im not sure if this gona work perfect…but im sure it will be minor adjustment on amps…i think my issue will be caps or my transfo computation….anyway let yoy know mte as alwys…your the boss here mate…
my pleasure mate!
For the gti..i would suggest this igbt. Gt30j324 or gt30j322 im sure with this…!!! Take cage with the gate,,,, as always….make sure bout the current flow…like i said a resistor needed..for unstable current…maybe 1 ohms 50 watts will handle….! Series to the positive of the battery or solar…
But my design will activated only when connected to grid…! So at the moment im on trial on switching via relay…to amke more my gti in safe mode…and after one ahour of used my big transfo become hot…but all the igbt in very cool… I thnk i need to search some good ventalation for my transfo…
IGBT gate is voltage dependent just like mosfets.
mate the transformer too should not become hot..I suggested the use of a center tap configuration at the output, did you do it?
before carrying out critical and dangerous projects like this one we should be very sound with the required info and knowledge, especially the basics, otherwise it could lead to hazardous situations….
Mate not too hot…is actually normal….kind of hot…but im thinking im going to use this one 24hrs..im sure..it will be hot…so need some air on the transfo same on the gti i both online…
Im not using center top…i found some dificulties with center top…when i look into scope it is hard to align the movement…i dont really understand…coz i really not focus to trouble shoot that thing…so i made it no center top much easy…and perfect..
Anyway i will look that one and try but so far without center top i found the waveform perfect…and no noise or interfernce…very nice square wave also compare with the center top the square wave is not smooth..and frequency somtimes become 600 to 10 khz while without center top..stable at 90hz or 80 hz juz put the caps boom 49 to 51hz..clean squarewave and when u put to the grid automatic very neat sinewave…compare again with the center top…you need to clean the power factor…mate u will not be happy if u see to mch scratch on sine wave..and after 5mins your igbt become too hot and transfo…
Hi Swag, this is my first time using SCR. According to my research, once the scr triggered it will always ON and the scr will turn off if we disconnect the source voltage. On the circuit diagram above, I can't see how scr goes OFF.
Hi RTO,
you are correct, but there's another way of breaking the latch, it's by providing a negative voltage to tge gates, that's what exactly would happen for tge gates of the SCRs shown above.
With every alternating voltage, the transformer taps will change polarity, supplying a negative voltage to tge ScR gates…according to my theory.
It will using optocoupler..it will switch and off..im sure..using 555 to be simple or if you want some noise used relay to switch it off and on very fast…i see that one in action actually its cool if u just want some fun on your gti used relay..but optocupler for the scr is the best..
Hi mate actually..i dont see yet the scr works..just have doubt on it…because i check the scr on the net i dont see a very fast switching which we need on gti…like what happen to me my scr juz blow mate i used high amps..aswell 500/60amps…i waste my 10 dollars mate
Not sure mate..but my friend here suggest it will possibly working on his theory if we used triac…
Thanks Swag, this information will help to me.
Btw, how about the gates voltage, if I will give it 2volts vs 5volts. What would be the result? It is the same with mosfet that the more voltage on gates will get less resistance on Drain and Source?