Monday, April 8, 2013

Designing a Grid Tie Inverter Circuit

A grid tie inverter works quite like a conventional inverter, however the power output from such inverter is fed and tied with the AC mains from the utility grid supply. As long as the mains AC supply is present, the inverter contributes its power to the existing grid mains supply, and stops the process when the grid supply fails.

The concept is indeed very intriguing as it allows each of us to become an utility power contributor. Imagine each house getting involved in this project to generate overwhelming amounts of power to the grid, which in turn provides a passive income source to the involved residences. Since the input is derived from the renewable sources, the income becomes absolutely free of cost.

Making a grid tie inverter at home is considered to be very difficult as the concept involves some strict criteria to be observed, not following may lead to hazardous situations.

The main few things that must be observed are:

The output from the inverter must be perfectly synchronized with the grid AC.

The output voltage amplitude and frequency as mentioned above must all correspond with the grid AC parameters.

The inverter should switch OFF instantly in case the grid voltage fails.

In this post I have tried to present a simple grid-tie inverter circuit which according to me takes care of all the above requirements and delivers the generated AC into the grid safely without creating any hazardous situations.

Lets try to understand the proposed design (exclusively developed by me) with the help of the following points:

 Again, as usual our best friend, the IC555 takes the center stage in the entire application. In fact only because of this IC the configuration could become apparently so very simple.

Referring to the circuit diagram, the IC1 and IC2 are basically wired up as a voltage synthesizer or in a more familiar terms a pulse position modulators.

A step down transformer TR1 is used here for supplying the required operating voltage to the IC circuit, and as well as for supplying the synchronization data to the IC, so that it can process the output in accordance with the grid parameters.

Pin#2 and pin#5 of the both the ICs are connected to the point after D1, and via T3 respectively, which provides the frequency count and amplitude data of the grid AC to the ICs respectively.

The above two information provided to the ICs prompts the ICs to modify their outputs at the respective pins in accordance with these information.

The result from the output translates this data into well optimized PWM voltage thats very much synchronized with the grid voltage.

IC1 is used for generating positive PWM, while IC2 produce negative PWMs, both work in tandem creating the required push pull effect over the mosfets.

The above voltages are fed to the respective mosfets, which effectively converts the above pattern into a high  current fluctuating DC across the involved step up transformer input winding.

The output of the transformer converts the input into a perfectly synchronized AC, compatible with the existing grid AC.

While connecting the TR2 output with the grid, connect a 100 watt bulb in series with one of the wires. If the bulb glows, means the ACs are out of phase, reverse the connections immediately and now the bulb should stop glowing ensuring proper synchronization of the ACs.

Assumed PWM Waveform (bottom trace) at the Outputs of the ICs

Parts List

All resistors = 2K2
C1 = 1000uF/25V
C2,C4 = 0.47uF
D1,D2 = 1N4007,
D3 = 10AMP,
IC1,2 = 555
TR1 = 0-12V, 100mA
T3 = BC547


After receiving a corrective suggestion from one of the readers of this blog Mr. Darren and some contemplation, it revealed that the above circuit had many flaws and it wouldnt actually work practically.

The revised design is shown below, which looks much better and a feasible idea.

Here a single IC 556 has been incorporated for creating the PWM pulses.
One half of the IC has been configured as the high frequency generator for feeding the other half IC which is rigged as a pulse width modulator.

The sample modulating frequency is derived from TR1 which provides the exact frequency data to the IC so that the PWM are perfectly dimensioned in accordance with the mains frequency.

The high frequency makes sure the output is able to chop the above modulation information to precision and provide the mosfets with an exact RMS equivalent of the grid mains.

Finally, the two transistors make sure that the mosfets never conduct together rather only one at a time, as per the mains 50 or 60 Hz oscillations.

Parts List

R1,R2,C1 = select to create around 1 kHz frequency

R3, R4,R5,R6 = 1K

C2 = 1nF

C3 = 100uF/25V

D1 = 10 amp diode

D2, D3, D4, D5 = 1N4007

T1, T2 = as per requirement

T3, T4 = BC547

IC1 = IC 556

TR1, TR2 = as suggested in the previous section design

The above circuit was analyzed by Mr. Selim and he found some interesting flaws in the circuit. The main flaw being the missing negative PWM pulses of the AC half cycles. The second fault was detected with the transistors which did not seem to isolate the switching of the two mosfets as per the fed 50 Hz rate.

The above idea was modified by Mr. Selim, here are the waveform details after the modifications. modifications:

Waveform Image:

CTRL is the 100 Hz signal after the rectifier, OUT is from PWM from both halve waves, Vgs are the gate voltages of the FETs, Vd is the pickup on the secondary winding, which in sync with CTRL/2.

Disregard the frequencies as they are incorrect due low sampling speeds (else it gets too slow on the ipad). At higher sampling freqs (20Mhz) the PWM looks quite impressing.

To fix the duty cycle to 50% at around 9kHz, I had to put a diode in.


For enabling the detection of the negative half cycles, the control input of the IC must be fed with both the half cycles of the AC, this can be achieved by employing a bridge rectifier configuration.

Heres how the finalyzed circuit should look according to me. The transistor base is now connected with a zener diode so that would hopefully enable the transistors to isolate the mosfet conduction such that they conduct alternately  in response to the 50 Hz pulses at the base T4.

Recent  Updates from Mr. Selim

Hello Swags,

I keep reading your blogs and continue experimenting on the breadboard.
I have tried the zener-diode approach (no-luck), CMOS gates and, much better, op-amps worked best. Ive got 90VAC out of 5VDC and 170VAC from 9VDC at 50Hz, I believe its in sync with the grid ( cant confirm as no oscilloscope). Btw the noise goes if you clamp it with a 0.15u cap. on the secondary coil.

As soon as I put a load on the secondary coil, its voltage drops to 0VAC with only a slight increase in input DC amps. The Mosfets dont even try to draw more amps. Perhaps some mosfet drivers like IR2113 (see below) could help?

Although in high spirits, I feel that PWM might not be as straight forward as hoped. It definitely is good to control torque on dc motors at low pwm freqs. However when the 50 Hz signal gets chopped at higher freq, it for some reason looses power or the PWMd mosfet cant deliver the needed high amps on the primary coil to keep the 220VAC going under load.

Ive found another schematic which is very closely related to yours, except PWM. You might have seen this one before.

The link is on http://www(dot)electro-tech-online(dot)com/alternative-energy/105324-grid-tie-inverter-schematic-2-0-a.html

The power handling circuit is an H drive with IGBTs (we could use mosfets instead). It looks like it can deliver the power across.

It looks complicated but actually is not too bad, what do you think? I will try to simulate the control circuit and let you how it looks.



Sent from my iPad