# DIY solar inverter project - synchronization with the grid

#### Henkdesteen

Joined Jan 19, 2018
5
Hi,

For a university project we have to build a solar inverter. A solar panel will be supplied to us, and the challenge is to design a circuit which converts the output of the solar panel to 20VAC which can then be fed into a 20VAC (not 220VAC because of safety) power grid. The main issue we are facing is the synchronization with the grid. We know that the frequency and phase of the power must be equal to the frequency and phase of the power grid, but what is the easiest way to synchronize a solar panel system with the grid? The only possible method we have found so far is the so called 'dark lamps method', but we have been unable to obtain a schematic and explanation for this method.

#### MrChips

Joined Oct 2, 2009
30,976
The synchronization method you have looked at is meant for mechanical generators which must be synchronized to the line frequency and phase.

With a solar panel, this problem is not so complicated. The output of the solar panel is DC.
You need to design the DC to AC inverter so that the inverter takes its synchronization trigger directly from the AC line frequency. You can do this entirely using discrete components or you can sample the line signal and then use microcontroller technology to trigger your inverter transistors.

#### Henkdesteen

Joined Jan 19, 2018
5
The synchronization method you have looked at is meant for mechanical generators which must be synchronized to the line frequency and phase.

With a solar panel, this problem is not so complicated. The output of the solar panel is DC.
You need to design the DC to AC inverter so that the inverter takes its synchronization trigger directly from the AC line frequency. You can do this entirely using discrete components or you can sample the line signal and then use microcontroller technology to trigger your inverter transistors.
Thank you for your quick reply. I have looked into the use of microcontroller technology for this purpose a bit, and it seems promising so far. For the conversion from DC to AC we will be using a circuit with mosfets, and we've already got that part figured out. Is there any specific circuitry you'd recommend if we were to use microcontroller technology to trigger the mosfets?

#### MrAl

Joined Jun 17, 2014
11,575
Hi,

For a university project we have to build a solar inverter. A solar panel will be supplied to us, and the challenge is to design a circuit which converts the output of the solar panel to 20VAC which can then be fed into a 20VAC (not 220VAC because of safety) power grid. The main issue we are facing is the synchronization with the grid. We know that the frequency and phase of the power must be equal to the frequency and phase of the power grid, but what is the easiest way to synchronize a solar panel system with the grid? The only possible method we have found so far is the so called 'dark lamps method', but we have been unable to obtain a schematic and explanation for this method.

Hello there,

Back in the 1980's we worked with Sandia Labs on solar grid tied inverters. One of the scientists there stated that the converters had to have a 10 or 15 degree lead. I argued that it should be zero degrees but the voltage must be higher than the line voltage for all points in time. We used the former for several years before switching to the zero degree idea after it was finally accepted. The idea is simply that the voltage should be greater than the line voltage and the phase the same, and after tie it switches into current mode. The line tie operation at that time was done with a rather large contactor.

There are different ways to get to this end though. The main idea is what method you use to generate a sine, and how you sync to the line phase. The sync is usually done with a phase locked loop. The loop is locked before tie and kept in phase after lock.
Error detection is of course a must too.

That was quite a while back though so there may be new ideas that came about some time after that.

#### crutschow

Joined Mar 14, 2008
34,692
If you want to synchronize using a PLL, a CD4046 IC should work.
The Type II digital phase detector in the chip has a 0 degree phase error at lock, which is independent of the signal's duty-cycle.

Last edited:

#### Henkdesteen

Joined Jan 19, 2018
5
If you want to synchronize using a PLL, a CD4046 IC should work.
The Type II digital phase detector in the chip has a 0 degree phase error at lock, which is independent of the signal's duty-cycle.
Thanks for the tip! We are most likely going to be using a PLL, and the CD4046 as you recommended seems like a good option. This seems like a good place to order it: http://nl.farnell.com/texas-instrum...-4mhz-4000cmos-16dip/dp/1106104?st=CD4046 PLL
So far, what I've understood is that the PLL will take an input signal (the 50Hz AC from the grid in this case) and produce an output signal with almost exactly the same frequency and phase as the input signal. However, how can it be made such that the frequency and phase of the AC power from the inverter are made equal to this output from the PLL?

#### Alec_t

Joined Sep 17, 2013
14,379
The inverter will need to have a sine-wave output of adjustable frequency. Analogue and digital solutions are possible. See post #2.

#### crutschow

Joined Mar 14, 2008
34,692
However, how can it be made such that the frequency and phase of the AC power from the inverter are made equal to this output from the PLL?
The inverter will have to be designed such that its frequency and phase can be controlled by the PLL.

Last edited:

#### MrAl

Joined Jun 17, 2014
11,575
Hello again,

I had forgotten that in the past (this is going back almost 40 years now) we had eliminated the phase locked loop.
I'd have to think back to remember exactly how we did this, but we went into instantaneous sine regulation also.
It was probably just like doing a linear regulator with switcher front end, where you just add a small voltage to the current voltage.