Grid Tie Inverter Schematic

[Hero999]

Plase read this thread from the beginning: the grid-tie inverter needs to be approved by the power compnay before you can use it.

I imagine that you could in theory use an existing inverter to scavenge parts from to build a gruid tie but it isn't possible to simply convert it to a grid tie inverter.

 

[abdulla]

thank you very muth

 

[Jules_Theone]

I am thinking of designing and building a grid-tie inverter for my final year project at uni. I've done some research into grid-tie inverters and stand-alone inverters (dismantled some stand-alone ones) and found the system seems quite simple but needs quite a lot of feedback / control because it could go horribly wrong pretty quickly. As far as i can see, the circuit consists of a step-up flyback circuit to turn 12v DC into 350v DC, then this is fed into a FET H-bridge which is controlled from a slightly higher than 50Hz oscillator. As the frequency is slightly more than the grid, the circuit will always try to lead the grid, but as the grid is essentially infinite current , this circuit is unable to change the frequency.
The main problems I am thinking is; failure modes (grid disconnect, anti-islanding, voltage or current spikes and large frequency changes) and the waveform I am introducing to the grid (modi-sine).

Is this still undoable as I attempted this project at home, then last year at college but didn't get a working prototype - lots of smoke and big flashes.

As Hero999 mentioned, the circuit would perform as a constant current source instead of a constant voltage source. What exactly would this entail (what's the difference between the two).

 

[Nigel Goodwin]

I'm sure people here would be very pleased if you post your completed design when you finish it

 

[Hero999]

As Hero999 mentioned, the circuit would perform as a constant current source instead of a constant voltage source. What exactly would this entail (what's the difference between the two).

You sounded very knowledgeable until this point.

A constant current source tries to keep the current the same regardless of the load and a constant voltage source tries to keep the voltage the same regardless of the load.

A grid tie inverter will inject a constant current into the grid.

A standalone inverter will supply the load with a constant voltage regardless of the current.

 

[Jules_Theone]

Thanks Hero999, that clarifies something. I wiki'd constant current source after writing that, which obviously explains it more! After a little more investigation I have determined there are two main ways to step up the voltage (I already sorta knew); high frequency flyback circuit like PC SMPS and most high efficiency power supplies nowadays and low voltage into modified sine, harmonic removal etc feeds into power IGBT bridge or similar, which then drives a large low frequency transformer. The cheap (and nasty) in-car inverters are high frequency, small units while most UPS's seem to be big, bulky transformer style. It seems to me that the latter is the way to go (low voltage but higher amp switching) because I am less likely to get electricuted, it seems more simple (not dealing with saturated core ferrite transformers) and also, I believe that UPS's synchronise to the grid's phase so when the power goes out, it keeps feeding without any phase shifting, then eases back into synchronising when the power comes back up. The UPS circuitry is a lot more complicated but I think I can decypher the pointless bits and strip most of the redundant stuff out...

So on the constant current route - I have a switching circuit I've used before but its constant voltage (like most power supplies). Can anyone see a problem with the way i have turned the constant voltage circuit to a current circuit attached?

 

[Jules_Theone]

From my initial investigations into existing stand-alone inverters, I have found most of them use a TL494 and some PC PSUs I've dismantled use the same chip too. I've found the datasheet for this chip and some limited design notes. Does anyone have and suggestions for this chip or a better alternative?

I think the way to grid-tie is generate 350vdc from 12v, modulate the high voltage to modified sine at just over 50Hz, filter to produce purer sine and add to the grid. It's obviously more complicated than this so I've broken it down into smaller parts.

Does that sound right or am I missing something?

 

[Hero999]

That's modified sinewave inverter plus filter. It will work but the filter will need to be pretty bulky to work effectively. If you use high frequency PWM then a smaller filter can be used to remove the harmonics.

 

[fernando_g]

What's the input voltage?

If it's a DC source another cheap and inefficient way is to use a motor generator set. This is basically the same idea except a DC motor drives the induction motor/alternator. Again, you need to shut it down when the shaft speed drops below the synchronous speed.

.

I have experimented wit the above approach and it works. However, as noted it was inefficient. In the whereabouts of 60% if my mind serves me correctly. Of course we were using a very small motor set, 1/6 Hp. Electric motors become vastly more efficient with larger sizes. But at the time that's all we had and we ran out of time and money, thus the whole experiment was just that, an experiment.

The key here is to run the AC motor at reverse slip equal to its plate slip. For instance if you have a 4 pole motor at 60 hz, its synchronous speed is 1800 RPM. But if the plate lists its rated speed of 1750 RPM, then to operate as a generator, you would have to drive it at 1850 RPM. Accurate measurement and control of speed is a must, generator output drops quite rapidly above the slip speed.

 

[mvocray]

Hi! I got interested as I'm thinking of building one myself, otherwise I won't graduate next semester. LoL. My idea is to build an H-bridge with the high side switches connecting at 60 Hz (it's 60 Hz here) and the low side ones running at high freq SPWM as Hero suggested.

The current harmonics will be filtered out by a low value sense inductor in series with the line. The focus of my project is proper timing of the control switches.

The problems I've encountered with the design so far are
1. getting proper line voltage feedback (I believe an unknown value lag would come from feeding back using a transformer)
2. I'm thinking of using FET switches with snubbers. With a DC current source feeding current into the H-bridge, overlap times would be necessary otherwise, there would be no path for the current, right? Following this argument, how will the SPWM scheme (requires off times) work with a current source?

As far as I've heard, most applications nowadays use MC. I would want to make one using only analog components though.

Also, this is my first time making the circuit, so any help/criticism of the work will do. Thanks.

PS. First time joining a forum. How do I post my schematics? LoL

 

[tcmtech]

Hi everyone! been reading up on your GTI unit problem. You are a bit off! I felt its time I jump in here before someone spends to much time and money only toget dissapointed. And yes I know my spelling can be way off! I am an electrical enginner not an english major.

GTI sytems are easier to build than a regular inverter! I designed one about 12 years ago just to prove someone wrong! I was 22 at the time.
I have several small units that are currently working now that are in the 300 to 900 watt range. they are running off a wind generator I built a few years ago.
I am currently working on a do it yourself CD that I plan to sell on ebay. Hopefully sometime this comming month or two.
I will cover the simplest design it has two transformers and two transistors and two resistors, thats all! The larger units designs will go into the actual opperating theories and the control and saftey interlock circuit designs. I will cover at least 4 and possibly 5 styles from 100 watt to my big 15kw. All will be photographed and documented through the entire build process. The big units have a PLC based control system for the saftey and control part. I will even include the software and basic program with the CD.
Video, lots of pics, true to life parts lists and actual places to get all of those parts in print! Plus trouble shooting tips and maybe a bloopers file!
If you check out ebay you will find they sell ready made units in the 250 watt rating. Plug in and go and yes they are legal! Way expensive for what they do Around $300 for a 250 watt. I can build one in the 1500w Plus range for that!
I have already talked with my local power company and they are happy to work with me on setting up a large unit this summer. 5 to 15 KW capacity! Even if its privately owned and constructed. There engineering dept has to check it out and sign off as its safe to work on thier system Then it should be legal!

So can I answer some questions now? Or are all of you going to snub me like my engineering professor did when he could not make a GTI that worked. And I brought him one I built out of an old battery charger and some misc power supply parts. I Got the blind eye deaf ear treatment. he couldnt do it on a $25000 research grant, I did it with dumpster divable components! ( stuff most people toss in the garbage)

I am happy to help any way I can! Thank you!

 

[Nigel Goodwin]

PS. First time joining a forum. How do I post my schematics? LoL

Click on 'Go Advanced', this provides you a 'Manage Attachments' option, which you then click to upload your file.

 

[Nigel Goodwin]

I am happy to help any way I can! Thank you!

Then post a circuit for it - rather than just an advert for your Ebay sales.

 

[Jules_Theone]

The simplest design has two transformers and two transistors and two resistors, thats all!

Is the circuit i have drawn up quickly what you mean as simple?

I have tried this circuit (slightly more complicated) and it gives a large amount of harmonics on the power line, which is unacceptable to the power company. Can you please post a schematic (or partial schematic) of your circuit?

 

[tcmtech]

Lookes like Jules Theones got it! one addition and you have the very basic GTI. you need a diode on the power output transformer. It goes in line with the input. this will keep the circuit from trying to use the switches as rectifiers and burn them out.
This circuit is horrible for switching harmonics and spikes the fets at every turn off point. the higher the load the higher the spike. I have also found this circuit will top out usualy at about 50% return efficiency give or take.
Heres the weak points of this circuit. It works but is only good for low volts aplications.

ONE: HV spikes will often get to 20x input volts! POP goes the fets! a simple RC across the switching side will help trim the spikes but if its big enough to trim them properly the RC tank will start to cause a new set of problems.
TWO: The Resistor in the RC tank filter will start to dissipate a lot of waste energy. Part of the efficiency limit.
THREE: With a large enough value of C every time the fets turn on they are dumping it and that dump can be a lot of amps! Over current level amps! Possibly 10X ave amps. POP goes the new set of fets! Thats where the resistor builds up the heat. The low ohms resistor = high amps, higher ohms resistor = less amps but higher voltage spikes!

FOUR: chose your drive transformer voltage carefully! more is not better. you actualy want to have a reasonable dead band at the zero cross point. I have found turn on or turn off to be best at the 20% to 30% of peak volts point a good starting and stopping level. below that the waveform looks like a dead short to the circuit. Loads of amps little volts and causes those Volt spikes at every turn off point of each half of the cycle. Check your turn on value for the fets. match your drive and fet turn on volts to be at the 20% - 30% of peak if possible. This will give you a better dead band for the zero cross part. it will also cancel out any inductance lag that is from the drive transformer. Remember your transformer will have only a few degrees of phase lag from input to output winding at no current load. That dead band will cover that little phase lag error!
FIVE: That drive transformer should be connected at ct common and the fets or switching devices should be N-channel. This will give the switches an automatic pull down when the circuit is off. no need for R3 on the control transformer. plus gives you a reasonalble negative going pull to turn of the fets fully.
SIX: Watch your Phasing! you will pop a lot of fets with this simple over site!

You officialy have a workable GTI unit! (if you build it correctly.)
Its rough, crude, low efficiency, electricaly noisey, function limited and often suicidal but it works!
Hope this helps everyone. You are on the right track! I will not spoon feed you my designs but I will point the way when ever I can!
Oh and dont knock my eBay sales... P in my cherrios and you will ruin it for others!
Remember, I already have what you are trying to get! I can take my toys and go some place else if you dont play nice!

 

[Jules_Theone]

It works, but your last paragraph sums it up, it's not really a viable product. If you can get rid of the harmonics and noise without the circuit oscillating, make it more efficient, safer, incorporate anti-islanding and other safety features, then you could market the product. I got this working several years ago, but gave up with this approach because of these problems. It seems to me that newer technology can be used now, eg. two stage PWM, high frequency transformers, intelligence (uC) etc.

I am currently reverse engineering a 150watt inverter to get the design for the 12v to 360v DC converter using the TL494 PWM chip. So i can build up the design as in one my earlier posts.

 

[tcmtech]

Nice!
I have several viable and working systems. I plan to sell it. to you! and anyone else that wants to learn how to make one. thats what the eBay part is. however keep me interested at this site and I will teach you all the info I have for free. I make money from lazy people but I make friends from interesting people!
------------------------------------------------------------------------------------
The circuit you are using now is only half of an H bridge. The full H bridge cures alot of the problems that part of the simple circuit. try using a second transformer the same as the first control transformer. Use one output winding to drive one side of the top of the H bridge and the other output winding to drive the other side of the top.
For switches if you are going to do some serious playing go on Ebay and get a pair of half bridge IGBT's. big ones! I use 1200 volt 300 amp units for my experiments. Powerex CM300DY-24H to be exact. Ebay around $25 or less with shipping is what I have gotten them for.
check out the specs on datasheetcatalog.com super website for specs and aplication notes and stuff for electronics devices.
------------------------------------------------------------------------------------
For all of you out there please dont think I am mad at you or a arogant SOB. I am kind hearted and very generous by nature. I have some practical experiance in the areas you are working in. I will happily share that with you when I can. I promise!
When I point stuff out I am not trying to hurt or to humble anyone, Ever! I just want my point to be clear about something. Most likely its because I have already done what you are now trying and failed or learned something usefull that will be helpful.

Now for a bit of reality that has been overlooked so far.
Why is everyone so conserned about islanding, grid noise, grid anything for that matter.
The national electrical grid is a mountain and your GTI unit even at a 50KW class level is still a old man with a stick poking at it. If you think your experiments at a even a few KW are doing something bad dont worry about it! I promise if the grid goes down while your at full output with no saftey interlocks it will drag your GTI down with it! a grid failure is a giant short circuit as far as your GTI sees it.
Also, There seems to be a lot of consern about electrical noise. again now your an old man shouting at that mountain. your bit of electrical babbling is nothing to that mountain you think you are going to move. Also have any of you ever hooked up a good Ossiloscope to to the power comming into your home? If you have you will see its not a clean pure mathamaticaly perfect sine wave! it varies from moment to moment and can and often is very messed up looking! On this note, If you have an O scope hook it up to your mains and turn on a vacuum cleaner! Those universal type brush motors, now they make some line noise! And your local power company doesent B slap you every time you turn on one of those. SO dont worry about the little static you may make. That noise your switching circuits make will unlikely go anywhere outside of your home electrical system. The big stepdown transformer you have feeding your home is a great inductive filter. its electricaly massive in comparison to your GTI. Think of it as the first big boulder in your way to that mountain that you are so worried about disturbing.

Hope I didnt hurt any feelings of superiority! The world is alot bigger than we mere humans give it credit for!

 

[Jules_Theone]

I understand that the national grid is massive in comparison with your little gti but I'm thinking if you're generating 50kW constantly with loads of harmonics mixed in your neighbours are not going to be happy if every appliance they turn on buzzes because of it. If the sub-station acts as a filter all the noise and harmonics are still going to be present on your side (and your neighbours) and you could get unbalanced phases before the sub-station. Also if you market your basic gti, and assuming a lot of people buy one, them combined will generate a lot of harmonics at high power that the grid will not compensate for. I'm sorry if I'm annoying you but I think that the simple way you are proposing is alright at low (<1kW) as long as there are not too many of them on the same line / phase.

Have you thought about a more complicated (smarter) system because I think that's the way to go. It needs to disconnect from the grid when the grid goes down and without any disconnect your little (in comparison) gti will try to supply the whole grid and blow up. The safety features are in comercial units for a reason, although I don't believe they need to be so expensive.

 

[tcmtech]

I understand completely what you are saying. I am using 50 kw as a reference because it isthe legal limit here with out special permits. And yes my systems all use feedback control and line monitoring circuits...
I try to keep this stuff simple. If i get too technical people get turned off or frustrated and dont partcipate! I want partcipation!
I am new here so I am starting out with the basic design and working up from there. You cant have a solid design with out a solid base!
Plus if you already know how all of this works why are you here asking questions?
HaHaHa!
It may be too simple for some, I know, but I would rather start at the begining and have a good reference base built up along the way than just jump in and give you a fully functional 15 kw multi source capable super GTI that cost about $1500 or less to build including the PLC system!
I know you and everyone else wants it. Earn it! I had too!
Give a man a meal and he eats for a day, teach him to cook and he will eat for a life time! No free meal from me! I worked hard to develope my systems.

Output line voltage monitoring. We use 120/240 60hz I use a +30 v -50 v limit in referance to the 240 peak voltages not rms voltages.

Wave form monitor. that uses a simple rms voltage referance this gives me a resonable waveform limit, too square and the rms will be to far off from the peak reference.

Frequency monitoring. I allow a 65 hz limit for high and a 50 for low. My units will stay online even when I am running off of my generator power. it lets the generator run a little high at no load and bog at heavy load.

Current monitoring. I use a active feed back loop that checks the current and expands the the zero cross dead band. if you turn your switches on and off higher on the sine wave you will get a simple but effective current control. Basic PWM!

Time delays. got to have them! all of my control and monitoring functions have small output state change time delays. This prevents alot of false triggering on the protection side of the system. they are not long delays. voltage monitoring works well with a second ot two time delay before its actual output state changes. Keeps those little pulse and static blips from triggering false shutdowns.
Same with frequency drift. a two second delay before shut down will let a blink pass with out problem.
And by the way if the drive circuits are designed right they will follow blinks dips and cuttouts with out any problem. Even if it comes from half wave short somewhere else caused by a shorted rectifier in a power supply somewhere else in your home.
(blew a diode in my welder once. Created a half wave short that ran for about ten seconds before the breaker tripped.

Output filtering. If you use a full wave H bridge for the switching circuit you will eliminate 90% of the noise right there. make sure the H bridge has reversing diodes parallel to every switching device. The IGBT's usualy have this built in. And never run a GTI without the diode in place that prevents the H bridge assy from feeding back to the DC source! This is what kills the H bridge!
The H bridge only needs a single capacitor to filter everything. It goes on the DC side of the bridge. Think of the reversing diodes as a regular full wave rectifier. the capacitor is your standard filter device but much smaller than what would be used in a power supply. about 2-5 microfarads per ave running amp. my one unit runs at 25 amps average. the cap is a 100 uf 250 volt electrolytic.
Too small and the switching spikes can get big at the high current end (top volts top amps input). To big of cap and the part of the sine wave when it is charging becomes a source for harmonics and then when it discharges at the lower part of the sine wave it creats more harmonics when it is at the low end of opperation(low volts low amps input). This cap is there only to cut the top off the switching spikes at heavy currents.

Output filter. AC side thats simple as standard emi/rfi filter just like wht you find on the ac side of home electronics power supplys. Just make sure its sized to match your GTI's capacity.
TV's, monitors,and computer power supplys have great multistage units that cover the 50 - 500 watt range. Microwave ovens have a good one for that 500 -1500 watt working range. Need bigger than that check out old welders and plasma cutter power supplys. comercial stuff has a different style than the home electronics, it does the same job but at the much higher power levels, 1500 watt on up.

love the questions! hope you are learning!

should we cover line syncronizing next? Or more detail of the control circuits?
How about usefull dumpster dive parts locating to get that good stuff for free! (or just that stuff you wont feel bad about when it burns up on the test bench!)
Still have a lot more I can cover yet.
NEC Code book, practical testing devices made cheep, real life F ups( got a bunch of those!) modifiying a watt meter for high resolution measurments, VA ratings of transformers and how the bend the rules on them.

I am here for you. And always happy to help!

 

[smanches]

Yes, first post.

The interference caused by a non-sinusoidal inverter on the grid will definitely be noticeable. This is the reason almost all current SMPS power supplies have both passive input filters as well as a PFC stage to even out current draw. And these are just pulling power off the grid, not supplying power to the grid.

This is true in America and the EU, not sure about elsewhere. There has been a huge push, especially in EU, to eliminate noise from the grid. And if they already have regulations for power supplies of a few hundred watts, I'm sure they will have regulations for a 10KW or higher GTI that are much more stringent.

The amount of power you are supplying has nothing to do with the harmonics you can inject into the system.