Tiny Grid Tie Inverter Unit, Maybe.

[tcmtech]

The tiny wind generator design got me to thinking about a new possible design related to a tiny GTI.

This one is intended as purely a proof of concept design. That is its just for the fun of it!

The intended working range is for the 25 - 100 watt 12 volt solar power systems. The intended highest efficiency sweet spot is intended to work at around the 12-15 volt area. With hopefully 50% or better peak efficency!

However I can either build it and debug it with the standard hands on method or I thought someone here that is good with the simulation software may be nice enough to debug it and correct the errors and mistakes and or change the values I estimated at and get and actual numbers to work with.

Like I said this is a just for the fun of it and proof of concept design.

Here is the basic specs.

T1 power transformer. 120:20 ct 25 - 100 VA
U2, U3 : transistor Opto coupler
Q1,Q2 : 10 - 40 amp 600 volt IGBT
555 timer: 5 -10 second delay
SSR,TR1 : standard 10 amp solid state switch.

Not sure if my R1 - R3, C1 - C3 snubber circuit numbers are close though.
Same with C7 as a voltage reference current limiting device for the U2 & U3 Opto couplers.

 

[SPDCHK]

TCM,

Maybe it's just me, but the output of the 555 (Pin 3) does not seem to be going anywhere except to fire an opto-isolator (you called SSR). The output triac is not connected. It seems to be connected to the TR1 on the output 120V line. Is this correct?

 

[tcmtech]

I didn't have an exact schematical representation of a full optically isolated solid state relay so I just did the note to show I was representing TR1 and the SSR component as as a full SSR device.

I cheated a little. Just asume the SSR device and the TR1 are a standard 3- 32 volt DC control - 270VAC 5 - 10 amp SSR in one package.

Sorry for the confusion.

 

[Mr RB]

I like it.

Tiny grid tie is cool.

I would change your opamp cct by adding a resistor in the positive feedback to add a little hysteresis, ie on=13.8v off=13.3v. I think you'll get best efficiencies that way.

 

[tcmtech]

Thanks!
I couldn't sleep the other night and pondered on how to improve a simple 100 watt unit I built a few years ago.
I just thought I would post a basic concept circuit and see what we all could come up with in refining it.
I Figured it may get some positive interest due to so many people working with smaller solar power set ups.

I think that by using CMOS IC's and a little tweaking the control and switching functions could possibly be reduced to under 50 Milliwatts.

If the IGBT gate drive circuit shows a reasonably sharp turn on and turn off transition the switching loss should be very small too. The only big efficiency loss should be the transformer itself.

If anyone has any schematics changes or refinements post them! It will help me and everyone else to see the evolution of this from theoretical concept to an efficient workable design.

 

[Martin Powell]

This Circuit seems very cool and I intend to build it
not sure if square waves are the best way to input back into the grid ????

 

[tcmtech]

The LC type tank filter plus the transformer work as fair wave form shaper.
I have never had much problems with any AC line noise if the filter is designed right.

I dont know if this circuit will actually work very well or if at all.
I have yet to build one and debug it.
There are a number of component values I have on it that may be way off!
I just went by theoretical formula calculations. And I never ever trust theoretical math. Its a big fat liar more often than not!

So be prepared for some fine tuning!

I will try and get around to building it one of these days too.

 

[Existenz_69]

Speaking of small circuits... On smaller circuits like this would the Texas Instruments TPS68000DBT be an alternative choice since it "claims" to have pure sinewave and drives a true h-bridge? I found a schematic for a full-bridge CCFL controller that looks much like the inverters we are trying to build only much smaller and minus the CCFL lamp. I found the analog dimming interesting. May be totally useless for our needs though (probably is since i found it... )Check out the PDF below... I tried building yours in cadsoft eagle 5.4 professional and it seems to crash my pc after autoroute command so i couldn't export it. I wasn't trying to SIM, but to check for errors and lay out the traces for the board. Thanks.

 

[Martin Powell]

Looking at your pdf file seems like this would be ok to step up the dc voltage with a few extra mods
Then you would have to convert it to dc then use a 2nd h bridge to convert to Ac and use a PWM to control 2nd H bridge.

Has anyone input mains 50hz into a PWM via a coupling transformer to control power Fets
And thereby put simulated sine wave back into the grid this way only needs a small filter and
Must be more efficient.

 

[tcmtech]

Interesting Idea.
I have also been thinking about how to use a Class D type audio amplifier IC to drive the H-Bridge in order to get a very clean sine wave just for greater overall efficiency though. This is the class D amp IC I was planning on experimenting with some time. I would just use the outputs to drive a set of IRF2110 or similar high/low side driver IC's that run the line side connected H-Bridge.

I am not one that cares much about output filter size. The output filtering that I have been using is not all that large in comparison to the rest of the GTI device. Just a the right size PFC capacitor and a basic dual choke AC line filter with the right amp capacity has shown it works just fine. Or at least with the low frequency transformer design it does.

I also think that by using a line side coupled op amp with all of its power and control circuits placed on the line side of the GTI it should be able to directly sample the line side wave form and feed it directly to the amplifier IC. That way the slight phase lag is brought down to less than a degree or so.

The input power source could then just be a simple fly-back or some type of switching inverter (Modified store bought inverter?) to take the variable input voltage and power available and then isolate and step it up to the line voltage levels that would supply the actual power to the H-bridge.

The down side to all of this is the overall complexity of the added switcher inverter isolation efficiency, 80 - 85% typical, plus the class D H-bridge drive and other factors dont add up to much greater overall efficiency gains or a overall improvement in line side wave form.
But the cost and complexity goes up.

If the input power source could be electrically isolated and was putting out a voltage high enough to run at line level voltages then the switching inverter isolation could be removed. I can see a theoretical GTI efficiency that would easily get into the 90 - 95+% overall peak operation efficiency. Maybe.

Theoreticaly speaking of course.

 

[killivolt]

Very Cool

Thanks, I sort of understand the process now. I'm going to have to get some reading in to understand how the ssr package how it works on the back end of things.

Once I get that then how the front end works and what the 555 timer is doing if it's an LC tank then it's doing something to the input voltage ? I just don't know what except maybe filtering the voltage smoothing it? or something is my guess.

I'll just keep following the thread. Interesting input on the Texas Instrument application seems your getting people thinking TCM.


kv

 

[tcmtech]

The SSR is just the line connection switch. Nothing more. it could easily just be a mechanical relay too.

The 555 timer is just a time delay for the on/off connection to the line. It lets the GTI stay connected during short input power drops of a few seconds.
That way its not constantly connecting and disconnecting if the input voltage is not smooth. (shadows moving across solar panels)

 

[killivolt]

I've gotta head out I'm glad you responded. Now I get it.


Back in the day. I went into Refrigeration and Air Conditioning as they said you will need this to go into the Solar Energy Field "right" Bullsh...... That put me on the wrong road from the git go. Now I make up for lost time trying to learn Electronics.

Rant: I should have had this all down back then. Just pisses me off when I ignorantly believe in lie's. But then again that's life.


Thanks for the help.


kv

 

[Existenz_69]

Looking at your pdf file seems like this would be ok to step up the dc voltage with a few extra mods
Then you would have to convert it to dc then use a 2nd h bridge to convert to Ac and use a PWM to control 2nd H bridge.

I was mainly thinking of using the only TPS68000 to drive a moderate h-bridge, or using it to drive some small fets or the like for driving a larger h-bridge. That schematic was a full dc/ac inverter, but was using a HV transformer as the power transformer. I was hoping that one might be able to use the analog dimming feature to control the output current/voltage, or frequency. One could easily adapt this IC to a higher power setup with a few small mods, and hopefully keep the true sine. The only trick would be the phasing.

 

[rx5]

The tiny wind generator design got me to thinking about a new possible design related to a tiny GTI.

This one is intended as purely a proof of concept design. That is its just for the fun of it!

The intended working range is for the 25 - 100 watt 12 volt solar power systems. The intended highest efficiency sweet spot is intended to work at around the 12-15 volt area. With hopefully 50% or better peak efficency!

However I can either build it and debug it with the standard hands on method or I thought someone here that is good with the simulation software may be nice enough to debug it and correct the errors and mistakes and or change the values I estimated at and get and actual numbers to work with.

Like I said this is a just for the fun of it and proof of concept design.

Here is the basic specs.

T1 power transformer. 120:20 ct 25 - 100 VA
U2, U3 : transistor Opto coupler
Q1,Q2 : 10 - 40 amp 600 volt IGBT
555 timer: 5 -10 second delay
SSR,TR1 : standard 10 amp solid state switch.

Not sure if my R1 - R3, C1 - C3 snubber circuit numbers are close though.
Same with C7 as a voltage reference current limiting device for the U2 & U3 Opto couplers.

@tcm

HOW do you drive the IGBTs if I just power it up from a 12V source? I mean I can see you get (isolated)gate drives from the output of the transformer.. OR is that it? I mean feeding its own output to sustain oscillation?

 

[tcmtech]

Well this design is set up specifically so that if the line voltage is not there it cannot run.
If you wanted to make a regular inverter and not at grid tie the circuit is considerably different.
Look around on this forum for things like push pull oscillator. I have posted some schematic for very basic ones in several threads.

 

[Alan Dedier]

grid-tie connection

I am new to Electro-Tech. Just got my NABCEP badge. How do you do your connection? There are specs on how long the grid goes down until you must anti-island. Xantrex used to make a block (for $450) that did this. I have 24/7access to dip solder pot about 8x10", washing machine, and lead cutting machine. Use mild acid solution flux. Orcad or Protel layouts. Could check or help layouts and if sent kits do the boards.

 

[ignisuti]

Looks like it's been awhile since anyone posted on this. Any improvements/progress since the last post?

 

[tcmtech]

Summer time for me. No time to play.

I work longer hours in the summer so I dont have to when its winter.

I did make a better working version of it a while back though. I just haven't got around to doing a proper schematic for it. I will need to take it apart now to remember what went where and to get the right components values.

And it does fit inside a standard computer power supply case too!

 

[summitville]

RX5, The AC Line Voltage from the Grid triggers & sync's the IGBT's to create a Square Wave inpulse into the transformer's primary for less than 100% of each half cycle of the Grid's Sinewave. There is no "sustained oscillation". No AC Voltage from the Grid will shut off the IGBT's.