Project Journals > Oztules
home brew 5kw simple inverter..... 20kw surge ( really), and seems bullet proof
oztules:
Yes and no. If you connect a grid tie inverter to the output, it will charge the battery and help run the loads.
You cannot put just any other 50hz signal into the output, as it will conflict with the native waveform, and something will blow.
But, a GTI will sync with the ozinverter, and as it will be trying to force more current into the output signal ( coz thats what it does as a GTI), that extra current must go somewhere, so if the house loads are not using it, it goes back through the transformer, then rectified by the synchronous switching of the mosfets, which also boosts the voltage so as to charge the battery.... it is a boost converter at that time. Supplying the GTI with timing, and then transforming and rectifying and boosting that.
It has no control of battery voltage, and if you have lots of kw of solar backfeeding the inverter, you will need a way to tell the GTI to stop when battery voltage is high. I will do one later on this site, and there is ample information on the backshed regarding the ozinverter and GTI control circuits.
I'm relaxing a bit today, as the weather is abominable..... , but there will a lot more writing if I am to justify this inverter. They are worth the pain.
..........oztules.
oztules:
Ok.... time to start the control board... and it looks like this:
My pic quality is crap... but it is what it is.
First and foremost, the 13v zener is 18v. It is this one here:
So don't use 13v zenner, use 18v zenner
Secondly this 4k3 is now 1k
this one:
and in the same pic the 1n next to the word current is now 22n.
Thats it for latest changes.....
Ok lets build the board. I will explain the use of most of the parts as we go, just to give you a idea of how it works.. it is simple, other wise I would not have been able to design it anyway.
First I normally istall the 10k resistors, of which there are three.
The two near the tip35c's are simple voltage drivers for the clamp of the 18v and 5v voltage regulators, made up from the tip35's and the zeners. The third 10k near the 8010 and current trimmer is for making the temp probe work for a 10k@25c thermister.
Then the 7k5 resistor. This is for calibration of the voltage feedback transformer. Use 7k5 if you have a 14v tranny ( I had a heap of them... thats why I used them. If you have a 12v transformer for feedback service, this value will change... to around the 4k mark etc. etc.
This shows them done.
now the 1k resistors
The three of them near the 500r trimmer, are for ( left one) on/off ground for switch, next one is for led, and so is the next one.
The 1k that replaces the 4k3, is a damper for the current feedback to the 8010., and the one near the current trimmer is part of the drive for the SCR for over current shut off. ( needs a reset to recover from this)
The 1k8 near this is the drive resistor to the SCR itself, and the 1k8 near the voltage trimmer is just a part of the divider for setting the voltage on the output.
The 1k4 up the top is to limit current to the o/load led.
The 100r near the temp probe pins, simply carries information from the temp probe to the temp feedback on the 8010.
The temp probe is handy if you want to shut it down using some other signal, as it soft starts from removal of the signal... just for interest. I did use it originally for the on/off, but the flashing leds on the front annoyed me, but the shutdown via spwmen pin 6 was a bit dangerous for me for a while... it is sensitive so beware.
It is ok in this current circuit configuration, but if you stray from here, it can cause wicked problems.
The 120r near the current trim is for damping the CT.
The 18v zenner provides power for the driver chips, the 5v6 for the 8010, and the 6v8 for the scr shutdown system... can't remember why I needed 6v, but it was necessary at the time.
The big 5w 120r is to slow the inrush from the battery side of things. Without it, the thing will blow up from spikes when I originally used 7805, and 7812 etc. They did not like instant turn on, and took a finite time to regulate... so the 120r and the 100uf63v was used as a simple time delay to get things running smoothly.
It also functions as a current block when we reset by shorting the tip35c side of the resistor to ground via a push button. This kills the power in the circuits, and resets everything, including the scr on current o/load, and the 8010 when it hangs from too slow a start up when you first connect the battery.
All the 104 and 10uf ceramic caps are used for noise suppression and ripple damping.
The 4 electrolytic 10uf are for damping and for current/voltage pump for the high side driver circuits. I tried ceramics, but the waveform changed slightly, so went back to electro's.
The fr107 diodes are used throughout,as I had thousands of them. They are required for the current pumps, but any diodes could be used to rectify the 14v from the tranny, and the CT.
The crystal is mounted with a small clearance to stop the case contacting the pins pads. Someone else designed their board slightly different, and suffered blow ups, and it took a long time to find this fault. It is a 12meg crystal... a few cents on ebay.
The 2 big caps on the right end are 2uf 250vac types. The actually shape the waveform, so must be present. You must not run this testing or whatever, without the 240v feedback, or the pulse width will go max, and probably blow up. It must have feedback connected. Thats what the 4 pins are there for ( 6.3mm ones) one set for 240v input, and I use the other set to drive the AC meter.
so now the board looks like this with all the diodes, resistors and ceramic caps placed
Now we look at the 8010. The boards I use to mount them are shown, as are the female strips to mount the boards, and the male strip for the pins for that board
Here we see how pin 6 is up not down. This will be used for on/off switch lead.
Now with the sockets mounted, and electro's etc
and now with the lot except the tranny and 250v caps
and we have run out of attachments again....
We can see from that short run through , that this thing is so very simple, yet works so darn well.
..........oztules
oztules:
And now the board close up of the left side of things showing the voltage regs and the driver stuff
And the right side
So thats it for the control card.... except for the small 240v:14v transformer, that will go on later ... currently run out of them :-[
The pin outs from this card are directly capable of driving the power jack power board if you have one and a blown PJ control card... with exceptions!
The 56r resistors on the PJ card must be replaced with 5r6 resistors, or the board will overheat very quickly. The bypass diodes on the power card of mine need to be added across these 5r6 replacements... and it will work perfectly well.
We will move on to the power card tomorrow perhaps, but it is simple as can be, and is just fets and driver resistors, bypass diodes, and gate source resistors, and a small filter on each leg.
Note when you make the power card, mount just 1 hy4008 on each leg, and DO NOT add the big bulk capacitors yet.
When testing, we will use a small 28v:240v transformer, and the lack of caps means that you will not have the power or energy available to kill the fets if you use a current limiting resistor from the battery. I use a 48v led driver power supply, limited to 10 amps ( about $35AUD on ebay), or you can use a current limiting resistor and battery. With no big caps, we limit the current seen by the fets to the power in the power supply. If we add the caps, they store a huge amount of energy that will be dumped into the fets in case of a problem you have managed to invent.
If it tests ok, you can add the caps and still have just 1 hy4008 per leg. You will be able to run up to 2kw easily in that configuration... then if happy, add the extra fets for full power.
You can use a few hundred watts without the bulk caps to test, and if all is well, then add the caps and extra fets..... will probably repeat all this later on when we come to the testing stage.\
Note, all the 0R resistors were for when I was testing the design with home done single sided boards. I left them there, as they provide the via's to the inter track connections..... I could have used vias, but was too lazy to change the board design......
.........oztules
ClockmanFrance:
Hi Oztules,
As usual my congratulations on showing your ground braking endeavours.
Nice articles with nice pics..
I know that you and i, and a few others, think that this is an easy straight forward project. But I have realised over the past few years, that what we think is understandable and simple is not understood by the majority.
Yes, I am still doing a New 2nd edition of the "How to make a OzInverter" book, but boy oh boy how difficult and awkward it is in putting down what is appropriate to get a finished project, especially as I am doing the book as a non profit project, and it sinks a lot of my time.
Present 1st edition is based around the PJ boards, and with a supplement with the new PCB's etc, ISBN 978-0-9935903-0-6
Most of the book is photos and explanations and block diagrams, and like before I have split the 2nd edition book into distinct chapters with appropriate sub chapters. ie, .... Intro, ......Tools Equipment,..... Materials and stock,...... Making the Oz-toroid, ......The Pcbs, Power, OzControl and the Ozcooling board. ......Enclosure, ...... DC side, ..... AC side, .....Meters, ......AC coupling,..... Tech stuff. ......
Got be careful, as the tech stuff Chapter, although very interesting, can overwhelm the book. Especially as I love your empirical evidence nearly 250,000 words and 'Warpspeed's toroid stuff is fantastically instructive.
I might have to add sections on actual tool use and limitations and construction methods.
Anyway my point is, that your OzInverter is so bloody important today, and where WE KNOW THAT THE OZINVERTER WORKS, is simple, robust and very cost effective.
So Hence me recording your works for posterity, my 3 lads, and other folk out there that have the abilities to do there own stuff and follow a publication that is a 'How To Do'.
Carry on your good works 'oztules' and hopefully I can make you proud with my humble publication.
My regards to your MRS, and that neighbour of yours, (I still have this image in my head of your Mrs attacking a faulty Inverter with a large felling Axe, haha!)
Best, Leslie Bryan.
oztules:
Thanks Les,
But the next part will be in a few days.. busy as all get out at the moment.
As you know this takes some time to do.
Hope all is well with you and yours.
Open source is cool, but someone still has to write it up :(
John Tulloch
Flinders Island
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version