Project Journals > Oztules

Building a 6kw pure sine wave inverter using power jack boards part2 the guts..

(1/34) > >>

oztules:
Well it has been a while, and a few have been built from France to Queensland and every where in between from the looks of it, and this is too little too late really.

All the info is in the previous posts in several sites... it is all out there now... so this is different to what I would have done months ago... it is just a simple... this is how I built this one kind of thing, and nothing new will be here I suspect..... thats the problem with lassitude or lack of time.. or both.

So we will start with the boards you can buy off the shelf, with a brief explanation.

These are not the best in the world by any means, but they are the cheapest very good cards you can get hold of easily, and cobble together a world class system

Much has been written about power jack crap, and a good amount of it is true.... but just as every compulsive  lair can tell the truth sometimes, there are a few jems in their inventory that deserve  credit where credit is due.... and total stuff ups too.

The Good:

They managed to stumble upon a generic board that seems to be in China in a lot of unheard of brands. They are the same basically as the W7 power star cards in nearly all aspects.... except the reporting to the outside world stuff..... lots of colored leds for the W7 style units, and a large digital interface on the later power jacks... which seem to be fairly useless... but I guess they tried. It reports incorrectly, and half of it is useless information..... better without it really...

The odd thing about these boards, is that coupled with a decent transformer, they are capable of showing off against the best in the world as far as just being a solid powerful inverter is concerned. They are amazingly powerful, and tough, and will drive any load thrown at them if your transformer and battery regime can handle it.... truly remarkable performance, and second to none that I have witnessed.

The Bad:

Just about everywhere they can mess it up, they have done so. They supply transformers capable of reasonable performance, but rate them astronomically high.... just sad.... at least halve or more to get a real idea of their power....... I say more, as they never build the same inverter twice... they change transformers willy nilly, use one or two for the same specs, then  supply from one to three fans to paper over the transformer swindles.... it is just sad.

Their QC is very marginal.... if it exists at all, and the first thing you do is go over their product very carefully, to find loose connections, and foreign matter left in the box... screws etc. You have no idea what will be in the box either. I have seen at least 4 versions of 8kw units turn up on the island... all different. The common thing seems to be trying to be cheap... and using as small a transformer as they dare, and then try to cool it as best they can to get some performance out of it...... in a lot of instances for particularly 12v units, this is probably good enough, as folks with them probably don't have the battery to bother it for long anyway, but in the 24v and particularly 48v range most folks can drive them hard for extended periods,..... and they will fall down from over heat.

The programming is woeful for the battery charger inter change.... not from inverter to grid sync, that is good and smooth transition, but from grid cut off to inverter stand alone .... it will blow up the fets from reactance if you have decent transformers in it... so if you roll your own transformer.... you MUST MUST MUST turn off unit before turning off grid power to it... or you will kill it spectacularly...... just a matter of time when the cross over is near zero in the cycle. The original W7 styles had lossy transformers, and so could stand the inrush currents.... but not  with big low leakage  torroids

The inrush current to the caps in the 15kw power boards will blow up 190 amp overloads just turning it on... so beware when you hook batteries to it, it will blow off chunks of the terminal.... so touch the nuts, not the threads with the lugs. ( I use a 15 ohm resistor for a few seconds, then there are no sparks.)

So with that in mind, we move on and build  inverters that so far have been faultless for over a year over several sites ( there are a few of them about the place over here). They all drive large loads during the days including hot water services of 3.6kw while running the houses with water pumps, kettles, washing machines and anything else at the time.. so going into the 5-7kw range during the daylight hours is normal.

They also reverse grid tie very nicely, and up to 6kw has been seen going into them for short periods...... battery banks are far and few between that can absorb in the 200 amp range for long @50-60v , so if you want to drive them that hard you will need to have big AC loads, as the grid tie does NOT repeat NOT care what happens to the batteries, and will not regulate. More likely, it will push the battery hard until it goes up in voltage, and this will in turn absorb less power, so the excess will drive up the AC voltage, and at about 258v it will shut down the grid tie unit from over voltage... unless you interfere first with some other way to regulate the grid tie.

So thats what we have to work with. We can make a very useful inverter that will run a decent sized household without costing much more than $500 ( even in France.... buying the cores from England). Other comparable  units capable of the same performance with the same idle characteristics are very much more than that.... probably 10-20 times in fact...... The W7 powerstar and copies...... inverters are well built, and have almost no bad habits except for the 5kwh/day idle losses... so if you have solar to spare, that may be a better way to go too.... but they will use up 5kwh/day, and the fans runs full time. There are some units like that here as well, and have performed flawlessly for over three years in their native state.... they were only $700 for 6kw and 18kw surge... very good units too.

So this is what we want to end up with:



And this is what we start with. A set of 15kw 48vdc 220v boards. They can be found on ebay, and are very much more expensive than when I first started out with these kritters... there are an awful lot sold since I asked them to put them up for sale. They have also since read some stuff on the net and improved their idle currents significantly..... so we have changed some of their poor behavior.




The newer boards have a black wire coming out from nowhere as seen here..... it goes to neg terminal, along with all the other black wires coming from the power card.... and no, I have not bothered to find out what t does.... my originals do not have it.



Here is a pic of the black neg leads, and a side on of the pins in the 20 pin header. If you buy the set, it comes complete with the start board that uses 12 of those pins. We can also see the later model replaceable fet boards... neat idea too.



Here is the front of the board, showing the driver sub board, and the CT and the transformer output interface to the control board.



Here are the pin outs of interest if you don't get the control cables supplied ( corrected by Glen at thebackshed.com )



While we are on this board, we need to stop another wicked fault that will shut us down every day, and that is the propensity of the board to track the current peaks instead of the mean, in oder to trigger the over voltage function..... we use a 5v6 zenner diode like this across R14 10m resistor. I have heard that a 1m will do well, but I like to make sure it will not shut off due to spurious voltages... that pisses me off.



Here are the donor grid ties. Aero sharp 1.5 units



And here with the new guts for it



next section  as we are up to 9 pics now.


.....................oztules

oztules:
First we need to make all the changes to the stainless steel cases. I want 4 fans, cut outs for the 180 amp o/load switches, mounting for the output o/load switch and a few other things.... so we do the metal work... mine looks like this for the fan and switches on one side.



These things are stainless steel, and hard to drill unless you use sharp tooling... be warned.
I use a 5" grinder with 1mm cutting blades to cut out the fan holes, and the o/load cut outs.
And from the inside our handy work looks like this




Next we mount the boards. I used two pieces of hard wood of the correct length to give me something to screw the main board too. It can just be seen there. I also decided to use the positive terminal on the top of the unit to the main heat sink ( where pos goes). It is the high side switched heat sink, and carries the 12 fets on it. The low sides switches have two separate heat sinks, and each heat sink takes an input to the main transformer.





Now we can see the unit with the AC output ( two heavy black wires at the top right on the control card.

The inductor ( three turns on a E65 ferrite former is also connected to one of the heat sinks of the low side, and terminates on a stand off ready to take the transformer primary. The other primary goes direct to the other low side heat sink.

The fans are seen, as is the AC o/load and the DC tripple 63A o/load switch. The Neg terminal is on the left side wall ( bottom of the unit). This will stop me from shorting out the spanner from the plus to the minus.... they are 2 feet apart on opposite ends of the inverter... that should stop me from modifying my tools using the splat method.

Neg goes from the terminal on the case to the main switch then to the stand off next to the connector wall, then to the 6 leads from the power card. So thats pos connected, neg connected, AC connected from the card output to the o/load and neutral is hanging loose and not seen at the moment... I am attempting to keep this one neat (ha ha).......

And a view from the other end of the box.



The fans are now connected in series for 48v with a dropping resistor, and the fan controller I built. Their fan control is terrible.
A note at this point, they will need to see a fan tacho feedback line. This I will supply by using a 2" fan with tacho wire, it won't be used to cool anything, just to satisfy the computer. All cooling will be done by my fan system.

The transformer AC goes direct to a massive terminal block. From there the neutral will go to the output and a sample top the control card. The active will go to the to the control card, through the CTtansformer, and back to the output o/load.. through 4mm wires



Now we can see the active come to the board, through the CT and back to the o/load. The neutral goes to the board, but nowhere else.... I am only using it to give the controller a sample of voltage, but the current goes through the CT for the active.



Here is the fan controller



and circuit board



and with it's bottom on



There is also a 25 amp 240v amphol fitting on the bottom for the AC out.


................oztules

oztules:
Were on the final stretch.....
Here we have gotten a sub front... never tried this before, but sick of having too many wires connected to a removable panel... this time we take the front off and no wire are disturbed...





So it looks like this now

[ Specified attachment is not available ]

And thats pretty much it I think....


....................oztules

Wolvenar:
What more can be said than.. VERY nice!

ClockmanFrance:
Here is my 'OzInverter'.



Standard 600mm by 400mm and 300mm deep, IP sealed metal enclosure. It had a dent on the door so it came at a good price.
The Inverter is wall mounted, with an extra support bottom bracket as it weighs 54kgs....... :o



The brass sheeting is my Ducting as I have separate air flows for the Torroid and the Boards, so I have 5 fans, with the 2 on the Boards, push pull, controlled by Oztules gentle Fan circuit.

Its been running the last few months, and is now having its final fitting, ie, its meters etc.

Oztules I have not had issues regards that zenner on R14. Should I do the mod anyway?

I have used very flexible 50mm/2 cable inside for connecting to the boards, but even so the boards are big and only have 6 fixing points around the outside, I did wonder about centre board support. Any comments ?

Yes spanner on the battery terminals is not good, so I fitted my insulated terminals High in the top of the case.

Battery cables not connected, but better photo of what's inside.



Thanks again Oztules for your excellent notes and helpful comments.   :)  :)

 

Navigation

[0] Message Index

[#] Next page