Building a 6kw pure sine wave inverter using power jack boards part1 transformer

[clockmanFRA]

Meters ,  As per your Notes Oztules, I have these on the way here.
For looking what is happening on the 220vac side ……..
http://www.ebay.co.uk/itm/181573167298?_trksid=p2060353.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT



And the DC battery side, this has a 300a shunt, I already have one of these on one of my 48vDC 30a Wind Turbines …………
http://www.ebay.co.uk/itm/171204358909?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT



And 3 of these Temp Meters so I can see whats getting warm…………..
 http://www.ebay.co.uk/itm/251637079552?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

[dgd]

Too many meters although all the temperature measuring is a good idea 

To monitor what is happening I am gathering together the various sensors for DC and AC voltages, currents,
AC frequency and temps at various locations then using an Arduino Meg2560 or DUE to monitor them and generate and serve a web page or three.

Might even include a few gauge displays (in the web pages) and get some Highcharts graphs included to look at performance over a few hours/days and maybe weeks.

dgd

[joestue]

those temperature meters are useable, but they may be off by 2C from one meter to the next.

Also you will want to hot glue the wire into the back of the case, it is very thin cable. they run from 1.5 volts so buy AA's to run them, or buy a pack of 50 of those lr44 cells to run them from ebay, they burn up a lot more current than they should.

[clockmanFRA]

Hi dgd, I want to see what's going in and what's going out.

joestue, the thermometers are really just temporary to see whats getting hot and where. So after a couple of months I can install a permanent Fan temperature controller. I think the PJ's come with some sort of temperature switching?


My 2off wound 190mm diameter toroidal cores. They came with core inserts.



Putting the woven glass fibre on the edges of the cores. I am using a thin-ish viscosity Epoxy Resin. About 12ml.
Here I just put the Epoxy on the wrap around, let it set overnight and Epoxy the turn over the following day.
I use 50 Micron Melinex clear sheet as a table top, as nothing sticks to it.



The two cores joined together with a woven matt fiberglass and Epoxy, and then Epoxy a 25mm wide strip around the joint. The core is now 20.5kgs.

[clockmanFRA]

3 layers of 20mm wide Mylar winding tape.

One of those fold away work banches,  with a spare bit of ¾ inch ply board screwed down. I cut a 95mm dia hole and cut it away so I can do a quater of the toroidal core before needing to lift and turn it. I have attached a single support leg so the whole lot cannot topple over.

[oztules]

Looks the part... nice ...... nice set up too.

What are the inserts made of?

Keep the pics and write up coming.

"so I can do a quater of the toroidal core before needing to lift and turn it"..... heavy little beggers aren't they.


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

[clockmanFRA]

Those 'Inserts' seem to be just smaller toroid cores, I think just to keep my big cores from deforming while in transit here.? Anyway they come in handy.

And yes oztules the 'beggers are heavy'.

Just finished the first winding/layer of 1.8mm diameter enamelled copper wire. I now have 'soar thumbs' and 'soar first fingers', as on each corner, 472 of the blighters, I have to hold the copper wire tight and form it ready for the next corner while keeping the wire tight.........

I have gone for 118 turns.. is that OKAY?...

My reasoning being, that as each of my 4 layers goes on, I test each winding.  I can reduce the later windings to match the first, as each will be epoxy coated.

But as you said "don't go below 114 turns".

The first quarter gets under way, it was so rubbish I went back and did it again.
The driver/drill is handy for bobbin winding, as the speed is very controllable when winding up those 42 meters. gulp!

I will try and get the bobbin wound a bit neater for the second layer, as its important to keep the wire tight when winding the toroid core.
Yes the bobbin fits nicely down through the core.



First winding just finished. It now needs a bit of loving care to tidy up some of those wayward wires before I epoxy.



 

[Mary B]

Most toroid winders do not use a spool type bobbin, they use a long flat form like a stretched H shape that fits through the core. Wind your wire on that and you don't have to twist the bobbin to go through the hole while winding.

[oztules]

Clockman, thats a good looking result thus far... making me look bad

Had a 1 year old 5kw PJ turn up today with a   noisy transformer. It turned out to be one of the totem driver transistors causing the problem... it went short to ground, so the h bridge was unbalanced to say the least..... first failure I know of from inverter duty only. It struggled to get to 240v in that state, and made a horrid noise.... all good now.... a 3 cent transistor can make a difference. Can't see why it went... will delve further to see if I can find anything else to report.

I think the 118 is a fair target, wish I could be more accurate.... but it sounds like you have more core than me or very close, so 118 should be fine. It won't hurt to do more if you want to get the magnetizing currents even lower. The further you are from saturation, the lower the currents required.

At some point, it gets fairly linear, and not worth going further as the copper loss rises, but up near the top of the curve, the results from lowering the current by increasing the turns, is very very handy. ( it is the extra turns on the primary to cater for the more turns on the secondary that matter really).

Beware, you may have to cut down your favourite bobbin to smaller size as your hole gets smaller. There is no crime in having to join as you go.... (twist and solder).

Handy looking little torroids you have there  to play with.... wonder how many turns/volt they would be....instead of volts per turn.



Thats true Mary..... but....

But i found with 2mm wire, it is much much easier to use the bobbin as clockman has done. If you push it through the correct way around, as it unwinds,  it places the wire hard against the core wall, if you use a flat bobbin with thick wire, it just becomes a nightmare to get it to sit correctly... and if you use reclaimed wire (me), it is so tensile as to be an absolutely painful experience......... unless you use the round bobbin.

This gets more problematic as you get to less hole size, but the flat bobbin offers no help at all in keeping it both taught, and straight and tidy.

For winding your 1800v using spider web comparatively, the flat bobbin way will do fine, but I would still use a round bobbin if i could, as I have found it so easy to remain tidy and tight.

But if your hole is too small, then the flat bobbin is the only answer.... sadly.



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

[SparWeb]

Still watching with interest.  Love the way this is turning out.
I don't actually have anything to add - except to Clockman, that round bale time has only just arrived here.

[oztules]

ooops...... didn't mean to press reply button..... thats why I am the village idiot I guess.....

[clockmanFRA]

Second winding going on. Again its 118 turns and will be in parallel with the other 3.....

As the centre hole gest smaller so my plastic bobbins are being down sized. In this photo I ran out at 3/4 of the winding so I have done a twisted joint, and soldered it.
Over the years winding the same diameter wire with my wind turbine PMG coils, I use the flat of the sharp knife to remove the enamel from the wire then a some emery paper, do the join and cover with glue lined insulating shrink wrap.



2nd winding done, and yes oztules the second is neat, especially the way it fits in the space of the first windings.

Testing time before I epoxy the second winding.
I followed your instructions. Light bulb is a night light, but the readings were the same for each winding, and Zero, well 0.01v on the final test across the two live outs.

[clockmanFRA]

Probably insignificant...............  The second winding used about another 1.1 meter of copper wire than the First so I see a slight difference in the resistance readings between them.



And the Second winding....

[clockmanFRA]

I will try and go for the Fourth winding, but it might be tight getting my 14 turns of 50mm/2 down the middle. ?

My first winding and fiberglass cloth took up 10mm of my 90mm original diameter centre hole.

I see what you mean about compromise's...........
 

[keithturtle]

Most toroid winders do not use a spool type bobbin, they use a long flat form like a stretched H shape that fits through the core. Wind your wire on that and you don't have to twist the bobbin to go through the hole while winding.

I built one that uses a split bobbin. Spent 2 weeks on the stupid thing and the best it will do is 26 ga wire. Gave up and wound 'em by hand.

But I proved to myself that it can be done

Turtle

[oztules]

WOW... how cool is that 
It certainly wont do 2mm tensile copper, but gosh what a fine effort that was... sheer class.


All going to plan clockman looks good too.... every micron counts.... and that is pretty good that you only spent another meter or so with the greater diameter. You could actually stop there for 6kw. Your 11.2mm wire probably has thick walls, you can strip that off and use heatshrink if you need more space... makes a huge difference... and some  heat shrink is high temp too... you could push the thing very hard with that instead of the pvc.... more flexible too.

You won't buy a better transformer than that. It will be all but silent with resistive loads , hair drier or any other phase switching triac controlled high power thing will still make some noise, but it is very quite.......... right up to 6kw or more you'll find..... ie if your running say 2kw, and then turn on an electric jug of a few extra kw, you won't hear a change in sound.

Looks beautiful.

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

[metalmangler]

Nice looking work,....."every micron counts", I can apreciate the idea,just cant work to those tolerances.
The size of the coils is confidence inspiring.
Building a winder for these would be quite a chalenge, simmilar to fillament winding in composite parts manufacturing,
where the bobbin has to be passed off over and over, or a rope factory where this happens in one dimension only,
still facinating to watch, a kinship to any weaving, just hard to wrap your head around....end up getting tied in knots.....
and loosing the thread

Mangler

[joestue]

In my opinion it gets easier with larger transformers.

I recently rewound my variac with 15 awg wire, I think I used about 2 pounds of wire.
http://johansense.com/bulk/variac.JPG

I pulled the wire off the spool about .7 pounds at a time and threaded it through the coil in a big loop about 7 inches in diameter (so perhaps 50 turns), then cut the wire off the spool and started winding the turns.. in the opposite direction as the wire I had threaded through the core.

So as I'm winding the turns where they need to go, the coil of wire is growing larger in diameter.. so I had to stop from time to time to thread the coil through the core to reduce the diameter to keep it from getting out of hand.

Anyhow if I had cut a slot in my desk as clockman had done, I suspect I could have wound the core with one piece of wire rather than 3 sections, but I was too lazy to try and calculate what weight of wire I needed.

Anyhow it was not at all difficult to get the variac wound.. the only challenge was I wound the core too dense.. you have to space the turns out more than you'd like to. I also found out how they get machines to do it.. the secret is the disk on top and bottom of the core has slots cut in it for the wire to rest in.
I was able to get enough turns to make a 0-240v variac, from 15 gauge wire. Originally it was wound with 18 gauge and was 0-280vac.

[Neil]

clockmanFRA , how many volts are going to the toroid from your pj board?

Neil

[frackers]

I built one that uses a split bobbin. Spent 2 weeks on the stupid thing and the best it will do is 26 ga wire. Gave up and wound 'em by hand.

But I proved to myself that it can be done

Turtle

I've been tempted to do similar when I start on my toroid - bicycle rim for the spool side and several sections of window winder mechs that have a toothed arc of steel welded into a circle for the winding side.
Too complex for the 1.8mm wire but I might use the (split) bicycle rim and wind by hand. Easier to get the 50m of wire (about 30 turns) on the large diameter than heaping it onto an bobbin small enough to go through the hole.

Got most of the parts (high power wheelchair motor, bearings etc) to remake the cores themselves first.

[clockmanFRA]

Thanks oztules, for your kind comments. Please do shout if I am heading down the wrong road.

I await your replies with baited breath, especially that last testing stage.

 My friend in the South of France and I, discussed your coil testing and balancing with several emails.   

 I have not soldered my coils in and out together yet as I have more testing to go.

Re the 50mm/2, 14 turns, Yes I see your point about the insulation thinning down so I still will try for 4off 118 turns.

Question oztules...... On the last of the 4 windings its likely that the core diameter will be to small to get all my 1.8mm dia wire to fit neatly along side each other, so they will have to overlap each other. Should the overlaps be evenly spaced etc around the inside core, or just bunch up ?

Neil....  Well the plan is to throw 48 to 60vdc in to the toroid, and get 220vac out of it.
However, I trust oztules will enlighten us on that part, especially the bit where I will need to call a Taxi to get the electrons to their correct place of work...... he he.

Regards toroid windings, I did a quick sketch of my double up cores.
As you can see winding that small inner core gets difficult as the hole diminishes, so if I did a triple core then winding it has to be done by hand bobbins, I think.



For me getting this size of toroid to do my work is brilliant, and doubling up, even tripling up of the cores means that perhaps this efficient design of oztules, is beyond normal commercial practise. 
 

[oztules]

Neil, we will effectively be winding a 30-32v:240v transformer... even a straight 8:1 ratio will be fine..... ... seems strange I know, but you have a massive variance in input voltage ie flat batts can be as low as 45v, and full may be 62v.... so you can see it is no point winding for 48. So we wind for the worst scenario, and with high power, give it even more  head room.... so 30 odd volts is what we want.... so to stop clipping at low voltage, and allow for less than 100% duty cycle, we are well below 48v.

Clockman... looks very good. It is a much more efficient way to do transformers this way, than doubling the number of transformers to share the load like PJ do... you use at least 25% less copper for the same turns.

It does not really matter too much whether the turns are bunched or spread, so long as they go through the hole, we will be happy. If it is neat, then your likely to have very low inter wire voltage, and the epoxy will negate wire rubbing , so you can then be happy with the breakdown voltage of the lacquer being over kill..... should be fine.

Measuring the secondaries..... well the system does work, and if you are even only 1 turn out, you will see it easily. 1 turn difference is a huge problem for tight transformers like this, and would create high idle currents due to the tight magnetic coupling. Looser coupling like EI transformers would be less dramatic, and probably absorbed into the other losses more readily.

To get an idea of your copper requirements... if you use 50mmsq for the primary, them 50/8= 6.25mmsq will be a balanced secondary.... ie 2 in hand 1.8mmm will be 5mmsq, and 3 in hand will be 7.6mmsq... so three is over optimal, 4 is over kill ( which I'm not against by the way), but for way way less than 100% duty, you will have created a monster for sure..... put your Sunny Island to absolute shame I suspect.

"so if I did a triple core then winding it has to be done by hand bobbins, I think.".... I have done a triple core as well, and it is painfully heavy to work with.... but gee I like the less turns.......



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

[clockmanFRA]

Thanks oztules for the clear information regards the bunching up of the wires, and the explanations on the  testing each winding for matching each other, the design concepts and ratios and copper used.

I have just finished the 4th winding of 1.8mm diameter wire with 118 turns. It got tight, with 10 wires having to double over, but I managed a card jig of my 14turns 50mm/2 of the primary and put the 1.8mm wires in the spacing's.....  Knowing my luck they wont be perfectly spaced but I have only used 5 out of the 14 spacing's.

Yes, the last winding leaves a 62mm diameter hole through the toroid. phew......

Question..... any tips about wrestling with 6 meters of 50mm/2 ? Out in the garden/field how many rugby players required.? Did you start at half way and wind each way?........What about wrapping the toroid before the primary goes on ?

Picture ... shows the simple winding testing 220vac light fitting, yes it works well. That PVC/plastic solid rod, I use as a resin pusher downer and just roll it around before the epoxy sets solid.



Picture..... The Powerjack 10000w 48vdc 220vac Inverter I purchased on fleeway. One of two Inverters toroids being stripped down for bits. These are the biggest PJ do as they only just fit in the case..?
Primary.... 14 in hand of 1.8mm diameter with 18 turns.
Secondary.... 2 in hand of 1.3mm diameter 4 layers, each layer of about 152 turns.
Core ..... 155mm diameter x centre hole 80mm diameter and 75mm high.



My toroid transformer with its final 4th secondary, all epoxy resin-ed and drying/setting.
The plastic bobbins are now a compact size and the last winding left me with 2 twisted/soldered joints.
The little card jig with the 14 primary windings showing, so I could space those 10 off secondary's.
4 copper 1.8mm diameter wires, each is one complete turn for each layer of windings, so I know how much I am using.
The core of the PJ toroid, look at that, they have rounded the sharp metal corners? 

[oztules]

Just a brilliant job Clockman.... 4x1.8 secondary is quite a monster.

Good to see the windings of the power jack. If you joined them together, it would appear that  150 turns secondary and 18 turns primary would give you a decent little transformer too, and 25% less wire.

It is comforting to see their turns ratios, as they are the same as ours..... about 8:1 ie they have 37ish  turns on the primary and 300ish on the secondary  ( series parallel )

Yes wrap the secondary with the tape... so it has at least double insulation.

Start your primary half way through... and I used hand clamps to hold the cable tight into the center and elsewhere, otherwise it is very difficult to get it to form neatly... so clamps and a hitting tool to GENTLY shape the corners.....

Your epoxy on the last layer can be a lot more generous, and it won't be a bad thing if it shapes the ends to a cylinder instead of the wire defining the shape.... I let it sit on the plastic, until near set, and cut away the excess so that the wire is almost completely immersed in the epoxy at the ends.... protection from me bending the primary too forcefully.

Once you wrestle the primary on, you will have one of the best transformers you can get for these purposes. Powerful, silent, no chance of wire rubbing... it should last forever.

curiosity... why did you destroy the PJ's you bought, I would have thought you would have left them complete for lightweight spares etc.


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

[clockmanFRA]

Thanks. more epoxy it is then!

With just its 4th winding, my 'Oz Animal Inverter Super Duper' is 28kg....... and I notice at this weight the transformer it self when just resting on the table, can easily damage the wire enamelling/lacquer.

The locals around here think I am already stark raving bonkers, so watching this mad chap wrestling with 6 meters of black cord, in the middle of the field, on a work bench, with appropriate machines/cars around as temporary holding equipment is going to be most entertaining.!


"curiosity... why did you destroy the PJ's you bought, I would have thought you would have left them complete for lightweight spares etc."

I purchased the 10000w and the 8000w, PJ's for not much money. I really struggled getting any decent size toroid cores from old PV GTI's.
 Originally I was going to use the 10000w PJ toroid's and do a Oz inverter on them, and leave the 8000w PJ alone as emergency backup. But upon opening the PJ 10000w case, I realised that the PJ cores were not that big, and I was going to struggle with the small core size, 155mm diameter, 80mm diameter, 75mm high, 7.5kgs.

 As the 10000w has the same boards as the 15000w, minus 2 caps and a few component changes, I dedicated the 10000w as strip for spares. Especially as a BILL has just arrived, a month later, from the French Government demanding 40euros for Importing from China those PJ 15000w power boards stuff.

However, now I have that 'OZTULES CONVERT & MAKE' ailment, warning highly contagious.
 
 5 days ago I thought, "Hmm perhaps I could do a conversion on those 2 PJ 10000w toroid cores."
And I had just run out of mylar tape for my 'Super Duper', used about 400 meters, very difficult just getting small quantities quickly here in rural Normandy. In the UK in a couple of weeks, so I can call in at the firm that made my toroid cores 2off 190mm diameter, 90mm diameter and 60mm high, .... http://www.airlinktransformers.com/bespoke-built-transformer-service/ .... and pick up stuff.
 
So the 10000w PJ is no more, Arise my 2nd 'OZ Super Duper' . he he.................

[clockmanFRA]

Been on a short break with my youngest 8 year old, and doing some work.

14 turns of 50mm/2     .................

I have hooked my winding bench to my trailer, and away I go......



Completed, phewwwwwwwww.

14 turns, 50mm/2,  through the centre equals exactly 6 meters, tum te tum, that was very close  .



Oztules , what do you do about wrapping and fixing those 14 turns with?

Thanks.

[oztules]

"Oztules , what do you do about wrapping and fixing those 14 turns with?".......... errr.... nothing.

The rubber bottom and rubber top washer and steel " washer/plate" stop all movement....

Nice to see some one else suffer instead of just me.....

Thats a serious transformer, and will give years of service at high power....never mind inverter usage only.

Well done........ too well done actually.... I've been out gunned ..........


Whats the voltage on the primary,  now with 240v on secondary??


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

[clockmanFRA]

Warts time.

Making that extra 4th 1.8mm diameter 118 winding left me on paper, just enough space to get 14 turns of 50mm/2 through the centre hole and lay tidy, in theory.

Nope it didn't work out, so I ended up with the 14 turns with 1 and a 1/2 turns overwrapping itself. 



However, I spent some time re=reading 'oztules' notes again.......
 "To get the thick primary in there... 14 turns will be difficult, maybe more iron, and less turns will allow you get to a stage where the turns are possible with the smaller hole you have.... note my photo. 14 turns of wire took up the whole hole with a single layer.... more would have made it messy, and hard to get a bolt through to hold the thing down.”"



I still have enough space to get a bolt down the middle of my toroid, and I read this........
 "Remember, a turn is counted as a wire going through the hole. Don't care how it ends up, but has to be through the hole... that's where the magic happens. So this takes away the confusion of half turns at the start or finish.... if it goes through the hole, it counts as 1 turn. If you only count the wires in the core, you will be fine."

And a photo of another 'oztules' torroid with more turns and some overwrapping through the centre hole...


So I think my toroid will be okay.? 

I also read this...... " Don't epoxy the  75mmsq cable... not necessary in any way. It is both silent because of the sheath, and vibration is not a problem for the same reason."
So no need for me to wrap and lock those 14 turns, although I just recently obtained some spare Potting compound used in underground Big cable junction boxes.
At the same time I obtained some 50mm/2 straight through connectors and I solder those.

 My first Pic of my toroid above, shows a joint on the last turn of my 14 turns, as although I started winding half way I ended up with half a turn out. 

I will use 50mm/2 Tri rated/car battery/welder flexible cable for the bit from the toroid to the PJ Boards.

[clockmanFRA]

"Whats the voltage on the primary,  now with 240v on secondary??"

I have just done the test.

220vac and light bulb on the 118 turns x 4off and 165.9vac going in the secondary.



And meter attached to the 14 turns, primary is 19.65vac.

[oztules]

Once you get the thing going, short the light out so you can drive the secondary with the full 240/220v. You need the light to get it going, after that it will draw almost nothing direct coupled to the mains...... very doubtful you will get it going without the globe in series though... inrush current is huge for these things anywhere near zero crossing conditions.

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

[clockmanFRA]

Yes, thanks for that 'oztules'.

'IT LIVES, IT lives I tell you'.... well once the secondary was taking the full 236vac, it was very very quietly softly humming.



The primary 14 turns was showing 28vac.



Hmm, Not sure if that's good or Bad?
 You did say " so 30 odd volts is what we want.... "

[oztules]

one more turn would be better, but it will be fine as it is.... you will do better with flat batteries ... you have more head room.

Your cores are slightly less cross section than mine from the looks. I would get 2.28v/turn, your  getting 2.0... so your actually running a 118x2=236v transformer.

I fully expected your cores to be bigger/better than my chinese ones... but not so.

The idle current will be interesting to see when you hook it up. It should be ideal for 220v use, as it will be further away from saturation then...... mine are 240v units, so you may have hit the right turns for 220v.

Next....... lets see what she'll do mister

Don't forget to use the ecore ferrite filter... you should do around 25-30 watts for a monster transformer and electronics draw..


......oztules

[niall2]

i,ve no idea technically whats going on here but its fascinating to watch .....good luck Clockman