Anotherpower.com Forum
Project Journals => User Journals => Oztules => Topic started by: oztules on April 26, 2015, 04:59:23 pm
-
It's time to put some practice down in a how to style rather than a Q and A thread.
To make a nice 6kw unit that will truly do the 6kw rather than the not so real figures bandied about by PJ, we need first and foremost a better transformer.
This is easily done in Australia by procuring "inspire" grid tie inverters off ebay for 50-80 dollars, and extracting the huge torroids from them. In this article we will assume you have procured two of them, and we will make a much better transformer than both of them put together, by making them a single core and then rewinding them.
This is a simple procedure, but one that requires patience and methodical action.... it is a true pain in the butt to do, but is simple, and a perfect foundation for your inverter.
The transformer you will create would cost around the $1800 region, if it was purchased in this country... this one I am winding will cost $120, and i will have two stainless steel boxes with ip55 rating as well left over, plus a range of very nice filters to quieten down the hash from the inverter if you want to.
Here are the two transformers after pulling them from the box, and knocking the center out of them... I use a smallmash hammer and a big drift to simply knock it straight through the center hole... you can drill it out if you wish like I did the first time... but it is both dangerous to the wiring and to you unless you have a good way of holding it very securely.... best just to knock it through... takes seconds, and no mess.
[attachimg=1]
Once that is achieved, it is time to look at what we have got. A huge torroid the like I have never seen before. The figures on the side say 130v:240v 50hz.
[attachimg=2]
The 240v winding is on the outside, and the 130v winding is on the inside.... there are several ways we can tackle this.
We can simply tape the two together, join the 240v windings in parallel, ignore the 130v windings, and wind a secondary around the two transformers, and were done.
This will work, but you will struggle to get decent wire through the center for your new primary... so not worthwhile.
Next option.... we can strip the two 240v windings off the outsides, and be left with the 130v windings on each. This is not so silly, as the 130v winding is 2 in hand, so if we stack them together, tape them up, we can series these windings and get a 2 in hand 260v winding, and a much bigger hole in the center. This is an ok arrangement, and will work well... and very easy to do.
This gives us a 260v secondary.... this is a good thing, as the reason we are doing this anyway, is to get the magnetizing currents down... otherwise we need only buy the Powerstar W7, and we get the same kind of honest power, but 5.5kwh/day losses because of the E I transformers they use... very lossy. It's that simple, if they used big decent torroids, we would not need to change anything.
So if we can "overwind" a bit, we decrease the strength of the transformer field needed to run the thing, so we can get the idle currents down even further... ie for a given field, it is amp turns... so more turns, less amps..... and the magnetic field in the steel can be decreased as well, and this will mean the magnetising currents will be less.
It is likely that the original configuration will be close to saturation, and a look at the graphs for this will show a non-linear relationship with current and field strength as we get progressively closer to saturation.
So if we can back off from that point even a little bit, we will get a more than linear decrease in magnetising current.. so all good.
One draw back with this approach is that you have to live with the transformer noise. This may not be an issue if you locate it well away from your living environment, but can be a nuisance if you work next to it all day.
The second draw back, is that you can get rid of a pound or two of copper, if you rewind it as a single transformer from the start........ ie those wires that are pressed together between the cores where they meet, can be gotten rid of entirely. Each transformer can then loose 1/4 of it's resistance from copper loss, and that pushes your power handling up.
Saturation: When you have designed the transformer, and it runs fine, you cannot saturate if from overload if you have a sine wave source... contrary to popular opinion.
If you look at the equations, you need a frequency change or a voltage change to saturate the core... input current makes no difference.. nix... nill nada.
So providing your inverter does not clip, you should not saturate the core, no matter how hard you press it. All this extra current put into the system will be working against the wire fields as MMF and back MMF, the magnetising current remains the same... so I think the load losses are in the copper... so if we can get copper R down, we are well on the way to more power for longer. ( eddy currents and hysteresis will also play a part, but not much in a torroid comparatively.)
ie any transformer ( because we know the core does not saturate from over loads) can take large currents...so the output current is only limited by the heating and regulation ie can be massively overloaded for short periods. The voltage will sag in proportion to the copper losses, but provided it is released before it burns up, it will survive... but it won't saturate.... so.... a 100w tranny will stand say 10kw..... but only for maybe milliseconds before the copper fuses.. this is extreme, but you can see what we are driving at. Basically the difference between a 1kw and a 10kw transformer is how fast we can keep temp from rising beyond safe. ( alright, the voltage regulation will be total crap too )
So a 2kw transformer may handle 10kw fine if it is...... say .....oil cooled, and so becomes rated at 10kw. The grid tie ones are inside a sealed box, so are "under rated" so to speak because they have little chance to get rid of the heat, but must run all day at full power or for a fair bit of it, so although they are rated at 1.5kw inside the box at 100% duty cycle, once outside, and with most of the windings gone, they are very much increased in power handling ability.
Edit: I edited the above, but it is still yuk, the idea is there, the expression is not.. will edit it again soon.
Thats why the PJ can do the 8kw for a short period of time, sadly, the transformer gets hot quickly... but the cores will do it fine... the copper won't. PowerJack has taken this to extremes, and sometimes uses a single transformer with a fan instead of two transformers... obviously hoping for a very low duty cycle.
So after all that, the next possibility is to unwind it totally... and start from scratch.. with three cores used, we could loose a complete transformer worth of copper, with 2 cores we get rid of 2 sides of wire..... so our copper losses get less, and if we want we have the option of more copper as well, as the hole size remains the same, but the turns is decreased by half or 2/3rds ( depending on core count)
I have done all three styles. A complete rewind is best for noise and performance.. so the suffering is worthwhile.
So, moving right along....
Here we are with some of the tape coming off. We use a small former of some type to roll the plastic on. This is to be reused, and is an easy way to unwind, and rewind... and is boring as hell....
[attachimg=3]
And here we are with the plastic off the outer layer:
[attachimg=4]
Here we can see the unwinding stick. It is about 800mm of 6mm x 38mm steel, with notches in the end to take the wire. It must be fully insulated with a few layers of tape so that the steel cannot touch the enamel as you weave the thing back and forward through the core.
There is a right and a wrong way to use this, and if you look at the wire as the stick is passing from right to left, you can see that the wire being unwound is on top of the stick, this will stop horrible kinking, and make it very easy...... you'll be surprised how unfit you are until you try this for a few hours.
[attachimg=5]
Notice how kinked the wire is on the stick... once we get a layer off, we need to straighten it out.... first step is to hook it onto something solid....and give it s stretch..... thusly
[attachimg=6]
Delicate is not my strong point..... unroll your wire off your stick, and stretch it a bit to get the main kinks out of the way.. then use this next thing... simple but effective.
[attachimg=7]
It can take two passes on the inner windings, as their corner kinks are far more pronounced than the outer wires corner bends.......
[attachimg=8]
Then wind it up on a spare former like this.... and try to keep it very neat, so the wires do not cross, this will help to not introduce more ripples in the wire.
[attachimg=9]
Then it is a matter of just plugging away until you find the core itself.
..............oztules
-
So we have now found our core and it looks like this:
[attachimg=1]
By the time you get it all off you will have over 7 lbs of wire there 2mmsq...
Now we strip off the last layer of plastic, and we get left with this after we pull off the glass layers from one end. There were a few manufacturers fo these for inspire, and they may be slightly different in how they terminate etc, but the cores will be very very similar
[attachimg=2]
So we get two of these cores with glass on one end only of each, so we can join the cores together with glass and epoxy like this.
[attachimg=3]
and then we simply rewind on two or three layers of tape that we have on our bobbins from the unwind... and we have this..... ready to start again.
[attachimg=4]
Well thats as far as I got today, will try to get some more done tomorrow.. it is the rewind time.
I think I will wind this with 114 turns on the secondary and around 14 turns on the primary.
......... oztules
-
Perfect timing
I've just last week ordered a set of 48v 15kw boards and (weather permitting) Wednesday lunchtime will be 'skip diving' round the scrap yards to find suitable transformers.
Must tap up a couple of old mates to get access to a decent crimp tool (and maybe even the cable and crimps) to handle the cable for the primary. I'm thinking 50^2mm as the (high quality) stuff I've used on my current system is good for over 200 amps.
Some detailed dimensions of the Inspire toroids would help in identifying suitable substitutes :)
Cheers
-
50mmsq will be fine for this application.. it is air cooled, and will rarely be driven for long periods where this amount of copper will be pushed.
dimensions:
outer diameter = 185mm
inner diameter = 85mm
hight of roll = 65mm
width of stack = 50mm
weight is 10.1kg with no wire, near 15kg complete.
The cross section is of interest only really.. although a big core hole helps immeasurably. The cross section can be calculated simply, and if you start stacking them, say smaller torroids 5 high may get the same results.... but the hole size may make winding them iffy.
The original power jack torroids would do 25% better if they were rewound and glued two transformers together too.
Apparently torroids are twice as easy to magnetize as EI transformers, so good old fashioned welding transformers may have iron good enough for maybe twice the magnetizing current. My testing bears this out too. A big EI transformer from the Hydro here runs at about twice the idle current as the torroids... it would be high quality.
..............oztules
-
Got some time to do a bit today.
I know this is a bit like watching paint dry, as the pics all look the same... but if your doing it for the first time, you probably want to see every stage... so here we go.
First we wind the secondary one in hand. There will be three of these... this is what the first one looks like. It is one in hand... even though it looks two. I followed their pattern of placing two wires side by aide rather than making each wire equidistant. This allows us to use the inter winding tape to advantage... ie the next layer has a depression to lay into.. two wires again.
[attachimg=1]
It took two reels of wire, so there is a join in there. It is helpful to put some tape/ texta/ whatever on the 45 degree positions, and calculate how many turns you should have by the time you get to each point. I wanted 114 turns so roughly 114/8=14.25 turns to each station. This is important if you want a tidy and easy to work with transformer. It means the second layer goes in easily between the first layers wires.
This next one is after the epoxy has been painted on all over. This sticks the whole mass together, but you don't want too much that is takes up any space... just glue everything together, don't let it run and gather anywhere.... space in valuable
You can see the tape under the epoxy I used as reference points for the 14 turns
[attachimg=2]
Now we put a layer or two of tape to make sure there will be no chance of wires touching the first winding. They would be almost the same potential, but we don't want these to have any chance of failure.. it looks like this, and we can see the outputs there too.
[attachimg=3]
and here
[attachimg=4]
And here is a pic of the second layer going on. Showing the former that we roll inside the center to pull the wire through smoothly. The wire is orientated so that it unrolls against the walls.. We can see how the winding lays between the first wires now.... very useful reference.
[attachimg=5]
here is the second layer finished
[attachimg=6]
Before you join the layers together in parallel, it is important you check that the wire turns are the same, and the phasing is correct.
So join one set of ends together as you would to parallel the windings, and put neutral on that, leave the other two wires open.
Now attach an active on one of the floating wires.........with a globe or similar in series as a fuse and soft start. Without the globe, you will blow any size fuse your house has..... it will blow any fuse if you get anywhere near zero crossing... once running it uses almost no power.
Turn it on. Now measure the input voltage across that live winding... then on the other winding that is loose at one end. Both voltages should be the same .. exactly the same.
Now measure between the active of the winding and the floating end of the other winding.. they should be the same potential, so the measurement should be zero volts.
If all good, then turn off and connect the windings in parallel for keeps ( twist and solder) ...... otherwise add or subtract turns until it is right. Minimum is 114 turns... don't subtract below that, add instead.
[attachimg=7]
For the benefit of Andy, I assume that the aerosharp has 1.6mm wire , not 2mm, so I wound this with 1.6 just to see.
It will require at least one more winding for power carrying. The aerosharp 3kw one I have has a bigger hole than the inspire, and that helps lots, hopefully yours does too.
It is much easier using the 1.6 rather than the 2mm stuff, but now there is an extra layer to do..... always something.....
So, more epoxy, then more wire tomorrow.... then the primary.
One interesting thing of using the epoxy, is that it helps the cooling, stops the noise, fills in any micro cracks in the second hand wire.... and also gets slimy when still runny. This allows you to fairly easily move the wires around on the plastic substrate very easily... for cosmetic purposes. So while the wire is tight and does not want to move in that picture, tomorrow when the epoxy goes on, it will be easy to tidy it up some... and hopefully make it symmetrical. This looks better... but importantly makes making the third layer the same as the second, as the wiring will simply fit in the gaps again, making it easy to keep the turns the same...
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... thats 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.
This will have to carry 8 amps for 6kw on each secondary wire. Should be no problem for 1.6 wire fairly well cooled. Thats at 100% duty cycle.
In this application, thats fairly unlikely, as the batteries will start to grizzle if we keep that up hour after hour, so full power will likely only be transient... maybe 30 mins max for 6kw-9kw cont... and even that is hard to see. This one will spend it's life at 500 watts average for 12 kwh/day, but should handle 10kw fairly comfortably for short bursts. For off grid, it seems 400 - 900 watts average is the normal usage. So providing your inverter can do very large bursts for say 15 mins, then it is unlikely you will ever feel like your off grid. So if the hot water is running, the microwave running, the toaster is going and the kettle goes on.... the inverter just runs as normal.... but we can see that these loads will not run forever. So providing our transient ability is huge ( 15-20kw), and our short term high power is in the 30 min range, it is highly unlikely that we will ever resort to overload shut down from temp, and the boss will never complain of not being able to vacuum while the washing is on and the water heating and boil the jug etc etc etc.
If you expect more, then another layer will give you better duty cycles. ...... ie 4 layers will mean that each secondary need only carry 6 amps for a 6kw load. 6 amps in a 1.6mm wire is pretty conservative, and only takes another hour to accomplish.... it's roll your own territory.
Your limiting factor will become how much copper you can get in the hole for the primary. This one will probably only get 50mmsq wire, but 6kw will be only 120 amps, and air cooled means that 50mmsq will be easily enough..
If anyone else has anything to add, please speak up, all this is from scratch, so I have had to invent every step. There are probably better ways of doing this stuff, and I'm all ears.
................oztules
-
Still unwinding the secondary. I measured the wire to be 1.8mm dia. including enamel. Its look like their bobin couldn't hold enough wire as there is a join. Keep up the great photos - much appreciated.
Andrew
-
Well.... different again. Interested in the hole size. So far the aerosharp I had (3kw) had larger hole than the inspire, same outer diameter, but higher profile to make up for the less width.
"still unwinding the secondary"..... are we having fun yet???
..........oztules
-
AeroSharp 2k5
The transformer was 15kg, I've taken off pretty much 3kg of secondary, if the primary is another 3kg that's leave the core at about 9kg. So yes, bit smaller. I suspect the inside dimension is a bit smaller than your core as well. We will know soon.
So i'm looking at 246volts /2 = 123 turns ?
I've haven't got any left over thread seal rolls (threw them out) nor empty solder rolls. I figured I might make some spools from 20mm conduit and 4mm plywood sides (use hole saw to cut them) - bolt through the middle
-
Hopefully your hole size is bigger than mine, and so a homebrew former will fit. What did you put the tape onto??... you must have miles of it running around the place without a former for that.
Yes, it looks like you will have 123 turns on the secondary, and 15 turns on the primary.
Doing a transformer rewind with these is quite a physical enterprise isn't it.
...............oztules
-
It really is :)
Admittedly I'm a slow worker, its taken 1 1/2 days of work just to get to where the secondary winding is off (and yes, other odd jobs). Yeah lots of mylar tape on the floor :) Would definitely be easier using a spool. I think the extra joins actually made it more work. I used a 3 plywood 800 x 35mm stick as unwinder former, duct tape on the V cuts each end. If some else it working one of these AeroSharps I'd recommend have 2 sticks available - I had to unwind one wire to continue unwinding another ::)
-
Ok This is where we have wrapped the first two windings up, and now ready for the last and final hv winding.
[attachimg=1]
And here is the final winding before epoxy
[attachimg=2]
and with epoxy
[attachimg=3]
Whew glad thats over for this one.
Now I we only have to wind the 14 turns of heavy wire and we are DONE :) :) :) :) :)
That may be tomorrow.
.............oztules
-
AeroSharp 2k5
More unwinding :)
Shield layer I presume. Do the inspires have this ?
[attachimg=1]
-
yep .. sheild.
No, the inspire has no shield thank goodness..... but my aerosharp 1.5's did.
now the primary.... hehehe...
Sadist Oztules
-
Okay, unwinding the 2nd layer of mylay tape, this time i'm using the bobbin when I should be.
I didn't have anything suitable so I made these. Seems to work, 20mm dia conduit x 20mm ish with 50mm round cardboard hot glued on.
[attachimg=1]
-
I found single width is best for this.. It will make a big difference for you now, without it, there is plastic all over the place..... and it makes rewinding easy .
............oztules
-
AeroSharp 2k5
I have blisters, sore shoulders, tired but the core is stripped of all its windings - just have to remove the final bit of tape.
I believe the core is exactly the same. 193mm , 86.23 hole, 68mm high - ancient bathroom scale weight ~ 10kg
You right john, the bobbin with the right size is best.
Now have to dismantle the other AeroSharp and do it again ::)
Andrew
-
Welcome to the club..... look at the final picture of my transformer for encouragement... it is worth it in the end.
The island seems to be scarce of heavy wire... so it looks like some very very hard to bend 40mmsq will have to do this time.... it is gonna be hell to wind.. stiff as a board.
.............oztules
-
You guys are animals. ;D Nice to watch, I must admit.
-
It certainly will be worth it - Keen as mustard :)
Miserable day in Brisbane today so a perfect day to pull my 2nd AeroSharp apart.
For all the bagging the AeroSharps have received I have to say I'm really pleased at the quality of everything inside the box - all good quality components, neat design, good assembly. I'm going to have lots of spare line filters, big volt caps, huuuuge heatsink and more :)
John, if I understand things right, by having a few more turns on the new transformer (as in 123 vs 114 turn) , will decrease the flux density, but will have no negative effect on the transformer action itself, just increased copper losses,reduced winding room ?
During my hours of unwinding I had a thought that i should have a tap on the primary for 24volt system, say at turn 7? (This transformer in destined for a new 48v system but my present system is 24v) may be useful for experimenting before I get my 48v system.
Andrew
-
"John, if I understand things right, by having a few more turns on the new transformer (as in 123 vs 114 turn) , will decrease the flux density, but will have no negative effect on the transformer action itself, just increased copper losses,reduced winding room ?"
Close enough explanation. less steel, more turns more copper loss.
If we think the idle current needs to be lessened further, then more turns on the primary and secondary .
This means for the same inductance, the core will be less magnetized.. so less iron loss... but more copper loss.... but at idle, the copper loss will be virtually nil, and if we get the iron loss down, then our standby losses will be less.... a case of deciding where to stop really.
So, if you expect that the thing will run at low wattage most of the time, then more turns is the answer.. but if you think it will be running high loads for longer time periods, you may decide copper loss needs the attention.... like everything else.. it is all balancing plates in the air.
Addition.... if you want to tap, perhaps wire in the primary in two half winds , rather than center tap. That way in parallel, you have 24v full power, and series for 48v full power.
...............oztules
-
Well had to use 30mmsq wire... hard as hell... might just as well have been solid core... it was very nicely work hardened.
After making lots of gutteral noises, we ended up with this.... thats after using the car and a 6 ton guillotine to stretch it straight....
[attachimg=1]
It will easy do the 6kw, as in open air 30mmsq it is good for 180amps.... and thats well over the 6kw we wanted.
Or at 6000w, we will have .66% loss in the copper primary, and even less in the secondary ( 40 watts lost@6000w in the primary and about 17w in the secondary ) .. more than I would like, but less than 60 watts is controllable and about 1% loss or less all up ... plus 25 watts lost in iron loss... 85w total from the 6000w so 1.4% losses total from the looks of it) The air spilled from the mosfets fan at that level will keep it easily cool enough. I suspect.
................oztules
-
nice ... :) look forward to reports on the testing.
Finally stripped both cores, I think I was just getting the knack too .. I'll join them tomorrow, then i have to leave it for a little while, meantime sort out the bearing wire straightener.
Andrew
-
You guys are animals. ;D Nice to watch, I must admit.
I'm with you all the way on this one Tom LOL
I decided with little 1.5" cores and like #16 solid that I really didn't care much for [un]winding toroids... This monster blows my mind... I envision plenty of cussing, lots of sweat, and the urge to throw things. ;D
Keep up the good work, Tom's right, this is fun to watch ;)
Steve
-
A quick mock up shows that the entire inverter will run at 30 watts idle current with no sneaky mitigation like the big boys do.. ie switching in and out small transformer for small loads, and big lossy ones for the bigger loads... so our iron loss is less than probably 20 watts no matter the load.... the other 10 watts is in the drivers and logic etc.
Pretty good for 60lbs of transformer I reckon.
Torroids really are the best, but very expensive steel, and a real hand full to wind. They can do twice the power of a good EI of the same size, and are half the weight for the same power.... thats means it would need a 120lb transformer of the EI type to catch up to this one.... something to think about when your winding... and imagine the idle current on a EI of 120 lbs.....
Steve....."and the urge to throw things".... when the transformer weighs so much I can hardly lift it up..... it slows down the urge to throw.... otherwise.... the thought had crossed my mind on a few occasions.
................oztules
-
LOL I hear ya there... hadn't even thought about the raw weight :O
Quick question on this, as it seems like a good spot to ask it... On the idle losses - Putting the little toroids in series with the EI having the same end result as using a toroid to begin with, I'm understanding that the mechanism is slightly different... the small toroid with EI acting as a HF filter, and less core loss at higher frequencies using just a toroid to begin with... Does this mean that there's more hash making it to the output that needs to be dealt with on the secondary side, or...?
Steve
-
The W7 powerstar has EI... and is a great performer... better than the PJ out of the box.... but a ferrite does not stop the 4-500 watt idle current.
I don't know why such a tiny choke makes such a 10 or more fold decrease in current.. but it works with torroids, and high quality iron cored EI. The EI is generally twice the current as the torroid ( twice the steel, twice the magnetic domains to switch back and forth).
But 13uh is not enough to explain very much, so it looks as if it is a switching thing ... don't know.
The hash is the same as far as I can tell, so the output is a composite of all the pwms that make it up... d class amp?
There is a 4uf cap across the output that shapes the wave to a decent waveform.
.................oztules
-
Andymack.
An idea for you here.
As you have the aero sharp 2.5, you will be able to do this:
[attachimg=1]
That actually has two full 1.5 boxes behind the facia. Once you remove those massive heat sinks, you can chop and bend and fiddle, bolt them together.....until you get something like this.
The power point is there not as the main output, but as an aux output if your working near the thing.
Once you get over a few KW, your best to direct wire, although there is a 25amps output on the unit that can be used for 4kw I guess.. It is a 4 pin amphol fitting I think.... I actually use it for the grid tie solar input..... which I woudn't do without.
The top box will house the 15kw power board and control card, and the bottom box will house the transformer ...... and possibly the solar controller for the 100 amps that the solar can put out .........
I guess that will be part 2
.................oztules
-
very smart looking :)
They're certainly good boxes. To fit the transformer into one box and the electronics in another box is a very worthwhile thought. I'm looking for more dead,cheap inverters and (beside the transformers) the boxes will no doubt be useful with various electronics I have around the property. I picked up a little 1kw transformer isolated grid connect unit and if nothing else I'm hoping to use the transformer as an isolation transformer. You started me onto a dangerous path ;D
-
It is done... the stuff it in part anyway... just some hooking up, and now it looks like this.
[attachimg=1]
and this
[attachimg=2]
Now you will have to get going...
The boxes are drop dead brilliant, and worth a few hundred dollars on their own....... all stainless and aluminum. Aero sharo have come in for some cririsism when they started over in Aust... but I suspect it was poor design and execution by the installers that was the culprit. They have odd voltage range, and stick meticulously to the aus design rules.. the german ones in particular seem not to. I heard an Inspector mention he had never had a Chinese unit fail the islanding tests, the name brand european ones fail the most. So they are the first to drop off because of line voltage or frequency problems.. as they should.
I have run a number of them for quite a few years now, and never a problem, and very well built too. The only mud that sticks from the comments I have read, is the ones with noisy transformers... and it's true. If you had one in the same room, you'd be unhappy, and you don't get that from the high frequency units..... I'll take low frequency every time... not all are noisy, but enough to earn them a bad name for that alone.
But you end up with lots of stainless nuts, bolts etc etc when you tear them down... great stuff.
....................oztules
-
An interesting idea has come up - the cores I've found so far are not really big enough either with the size of the hole in the middle or the overall cross sectional area. How about rewinding the steel of the core itself and incorporating 2 of the 1" cores into one, starting with a larger inner diameter and then stacking 3 or 4 of these.
I reckon a jig that uses a sprung loaded bearing pushing down onto the steel as its wound to make it tight and the occasional splash of polyurethane varnish should stop it buzzing. A lot depends on whether the steel in the cores is even the correct spec - any ideas on how to test the basic characteristics of the cores?
Is this a really silly idea or what!
I'll retreat back into my cave now...
-
Worth a shot. If you have a lathe handy, it will be easy of you can get the steel to unroll form one former to the next, tension from a bearing bar will make it pretty tight wrapped. I expect the steel to be fine.
The funny thing is, as we make the hole bigger, the same amount of steel will have much less cross section, so a few cores will be needed... but at that price. nothing to loose.....everything to gain.
As CT's they would have to be fairly decent, or the eddy currents would skew the figures a lot more than needed.... maybe... possibly....
.....oztules
-
Hi Oztules I don't know if this is the right place to put this but its the only way I can get some info from you. The last time I talked to you was to tell about the inverter on light on 24/7 remember. I got on to power jacks and waited a long time what with the holiday in China in February, In the end They said I needed a full set of boards this will most likely cure the trouble for £100.00, was this a good deal ??, as it took so long to come in the end I traced the fault to the control board with the three leds on, the problem was very bad soldering shorting out the control transistor, ah well live and learn.
I'm now trying to run off grid and am having trouble with over voltage tripping out the PJ as the sun get a bit higher in the sky, hoping you can come up with a method of controlling this, thanks if you can.
Do you remember we were doing tests on my 8000 PJ, we put a link wire in one of the plug tops linking out n to earth as my sonny boy inverter would not start, I have since modified this with a 1100 ohm resistor, this seems to stop the E Trip going out at times. Thank for all your help M8y, Steve.
-
What I do...
There is a resistor near the main chip. It has the words cut if high (Ithink) will check again later today ... anyway it is a normal 1/4w resistor through hole... of a few megs from memory.
I use a 5v6 zener in parallel with this resistor, as it is just a piggyback and easy to do.
Picture of the resistor is here #59 http://www.anotherpower.com/board/index.php/topic,902.45.html
....oztules
-
Hi Oztules that sounds a simple mod, just a couple of points,( should the resistor and the zener diode be changed) ( is the diode looking down towards the big chip. How do you find out all this stuff M8y. Many thanks. Steve.
-
Hectic here at the moment steve... sorry...
I just piggyback the zenner across the resistor. No reason to try to get it out of the double sided board amongst those fine tracks.. nothing to gain really. It is a 10meg resistor on this board
[attach=1]
I find it out because I need to .... mother of invention
I expect it cripples the voltage sense for charging with the unit unattended.... but I don't use that function, so never bothered to find out... lazy...
........oztules
-
Hi Oz I've said it before and I will say it again, your a Hero and a Gentleman, that's a much better picture for my old eyes, will let you know how it works, many thanks, Steve.
-
Hi Oz It was hard to find a 5.6v zener as it was a holiday over here, but in the crap box I found one, popped it in, things seemed more stable when off grid charging, as the batteries came up so did the voltage, when it got to 31v the PJ made a bleep, the voltage went up more to 32.3 at which point I lost my nerve and switched off, was this mod supposed to stop things going over the top ?, any more help John, If you can crack this one for me I can stay off the grid a lot more, and that the hole aim of things isn't it, many thanks M8y. Steve.
-
No. This mod ONLY stops the PJ from turning the power system off when ever it gets nearly charged from an external pwm or mppt controller charging the batteries.
It does not have any effect on the charging of the batteries by whatever method you choose.... and it will render the internal charging by the pj itself as dumb, not intelligent.
If you use the PJ as a charger, this is not for you. ( and it would not have caused any trouble anyway without the mod)
If you are charging from some other source ( what I thought you did ) it will have no effect on charging, but will stop the annoying OV shut down.... I note that the latest board I bought has no bad effects at this time, and requires no fix... good to see.
If you explain exactly what and how your charging with, we can help.
From your comments thus far, I suspect it is using a grid connect inverter, reverse charging.
This brings it's own set of problems to deal with, and the PJ has no part in this.... ie you don't want the PJ to turn off... just to turn off the grid tie inverter.
So the zenner stops the PJ from turning off under any circumstances... what we need for uninterrupted power, but we must control the over voltage some other way.
If using a grid tie, then you need a voltage switch to shut the grid tie off when V= whatever you are comfortable with... 29v@7 amps perhaps.... but you can't expect the PJ to do this for you.... even a windmill style dump load wold be fine.
As a generality, the grid tie will shut off from over voltage AC when the batteries can take no more, and shut themselves down. For flooded cells this can work ok. This is because when the batteries can take no more heavy current, the power from the grid tie must go somewhere..... so it pushes up the AC voltage... and when it gets to ( whatever your design rules say), it will shut off... primitive but works with properly programmed and islanded inverters. The european ones fail this regularly, so are probably going to cause problems, the chinese ones seem to follow the rules more carefully.. SMA will cause all sorts of problems in this way..
So your best to make up your own OV switch, that interrupts the grid tie by battery voltage relay ( op amp and a relay).... or a very large dump load.
Before it sounds like you were relying on the pj turning off.... which is a no no.... if your relying on it to drive the house at all.... so you don't want to be doing it that way .
If you don't care if it turns off, then remove the zenner.... then the PJ will act as a voltage switch... but it will be useless as a steady load provider.... always something.... but it is not PJ's problem really, and needs another solution from the start.
It is interesting that the PJ could still see the 31v as too high, but now does not act on that information... interesting... maybe you could use that as the voltage switch signal????... ie any time it beeps, shut down grid tie relay.... ...... not so silly now I think about it,,,,
........oztules
-
Hi Oz thanks for taking the time to sort things out for me, normally I run on the grid and use PV power via a big battery charger to fill up the batteries. When I run off grid which I like very much, The PJ as you know turns on the Sonny boy inverter in the loft, I then charge through the PJ which is very good way of charging Im not going through so many things not so many losses, I'm thinking the way to go maybe, your mod with the zenner moves the tripping point higher I'm thinking, If you could move the tripping point lower to say 29.6v (my top charging voltage), maybe make it more stable switching ?, using this tripping point maybe it could be done without relays, When it gets to the top level I don't mind it tripping out the system, I would then have to go back on the grid and finish charging. Oz I will be guided by what you say is best. Steve.
-
You'll have to make your own bed to lie in here, as you have a problem of mismatching with no attenuator of any kind... only bang bang type results with what you currently have.
To make this work as you envisage, you would simply need to control the mppt point of the SMA, and all the problems would be gone.
SMA does this by varying the frequency a bit from the inverter to the grid tie inverter.. but they are matched ( and don't always play nicely either ):)
In your case, unless you want to make a voltage attenuating device between your panels and your grid tie, controlled bby the battery voltage, then I'm afraid you back to bang bang... unless you have a swimming pool handy that you want to heat up.
The swimming pool idea is the best, or even a few thousand liters of tank water. Then you simply use a 5kw water heater, and a triac control ( 5kw light dimmer?? thing) controlled by the batt voltage, and you have a 5kw dump load. This would work seamlessly... but you need the "heat sink" available... ( a few in line water heaters from dud dish washers would do fine).
This would eb simple, and very effective to control batt V.
If you can't control the bottom end like that, then you need to control the input voltage/current to the grid tie inverter... much more difficult, as it will need to be a 5kw pwm@up to 4-600v.
Those are the only real seamless ways I can see at the moment.
Other than that your in bang bang territory. If you don't mind power interruptions, then remove the zenner, and let the pj die, or use a comparator to shut down the grid tie via relay in the ac feed line from the grid tie to the inverter.
I don't see any other useful way to do this, maybe Rossw or some of the other folks here have a thought on this.
.......oztules
-
Thanks for trying Oz, I will play and see if I can work something out, will let you know. Steve.
-
I don't see any other useful way to do this, maybe Rossw or some of the other folks here have a thought on this.
Sorry, Oz, not much to offer here.
Unlike you (who threw brains at the problem), I was time-poor and threw dollars at it instead.
When I built, I had a Power Systems Australia RAPS5 inverter. It was expensive, but it did everything I needed out of the box.
When it let the magic smoke out 10 years later, I was deep in a hole and needed something seriously overnight. Again, rather than brains, I brute-forced it with dollars. A Selectronic SP PRO 481 is now doing the job (and doing it very well). http://www.selectronic.com.au/brochure/BR0007_03%20SP%20PRO%20Series%20II%20Data%20Sheet%20Web.pdf (http://www.selectronic.com.au/brochure/BR0007_03%20SP%20PRO%20Series%20II%20Data%20Sheet%20Web.pdf>Specs here). It's far more expensive than using your brains and some effort, and if circumstances were different I may have done things differently, but they weren't.
So... after all the words, my comment is "I haven't played with the PJ and really am not qualified to make any suggestions!"
-
May I suggest you use the carcass and transformer from the first unit, and add a 15kw powerjack card and controller as a cheap backup unit.....and you get to play along too:)... http://www.ebay.com.au/itm/171760626915?_trksid=p2055119.m1438.l2649&var=470713282062&ssPageName=STRK%3AMEBIDX%3AIT
The $300 is not a fortune to avoid another hole.... make no mistake, the selectronics are very good units, but there have been early failures over here for them too.... Their back up is commendable, but comes at a time price.... particularly here...... (and when you start flying 50kg units around the place, it has a fiscal price too)
Then you will be able to compare them as well. I use a neighbors selectronic as a comparison unit... you may be surprised at what these cheapies can do.
.............oztules
-
An interesting idea has come up - the cores I've found so far are not really big enough either with the size of the hole in the middle or the overall cross sectional area. How about rewinding the steel of the core itself and incorporating 2 of the 1" cores into one, starting with a larger inner diameter and then stacking 3 or 4 of these.
Still looking at this idea, fettled up the figures a bit having found a website calculator for winding stuff onto a roll. Punching in the inner and outer diameters of my 1" thick rings and an arbitrary thickness it comes out with a length of material required. Putting in that length figure times 2 and starting with a 90mm inner diameter I should end up with a ring with an outer diameter of 205mm. Stack 4 of them (i.e. material from 8 of the original rings) and I get a cross sectional area of ((205 - 90) / 2) * 102 = 5865sqmm and a weight of about 22kg. The material for the cores should be about NZ$20, the 10kg of 1.8mm wire was NZ$110 so only mylar tape and the primary to source now :)
Searching out an old trailer wheel hub and getting a CO2 bottle refill for the welder to make the winding jig.
Sounds to me as though a 6kw box isn't far away! Ah - got to find a box to put it all in!!
-
Now thats building a transformer... makes my effort seem a little less old school for sure....
Can't help but feel a lathe would do well in this application, rather than a wheel bearing arrangement... any one handy got one??
.......oztules
-
hi oz i have started up a new thread about modding a couple of powerjacks and could use your help if you don't mind
cheers :D
-
A Big Thank You to oztules, for getting me this far. :)
My Oz' Inverter gently came to LIFE yesterday evening with a bench test down at my Power Station Building/Shed where my 1100ah 48v 2 tons of batteries and control equipment live.
Here's a photo with 4.8kW of load after 30 minutes, the meter is showing battery voltage.
[attachimg=1]
Load reads 230vac, No load is 235vac, HZ at 50HZ.
Primary wires get warm see Temp gauge, the closest heat sink in the photo gets the warmest see temp.
The PJ Fan comes on and off at the correct time.
The Ecore/choke gently hums, yes its superglued, in fact it hums more than the toroid does.
Main cores are 2 off 190mm diameter, 90mm hole and both at 60mm thick giving me a 120mm thickness.
4 secondary windings of 1.8mm dia at 118 turns. Primary is 50/mm2 14 turns in the core.
Toroid total finished weight is 32kg.
15kW Power Jack boards.
On Off switch for the Boards on mine is, NOW CORRECTED, Pin 11 to Pin 13.
[attachimg=3]
Again a Big Thanks to oztules for his knowledge and help, and his kind words of encouragement, thanks.
-
[attachimg=2]Very good batteries.... still 50v after 30 mins of 100 amps is good in my world.
Only one thing to niggle...... for the sake of other folks, it would be useful to label the start pins as 11 and 13... not 8 and 10...... notice the numbers on the PCb are in the opposite direction to that you have used... ie 20 is closest to the main chip, and 1 is near the power supply... I know it is pickky, but it may cause a confusion for someone some where some time.... Your post will be referenced elsewhere on the net... and without the photo, would be a disaster.
OK that aside...very well done. 8)
The temp is pretty good for that power for that time with no real air movement near the tranny. I use a push and pull fan ( one either end) for the heat sinks, with a dedicated fan circuit... their fan never comes on........ I don't like their settings. My heat sinks hold at 30-35C for most of the heavy parts of the day ( hot water etc for a few hours plus normal house useage ). My fans are speed controlled , and at near 25c they turn on slowly, and remain silent until near 30c.... but this slower rpm range moves air around the interior of the case, and this helps to keep the whole thing cool, without actually hearing any fans until the temp gets to 30c. The spill cools the transfromer, which can get up to 50-55c under prolonged heavy use ( hours).
Your start jumpers scare the life out of me, as pin 11 is at full battery potential, and any mishap to other pins nearby ( except 12 and 13) may inject wicked voltage into places you would rather not see it.... at least 8 or more of those pins go direct to the main chip... and anything over 5v will be fatal to the control card.
To say I'm pleased to see your work running is an understatement... well done...
I still find it amazing just how good those cards work... and the tranny.
Now you know why we epoxied the tranny, it's quieter than your superglued and plastic suppressed ecore... pretty flash for near 5kw. Don't mind it in the same room even.
Just noticed you say 235v@60hz...... 60hz.... did you order the wrong control card??? or is that a typo. I thought 60hz was what you really didn't want.......
I'm impressed........oztules
A quickie I did for another fellow...
EDIT......this has been updated
[attachimg=1]
-
Yes , My Mistake, late nights. I have changed the photo to show PINS 11 and 13, and I will put some leads and a nice switch on those 2 pins.
Yes typo error, 60HZ on the Brain nowadays. I am 50 HZ, but noticed that my simple meter was struggling with HZ readings when that 4.8kW load was on.
But the HZ was still okay as the washing machine and microwave were running sweet as a button.
-
May need hf filter on the 240v outputs to get rid of all the switching noise.
For the meter to not track properly I assume it is an auto ranging meter... they are a pain if there are any harmonics around, or the fundamental is not very very clear.
The lack of shielding ( spread across the desks ) will allow a lot of RF radiation too at high power, and these frequencies will range up to and over 20khz... a big filter will help control this when it is in a metal box.... keeps the hash inside...
..............oztules
-
Normally I can not get on this AnotherPower Forum as the French Authorities, Orange, Nordnet etc, for some reason block access. So I am using a Proxy Server from USA.
Thanks about the HF filter, any recommendations please? .
Also good quality Cooling fans, ? as I am not sure about those 2 spare PJ's I have. I am fitting a temperature switch unit as well as leaving that PJ 3 wire fan connected in to keep the PJ control board happy.
http://www.ebay.co.uk/itm/261735347690?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT (http://www.ebay.co.uk/itm/261735347690?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT)
[attach=1]
I am a simple sole with not much knowledge of HF filters, except I did install a filter plug unit on my HiFi Amp.
I have a Metal enclosure 600mm high, 400mm wide and 300mm in depth, so everything will fit inside in a neat and tidy way. Although, at 40kgs I might have to weld up a angle bracket with struts, as the box fits to a wall.
As I am a newbie here Not sure about the next paragraph.........
Blast, Fieldlines is off line for nearly 2 days now, and I would like to correct the Pin 11 & Pin 13 mistake on my photo that you mentioned/spotted. Plus I normally copy and paste the topic into a Word document so I can get my construction booklet together for my eldest boy, he's studying RE technology at college, then he can carry this project on.....
.. Shhh, but really..... as I get older my brain seems to leak and I forget stuff.
The last post I copied and paste, was the 31st July the Cat one.
Thanks.
PS, Yes , you are right those PJ Power boards and control board sure are magic. Anything that you think needs modification before I install into my cabinet.?
-
Cooling fans...
That needs some discussion for these boards.
I use beecroft ball bearing fans 5". They are good value for money, and i like the ball bearings too. The following pics depict the fan I use, and how to put a 2k2 resistor onto the circuit board so we can get a pulse to replicate what the board expects for the tacho readings.
Without the tacho readings, the board will assume fan failure and shut down.
It won't do this straight away, it is when it looks for the fan, or the temp sensor tells it it is hot, then it looks for it.... if it cant find it, we stop.
So pic one ... the beecroft
[attachimg=1]
Now we remove the stick on disk on the rear, and we find the circuit board under neath... a careful solder with the 2k2 resistor and shrink wrap, and we now have a 3 pin fan with plus, minus and tacho.
[attachimg=2]
This will allow those who just want to use the boards as they are supposed to be used........
But...
For the rebels.....
Next how to control it..
That image you posted of the temp switch is fine, but I prefer to use a gradual driver, not a bang bang..... I need to live/work in the same space too sometimes... and a fan on.... fan off does not do it for me... so we build this for a few cents.
[attachimg=3]
This has two pots, one for the ramp rate, and one for step off temperature setting... the temp one is marked temp, the ramp rate is the other... devilishly cunning really...
So we can bang bang it if we use no neg feedback/ramp, and as we add feedback we start to make a ramp up to the top speed.... so as the temp approaches the set point, the fan will start turning, and disturbing the still air, if thats enough then good... if not then the temp will rise, but so too will the fan speed, until it gets to full speed if it cant stop the temp going past the set point with lower speeds... its simple as..not very often are they running full speed.
I use two fans, and they have no tacho needed... so no conversion on them. They push pull the main heatsinks.
I use a tiny fan with the tacho wire to fool the board... it never needs to come on anyway.
It is wise to still use the temp sense they supply as is where is... so if your fan techniques fail for some reason, it will turn off from over temp and save the day.
We can get the 12v for the fans from the front two pin socket. It will run two fans comfortably.
Here are half a dozen being built for this and other projects.
[attachimg=4]
Just the pins and odds and ends to add.
The noise... This is the sort of thing we want.... a dual filter.
It has capacitor, torroid, capacitor, torroid, and capacitor.
[attachimg=5]
These are too small, but the ones from the aero sharp 3kw units are fine doubled up for 40 amps.
Plenty on ebay in various guises..... look for EMI filter. You may need a few to get the current rating. I see a few from china 20A fir 10 AUD... four fo these series parallel would be fine. They are single stage, so one after another will be ok, and doubled up in parallel to get the40amp rating we need.
Perhaps try these...
[attachimg=7]
....................oztules
PS somehow the pics are all queer at the end ... I can't get the pic to display properly
-
Thanks oztules for the information.
For the time being I will try my bang bang arrangement on the two new fans in series with that adjustable settable temp meter of mine. I have a separate independent 24v supply in my enclosure.
I like your fan speed to heat rise arrangement.
Filters, thanks for that, not sure on backfeeding implications with the filters when I am getting power from the AC Coupling?
I seem to have made a write old mess on the Fieldlines Forum, everything is now out of time sync, as its not possible to re-edit ones posts after 15 minutes.......
-
mess on the Fieldlines Forum.
It's been a mess for a long time. Incompetence drove most of it, lost posts, scrambled posts, admins that couldn't admin, owners that didn't care... I'm surprised it's still running TBH.
-
Hi oztules,
I have been thinking about the cooling fans and the situation in my Enclosure box.
I believe your circuit that regulates the speed of the fans to the temperature is going to be a lot more gentle on things and be more effective in the long term.
However, looking at your 6 circuit boards, I am starting to drool........ the components bits I think I can get, but I could do with a bit more info on the TR that drives the Fans please?. And what heat sensors are you using? I could rob one from those 3v temperature gauges I got?
Will your PCB be available,? ...........my PCB making stuff, just looked, is dodgy to say the least, and goodness knows where my mask/stop material has gone.
Probably revert to breadboard......
-
I used IFRB4110 fet as i had a heap of them next to me, but any fet will do, small heat sink wont hurt but not mandatory.
I don't sell anything, but can send you one if you need to do that, you have suffered enough... however, you are better off just bread boarding one, or using Kinsten board, already has the positive emulsion coating, and then only needs a mask.
I used to use transparency and laser printer, but never could get modern printers to work properly with the transparency... now use a cheap $40 inkjet, print to A4 normal copy paper...... then olive oil spread over the paper to make it transparent, and then place on Kinsten, put under flouro desk lamp for 25 mins at 3" distance from board to flouro.
Then use NaOh or caustic sode or lye or whatever they call it over there in France to develop the pattern... will need a fine brush or sponge to help wipe the oil and sacrificial emulsion off.
I now use HCl ( hydrochloric acid , myratic acid, brick cleaning stuff, plumbers soldering fluid or swimming pool ph adjuster etc.... ) mixed with copper sulphate.
Not as messy as ferric chloride, just need oxidation between uses to revitalise it... very cheap and easy to get hold of.... chemist or plant nursery will have the copper sulphate. ( I got 25kg sack free from next door farmer... last 100 lifetimes)
The controlboard is very much superior from every aspect. I strongly recommend it. It will keep things cooler quieter, and everything will last longer... for about a dollars worth of components.
Note on fans.... blowing onto heat sink is infinately more effective then just evacuating the box with fans, Air directed onto a warm surface is much much much better, than an exhaust fan... so use push pull ... one to push air directly into the cavity between heat sinks, and one at the other end trying to create a low pressure area at the far end of the heat sink so the pusher has the best chance of keeping a decent velocity all the way through the center. We want to cool the capacitors just as much or more than the fets. Fets will stand a much higher temp than will caps for long life....so we keep the caps as cool as possible too. Thats where this control card really does it's stuff. It will silently try to keep the temp lower than the step off point the PJ would start.
Hopefully the PJ fan will not start at all unless we are squeezing 3kw or more for extended run times.... even then unlikely if you push pull. I have not seen over 35c yet, so the PJ has never started it's little 40mm fan I use to satisfy the controller.
...................oztules
-
Thanks 'oztules', as this is just a one off, for the time being! gulp....... bread board it is.
Sometime today I will disappear into my workshop and dive into all my electronics stuff, racks and racks of stuff, untouched for 20 odd years, and I could be gone days............... :D and see what I have.......
Thanks about your wise words about the cooling, push pull it is on those PJ Heat sinks.......
What do you use as the temp probe from your circuit? could I use those on these ?..........
[attach=1]
http://www.ebay.co.uk/itm/251637079552?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT (http://www.ebay.co.uk/itm/251637079552?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT)
Bare with me 'oztules,' for me it's all Virgin territory again, thanks for your understanding.
-
Yes they are fine, and so are these..
http://www.ebay.com.au/itm/10pcs-Thermistor-Temperature-Sensor-NTC-MF52-103-3435-10K-ohm-5-DIY-New-/191204294415?hash=item2c84a9ff0f
I buy them by the hundred for only a few dollars............. if you look hard... ntc 10k@25c... will work fine.
It was only a quick circuit to test the fans... but seems to work fine.
The first pot ( that says temp) sets temp, and the second one sets the negative feedback, and so makes the set point very fuzzy, and acts as a ramp for the temp. As we increase the pot, the feedback decreases, and so the set point becomes sharper. Full resistance will get it to near bang bang territory, and low resistance setting will make it a very sloppy set point, effectively making it ramp up to the set point, as hard or as soft as you want it to be. A higher R pot will increase the sharpenss further if you want.
................oztules
-
hi Oz,
i have been flat out lately but finally have some time to do some work on my inverters. I will update my thread tonight or maybe in the morning.
just have a quick question for you. would it be possible for me to get a copy of the protel file for the fan controller board?
cheers and thanks for the help.