 Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me. |
Forum Index : Electronics : Calculations for Toroidal Inverter Core
| Author | Message | ||||
| Solar Mike Guru  Joined: 08/02/2015 Location: New ZealandPosts: 1138 |
Hi all Have managed to locate an old variac to use as a potential inverter core, I pulled it apart to the point where valid measurements of core size and windings count, if its no good then it can be re-assembled and put to use as a variac again. 
Variac transformer: 230 vac input 0 to 270vac output @ 28amps, 7500VA Overall width 3000mm dia lamination Ring = 60 * 60mm = 3600 mm^2 Centre Hole = 160mm dia including existing primary Existing 230v winding = 230 turns 12g wire, 3.3mm^2 and exactly 1 turn/volt I note this does not match the turns calc that Oztules has outlined where mm^2/2800 is used to get the volts\turn, if I use this the secondary works out 3600/2800 = 1.285, so for 230 volts = 230/1.285 = 179 turns rather than whats there now 230t. It seems the designers of this variac are being very conservative and using a low flux density thus the higher number turns, that 28 amp rating is continuous according to the spec. Should I use the 179t in the calcs below or stay with whats existing, Its easy to unwind wire... I want to create a 6kw output inverter, hopefully this core will do it, the large centre hole should allow good thick wire. As my Batteries are Lifepo4, 50v x 800 AH, voltage will never drop below 50v no matter what the load, so I have used 50v for the primary, should I be using a lower figure of 30 volts mentioned by Oztules, noting the voltage will hardly sag at any load, which isnt the case with lead acid. This will affect the calc below. Proposal is for 230vac output @25amps from 50v input approx @120amps around 6 kw continuous Secondary: Add additional 230 turns 3 in hand of 2.0mm dia wires and parallel with existing winding, this will use 73m per winding, giving a total resistance = 0.095 ohms for the combined secondary. 25 amps in the secondary = 2.37 volts drop = 60 watts heat. Primary: 50 volts @ 120 amps, each turn length = 40cm * 50t = 20 meters If flat copper wire 10 x 3mm is used 2 in hand = 60mm^2, total R = 0.006, at 120amps = 0.72 volts drop and 108 watts heat Copper loss = 108 + 60 watts = 168, not sure what the iron core loss will be. Comments please, do these figures look right, or perhaps a lower primary voltage should be used and the lower turns ratio ? Cheers Mike |
||||
yahoo2 Guru Joined: 05/04/2011 Location: AustraliaPosts: 1166 |
I think the lower primary voltage is a calculated number. converting peak voltage to rmsV or something like that. I cant look it up, short of time this morning. But from memory it is rms voltage divided by peak voltage AC eg 230/325 = 0.707 then multiply that by battery voltage, as you say, at its minimum working voltage. I think there was some allowance for headroom so that the inverter can adjust the voltage higher under load and compensate for voltage slump. not sure if that is factored into this number or the target voltage is raised to 243-248 volts. I am sure someone has the official explanation at their fingertips, gotta go, chow! I'm confused, no wait... maybe I'm not... |
||||
Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
If you are building your inverter with the EG8010 chip you need to calculate the primary using 26v, if you are buying a premade PCB then I think you need to aim for 28V. The voltage is lower as what is fed into the transformer is PWM wave form and not straight battery voltage. There are only 10 types of people in the world: those who understand binary, and those who don't. |
||||
| Solar Mike Guru Joined: 08/02/2015 Location: New ZealandPosts: 1138 |
Thanks Madness, makes sense, I have purchased the book + PCBS from Clockmanfr, waiting for them to arrive, which no doubt would explain everything. So 26V it is, I will do a re-calc and see what the numbers are. It is a shame to destroy a good high power variac, but other sourced toroidal cores are not to be found in NZ and shipping costs from China or Ebay are almost double the purchase price so its an expensive option to import, as there seems to be no sea freight options. NZ manufactures dont seem to be keen to make any, out of four I sent emails to, only one has bothered to reply saying they dont make them, perhaps I will try phoning them. |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
Check with Clockman, he has some options, also look on Alibaba.com, I was going to order from there till I found a supply of dead Aero Sharp Inverters. I enquired on Alibaba and have had one supplier offering free samples. Also would be worth talking to all Solar suppliers in the area. If you were not so far away I would swap some toroids for that variac. There are only 10 types of people in the world: those who understand binary, and those who don't. |
||||
oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
Yahoo has summed it up for the most part. The PJ has better dynamic range to deal with low voltage in the battery lines, so 30v is fine. Testing shows that the 8010 is better around the 28v mark, and some commercial sites using the egs002 claim 26v for 48v systems. For 50v min, then probably get away wih 28-29v primary. Mine using the 8010 runs each day from 48 to 57... most of the day near 56-57 unless I give it a bit of zip after a few woeful days, and it will spend 4 hours up in the 59-61v range. The auto transformer looks like it is wound for volt steps so provided it is more than the 170 mark, the saturation will be far enough away to not matter, and 230 turns will make it very efficient with magnetzing current.... it may also be different grade steel... I don't know from here. Do a turns test to make sure the theory matches the facts... It is a big core, and if you find that 230 turns is way over the top, but you have ample room for the winding thickness your looking at, then go with that, as the magnetising current will be less, so idle current will be very low. They have excellent cooling with that "auto transformer", as i don't recommend trying to run 28amps through 2mm wire without getting pretty darn hot in a normal format. .......oztules Village idiot...or... just another hack out of his depth |
||||
| Solar Mike Guru Joined: 08/02/2015 Location: New ZealandPosts: 1138 |
Thanks Oztules, yes that existing secondary winding is too light, so will add 2 more layers of 230 turns of 2.0mm wire. Nothing is that simple thougth, contacted 6 motor re-winders in Wellington and Auckland to get some wire... no one has any over 2mm in stock, one did, but quoted me $50/kg for the 7 kilos needed (he sounds like my dentist). The one wholesaler willing to sell any has a min order amount of 45kg at the more realistic price of $19/kg. So my question to the forum is, does anyone know where to purchase copper wire in NZ without being robbed blind. |
||||
| Solar Mike Guru Joined: 08/02/2015 Location: New ZealandPosts: 1138 |
Persistence pays off, have located a couple of good sources for winding wire in NZ and not too expensive. Winding and Insulation Supplies CASA Modular Systems Casa are surplus dealers and have lots of smaller gauge wire 1.6mm and below and a huge roll of 10x3mm flat cable for primary windings. Back to that variac core, after powering it up with a 2KW jug in series with the existing primary the input current was very high, added some extra turns and graphed the results, very interesting... Looks like for 240 volts 320 turns are required for a current of 600ma, not sure what it should be, the thing weights about 24kg.  |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
Thats just all wrong.... idling current in original mode is over 500w.... the steel cannot be any where near that bad... I think it must have some shorted turns in there somewhere, the figures are very very poor... and it looks like it was designed and built for better results than that..... something is afoot..... just wrong. You have 100 turn on there that are known to be sound, but even driving it with say 100v from elsewhere will still be bad, as the wiring in the original seems to be ... well suspect. Perhaps check the winding for consistency ( very hard to do I suspect), or bite the bullet and strip and start again... I just find it impossible to believe that it should draw over half a kilowatt just to idle in it's original guise. ..........oztules Village idiot...or... just another hack out of his depth |
||||
| Solar Mike Guru Joined: 08/02/2015 Location: New ZealandPosts: 1138 |
The turns are all exposed as per the original photo, just a single layer, no obvious shorts anywhere, as you say something seems very wrong, I wasnt going to use the original anyway as would prefer to wind parallel strands of fresh wire of bigger area, so will strip them off and quickly wind on a single strand of thin wire and repeat the test. |
||||
Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
If it was mine then I would just wind on 10 turns of any odd piece of wire on top what's there. Then power it up, ignoring the high idle current (as long it does not blow the fuse  .Then measure the AC voltage on that 10 turns and use that result to get your turns ratio. Then strip the old wire (which most likely has wiper brush residue between the exposed parts of the winding) and wind on your calculated turns for your chosen mains voltage. Klaus |
||||
| Solar Mike Guru Joined: 08/02/2015 Location: New ZealandPosts: 1138 |
I have stripped off the old wire, was wound 1:1 ie 1 turn per volt, this maps to the one volt graduations on the dial, it did have a bit of gunk where the carbon brush tracked. Apply mains volts to it and a ten turn winding gives 10 volts as expected. However after cleaning it all up, putting a temp layer of masking tape for insulation, then winding a tapped test winding to measure the currents at various ratios the graph is almost identical, so obviously the steel is either a lower quality or because of the large magnetic path and big 18cm centre hole the losses are higher. Looks like if I use a 1.4 turns per volt the losses are lowest.  |
||||
| Solar Mike Guru Joined: 08/02/2015 Location: New ZealandPosts: 1138 |
My build continues... Have decided to not use the old core for now as its too lossey, approx 350 turns would be required on the 240ac side to be efficient and will require more wire than I currently have; so have purchased a new modern core from the manufacture in Christchurch, 240mm dia x 120mm hole x 100mm height, they annealed it after forming the core at 700 odd degrees and coated it in a high temp plastic. Its magnetic properties are so much better than the old one, wound a quick 150 turn 0.5mm test winding and ran a turns\volt saturation test at various tappings to see what the magnetising currents are, see graph, only draws 30mm at 120 turns, 240 Vac, so 2 volts per turn seems about right.  Decided to use 3 in hand of 2.36mm wire for the 240vac winding, stuff is quite hard to wind on, used a shuttle with pieces of plastic pipe as formers to lessen any sharp bends, 3rd hand is required to hold the previous wire in place, used a springy stainless ruler as a bendy clip, works quite well; photo shows first layer.  Mike |
||||
| Solar Mike Guru Joined: 08/02/2015 Location: New ZealandPosts: 1138 |
Is is necessary to place an earthed foil screen between the two main prim\sec sets of windings, cannot see it making any difference in this application as the switching noise will be present on the output anyway. Mike |
||||
| Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Very neat work Mike, I like it. I re used the static screen that the manufacturer put between the primary & secondary windings of the AeroSharp toroid transformers I have recycled in my inverter build. If one considers that a manufacturer of a commercial inverter would not include such a screen if it does nothing (after all, it costs money to put there)I re used it. In your case its a bit of a dilemma as you started from new and do not have such a screen foil. It has to be insulated, of course, to avoid a shorted turn and I would not know if its available to purchase where you are. Perhaps the maker of your core might know since they deal with toroid transformers. Since this foil is earthed its an important safety feature by providing an added separation layer between the secondary at mains potential and the primary at battery bank potential. Klaus |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
Experience shows the foil is not necessary, or maybe even not beneficial. The Aerosharps have it in all of theirs sizes, the inspire inverters have none.... so I would not bother.... the safety angle is not a real concern for me.... you will have a few layers of mylar there and insulated wire on the secondary... and epoxy. If your using only enamel wire on the outer, you may sleep better with the shield, but for 1mm wall thickness welding wire... not so much. I note only those with this shield have had problems with their inverter... possible short turn would explain that too maybe...or it could be the capacitive coupling that may occur with earthing the neutral... may be a possibility.... whatever it is I don't do it. On the rule of thumb, your core would be 2.14v/turn ( 112 turns), so we were close, and well away from saturation... commercially, they would probably have gone for 105 turns  120 looks very good.(Normally I wind for closer to 260vac... so would have used 121 turns using the rule of thumb.) Can't get over how much better everyone else does this than me  , that looks truly professional........oztules Village idiot...or... just another hack out of his depth |
||||
| Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
That's interesting oztules, I did not know you wind yours without the shield. There's definitely no shorted shield turn in my toroid. I had powered up the secondary from 240V mains and if the shield had a short the idle current would have been different from the one I measured before winding that shield on. So, to test your capacitive coupling theory I will test it with the shield floating next time I fire it up (still waiting for parts). Klaus |
||||
| Solar Mike Guru Joined: 08/02/2015 Location: New ZealandPosts: 1138 |
Thanks for the comments, I won't use the static screen, seems unnecessary, no audio pre-amp stages here... Here's the core after the 3 in hand of 2.36mm secondary and light epoxy\kapton on each layer; 13.12 mm2 total. Hole in the middle will be 90mm after a good cover of mylar; using that Kapton tape as its way cheaper than the no adhesive mylar obtained in NZ. Will be a struggle to wind the 50v primary, 10x3mm flat wire, see how it goes... My hats off to others that manage double or triple cores, its hard work.  Mike |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
Double core is probably easier as there is less windings and just a little further to pass the wire through. However much heavier to pickup, certainly would make a very good door stop. There are only 10 types of people in the world: those who understand binary, and those who don't. |
||||
 Print this page |
