 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 : Building a new Inverter.
  Page 5 of 8   |
|||||
| Author | Message | ||||
yahoo2 Guru  Joined: 05/04/2011 Location: AustraliaPosts: 1166 |
Hi Billy, I am guessing here. It will be a 3 wire 12v brushless computer fan, yellow wire tacho to the centre pin probably 4000rpm and 120 mm. I dont know the rated amps. the pulse signal from the yellow wire would stop the inverter from shutting down. the powerjacks I have seen have one fan and it works v-hard to keep the heatsinks cool. I read in the energy matters 4m that someone got theirs running with a silent 90mm 1500rpm fan and others have had problems with 2500-3000rpm fans so it might be a suck it and see situation. Lucky I dont have a lithp with that last sentence!  I'm confused, no wait... maybe I'm not... |
||||
mason Regular Member  Joined: 07/11/2015 Location: CanadaPosts: 86 |
Thanks yahoo2, I thought they might of been 5v, yeah the lithp would be funny. Billy |
||||
| yahoo2 Guru Joined: 05/04/2011 Location: AustraliaPosts: 1166 |
I just had a look back to page 3, I see that Glenn had to wire a pull up resistor I assume it was a 10K resistor, from the +12v pin to the fan signal pin. I guess the original fan has one installed across the wires. I'm confused, no wait... maybe I'm not... |
||||
Gizmo Admin Group  Joined: 05/06/2004 Location: AustraliaPosts: 5078 |
Yeah not 100% sure I needed the pull up resistor. I was having a few weird shutdowns that I couldn't explain, and while testing the fan with a cro and a bench power supply, I noticed the output signal was very yuk, adding a 10k resistor cleaned it up. But I now realise many of these fans have a open collector output, so chances are the PJ board has a pull up resistor anyway. I ended up building one of Oz's fan controllers with a large fan to keep it cool, and use a smaller 3 wire fan to keep the PJ circuit happy. Glenn The best time to plant a tree was twenty years ago, the second best time is right now. JAQ |
||||
Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Sorry Glen for posting this here but I see Oztules is reading this forum and I have a question or two regarding the Inverter conversion discussed here since he mentioned a similar inverter in his fieldline articles. Anyway, I managed to do a very reasonable barter (with a local forum member) for a 3Kw Aerosharp grid tie inverter. I would like to make a LV inverter from that, as big as its 15kg (incl windings) toroid can handle. My battery bank is going to be 50V/ 200Ah LYP (presently connected as 25V/400Ah) to a smaller LV inverter. I have ordered the 8Kw PJ boards 48V-220V AC mentioned here. I also ordered the ferrite E core mentioned elsewhere here. The aerosharp specs were: 
The toroid specs: 
The 230V winding connected to the grid is on the inside, the 250V winding is outside and is 3 in hand of 1.5mm dia wire ea. I did a test by feeding exactly 230 VAC (via variac) to the 230V winding and measured 250VAC on the other winding and the picture below shows what's on a temporary 10 turn winding: 
Now my question is how to best proceed. Should I unwind the 250W winding altogether? Should I unwind just enough turns to make the exact same voltage as the 230V winding and then parallel these? The toroid hole now is 65mm and I could fit 45 turns of 8mm OD cable (120A)max. The long fieldlines messages suggest a lot less turns but they are for worst case LA batteries, what works best for my lithium batteries who will never go below 48V? Any other suggestions are welcome, no doubt I'll have more questions when the PJ boards arrive in a months time or so. thanks. Klaus |
||||
oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
For the best power handling we want the secondary to be solid.... so winding 18 or so turns onto the outside connected in series with the 230v winding should get you a very heavy 250v winding when paralleled with the original of 5 or 6 in hand... bigger than anything I have done thus far. The 250v will give you less magnetizing current, so better idle. 250/8=31.5v primary.... 31.5/1.12=28 turns for the primary... so maybe 35mmsq for that will fit? This will keep the sag out of the secondary, and the 35mmm will do fairly well, and is easier to cool than an internal winding. Is this what you were after? The limiting factor for a transformer is the copper loss heat and sag. Once the proper turn for the voltage and frequency have been established, using a sine wave you cannot saturate it by driving more current through it... not even a short circuit. The extra amp turns is not used to drive more field into the core ( which would saturate it), but is involved with the current transfer between the wires involved via induction. So in theory if we used super conductors, we could use fishing line width wire and have massive kw transformers in a very small package. There are many people who think shorting or overloading a transformer leads to saturation, and that is what kills it... but it is not so... it is the copper losses adding up that fries it in o/load conditions. It is useful to remember this, as using two cores as I have ( and even three) does not push up the power handling in itself because of core size, but allows thicker wire, that stops the voltage sag, and copper losses... this is what dictates the power handling and heating problems. If you use both primary and secondary winding off the tranny, you will probably have a better transformer than my twin core for that part..... then it is how much copper you can get into the primary will dictate the power handling ( duty cycle dependent). ............oztules ps my battery is normally between 50v and 58v. Most sunny days it is 58v even with 5kw loads. Village idiot...or... just another hack out of his depth |
||||
| Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Thanks oztules, excellent explanation. But something puzzles me (and as a retired electronic technician I ought to know this  ) where does the "8" as in 250/8= 31.5 primary come from? Perhaps you explained it elsewhere but that fieldlines post is rather long....
With my measured voltage of 1.16V/turn I would assume the 230V winding being close to 200turns and the low voltage primary requiring around 45 turns ???? I think I'm missing something here. I had another look at that massive toroid just now, unfortunately I can push only a 60mm diameter wood disk through its center, must have gotten the first measurement wrong. Drawing a 60mm circle and finding out how many 10mm circles fit inside (10mm = 25mm sq 200A multistrand cable as used for welding cable) will now only accommodate 27 turns max, leaving zero space for a mounting bolt. No chance of 28 turns of 35 mm sq., not sure where to source 35 mm sq cable either, the flexibility of multistrand does appeal.... So I'm now thinking to rewind the entire toroid 4 or 5 in hand with the 1.5 mm wire, a lot more work though but at least I could epoxy up each layer then. Pity if it comes to that but I prefer to do it very well and have plenty of time for this project. Further comments please. P.S. Using a variac via a 1:1 isolation transformer to test that toroid completely negated the requirement of requiring an incadescent lamp in series. All my house fuses are still intact 
Klaus Klaus |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
8:1.....Buried in amongst the 130 odd pages over at EMatters forum, we found that that was a useful rule of thumb. It takes into account losses, low battery, sag scenarios etc... and seems to work pretty well. 45 turns would probably clip under high power, low batt volts, from winding sag and lower ratio. If your going to rewind, there are some things I would try now... experience leads to more ideas..... The hole size is the problem, and another rule of thumb ( very rough) is that about 2800mmsq of steel per volt per turn seems to be a workable unit. From this, if you tear the torroid down to it's core, it will be a very useful thing to pull probably 10mm of inner core.... out. so your current core will be about 1.12 x 2800 for about 3136 mmsq... when you get it bare... check to see... adjust rule accordingly. I think from memory, the core is about 70mm high... so if we lose 700sq mm of core to get a 20mm bigger hole to play with, we will lose about 700/2800 or .25v/turn.... significant yes, but countered by having a huge winding hole that will take lots and lots more copper to compensate.... so probably .87v/turn or 270 turns instead of 205 turns for 230v winding. If your real keen, you could wind that steel onto the outside and get maybe .12v back, possibly worth the work. The big hole will allow decent bobbins to be used, making it easier and much more enjoyable work.... more turns but easier. Keep in mind that you should... ( but wont) have balanced copper in the secondary and primary re cross section..... ie if you used 3 in hand of 2mm wire for the secondary it would be 9.5sqmm of copper total. To balance the primary...... the primary should be 9.5x8 or about 75mmsq... not likely to happen, but gives you an idea of how strong the secondary will be... cooling on the primary is very good, so 35mmsq would be fair. 50mm better.... etc...... 75mmsq won't happen unless you are doing 24v. So many compromises it;s a wonder it works at all.... .........oztules Village idiot...or... just another hack out of his depth |
||||
| Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Another question to oztules if I may ask: Is there a circuit posted somewhere for that neat fan control in your inverter modification page here ? I see the PCB layout but as I brew my own PCB's a circuit would be nice to have. And, do you have a source for that neat 4 x DPM red digit display that is shown on your monster inverter front panel? Thanks Klaus |
||||
Grogster Admin Group Joined: 31/12/2012 Location: New ZealandPosts: 9308 |
@ Oztules: If I was to have my own custom toroid wound to save me all the hassle of unwinding, rewinding, what would your recommendations be? Here in NZ, we have several tranny winding companies, and I can just get one made to the right size, with a centre hole nice and big or small as is needed, without having to force things into a smaller centre hole etc. This means you can custom-build the "Perfect" toroid.  
You pay for this, yes, but I have looked around, and there are just none of those dead inverters here in NZ to be had - they must be easier to find in Australia then here. ...and you should only ever need to get one made... See last post of page-2 of this thread, for the price details. Making a toroid with a bigger centre hole for easier winding of the primary would alter the mathematics for the secondary winding and number of turns though, would it not? Any reply gratefully received, as you seem to be the toroid winding guru!  Smoke makes things work. When the smoke gets out, it stops! |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
Tinker, I tend not to do circuits, I just get the protel pcb program and design on the run...and then make the board. I am not an engineer, just a hobby hack, so I get the pin outs of the chip I want to use and just go.... not very usual I know, but I don't have circiuts of anything I make... I just build it... stuff it up, change a few things and then there it is.. There will have been a scribble here and there of part of some of the things as they are complicated sometimes ( pwm power supplies, etc), but just a scribble here and there to incorporate into the on screen design.... lazy or silly, not sure which... Just reverse engineer it if you want to see how it works, but it is just a few op amps and a fet. However, I think someone has done a circuit for you... possbly clockman on fieldlines or anotherpower Grogster, A fellow in NZ is making a torroid ( frackers ( aka....Robin down under)) out of current transfromers he found over there at a scrap yard..... he is unwinding the steel strip, and rewinding a heap of smaller torroids to get the mass he wants. You will find him on anotherpower and fieldlines... pm him to see where he is up to... but I think he said he found a 44 gallon drum full of the things at scrap cost.....very cheap way to go... he very probably has spare cores. I am doing a bit of this core reshaping too to get bigger holes to work with.... loose a few tenths of a volt of "volts per turn", but happy to scarifice that to get a 100mm hole to work with. As a rule of thumb, every 2800mmsq of cross section you end up with, gets you around the volt per turn from my experience. The energy matters site has details of commercial wound tranformers worked out by an engineer, and that would be the best person to ask. He got 3kw torroids wound for his project. Clockman got the cores from england and wound his own, also on fieldlines and anotherpower. Failing that, and to answer directly, get them to wind you a 240v winding for 3kw or more continuous, with as big a hole as they will do ( 100mm is ample, and 80 is difficult but doable)...... then you only have to wind the primary into that hole...expect it to be expensive. They are the experts, so will be able to tell you what you require really. If they are going to wind the primary as well, then a 30v:240v will work for the 48v version. Edit: as they are doing the work, a 260v secondary would be slightly better, with a 32v primary. As a guide, the W7 units use 40 x 1mmsq wires for the primary, but they are EI transformers. Not sure if that helps or confuses..... A 6kw commercial transformer would be very very expensive, and your 3kw would probably be more than enough to get by... ask Glen how his is going with his single core unit. You can push the 3kw Transformer very hard for shorter periods if necessary. The 15kw PJ should have near 3kw of transformer in there also... maybe a way to go to..depends which one they send... 2 or 3 transformers...who knows with them.... .............oztules Village idiot...or... just another hack out of his depth |
||||
Clockmanfr Guru Joined: 23/10/2015 Location: FrancePosts: 429 |
Hi Tinker, Here's Oztules circuit put on breadboard. Its compact but tidy. 
See fieldlines for component values etc. Everything is possible, just give me time. 3 HughP's 3.7m Wind T's (14 years). 5kW PV on 3 Trackers, (10 yrs). 21kW PV AC coupled SH GTI's. OzInverter created Grid. 1300ah 48v. |
||||
| Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Thank you for posting this Mr. Clockman. I sort of drew something similar on my scratch pad but could not read the component values on that PCB screenprint of oztules. Then I went to e-bay and searched digital temperature control switch and there are lots. I ordered 2 without a display, they have a DIL switch to preset the relay trip temperature. And at $4.45 posted there is no way a homebrew controller could compete. Lets see how these things perform, if unsatisfactory I'll fall back on oztules controller. Klaus |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
The only problem with those sort of controllers are that they are bang bang things. The best way to control temp in normal operating parameters ( low to medium power ) is with variable control, as the fans will run at slow speed, and be silent for a lot of the time, instead of on -off-on full bore..... I detest that in my work area, when a slow stream of air would prevent the temp rise in the first place. However, if your gpoing to push it hard, then the difference is mute. forgot.... the meters are from ebay. .............oztules Village idiot...or... just another hack out of his depth |
||||
| Gizmo Admin Group Joined: 05/06/2004 Location: AustraliaPosts: 5078 |
I built my fan controller based on Oz's controller, but using a 741 op amp, 20k pots and a 4k7(?) thermister. It works brilliantly, but took a bit of adjusting of the trimpots to get it working nice. The fan speeds up and slows down depending on load ( heat ), very smooth. I used some heated water in cups and a thermometer to adjust it, at 15c the fan is off, at 25c its slow and at 35c its full speed. The fan in my Latronics inverter is either off or on ( and loud ), very annoying, I'm tempted to fit a Oz fan controller circuit to it as well. I'll draw up a circuit later today if I remember. Glenn The best time to plant a tree was twenty years ago, the second best time is right now. JAQ |
||||
| Gizmo Admin Group Joined: 05/06/2004 Location: AustraliaPosts: 5078 |
Drew this up. Its really just Oz's circuit with some component changes. Next time I have the inverter off ( unlikely to happen soon, been running for weeks now without missing a beat ), I'll measure the values of the pots to give a starting point. 
The mosfet can be any power N channel fet, I usually source mine from retired computer UPS's. Glenn Edit: I may revise the component values later, I did remember one of the pots was pretty close to the end of its adjustment. The best time to plant a tree was twenty years ago, the second best time is right now. JAQ |
||||
| Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Thanks Glenn, brilliant! I did manage to draw oztules circuit from Clockmans sketch, you are right, very similar. What puzzles me though is that oztules uses a comparator and to my experience these things are either on or off - and should give a "bang bang" control but your Op Amp is more gradual of course. As I have all the parts to build your version (except the NTC) I might just do a little PCB to give me something to distract me from re winding the secondary from scratch onto my toroid, very tedious task indeed and time consuming too since I have to pause at each layer to wait for the epoxy to go off. Good news today, the ferrite core arrived and it is intact, it was packed very well indeed so the seller might have learned something from the many broken ones that other inverter builders received last year. Klaus |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
"What puzzles me though is that oztules uses a comparator and to my experience these things are either on or off - and should give a "bang bang" control" Both are just op amps. "op amps" generally have their own output stage, and comparators generally have open collector outputs. Op amps have slow slew rates, and compartors are op amps optimised for high slew rates. So either can be used in just about any application, it is the surrounding circuit that takes advantage of the different output stages and slew rates. I used the comparator, as I wanted either fast response ( bang bang style ) tight cross over control, all the way to vey mild gentle cross over from full drive to no drive and every where in between.... just adjust the feedback pot to get whatever you want. I was going to use pwm so the 339 was the obvious choice... but the fans use so little power, that linear use of the fets was simpler... but the 339 stayed. Both Glens and mine use the same feedback path to achive the same thing, but the difference is the op amp has it's own output stage, so no pull up resistor is needed. They will both perform the same if you choose the right components around them... (and the pots do that for you.) Feedback changes the rates of change in both as expected, but the slew rate in the 339 is much higher, and so feedback needs to be a bit more savage to control it's gain. Put simply, comparator is a high slew rate optimised comparator with open collector output ( needs pull up's), op amp is a differential amp not optimised for max slew rates, but rather a more linear feel, but both can do the same job in essence, but do their particular style of amplification optimised for different purposes..... feedback changes both from an amp to a switch I guess is how I look at them both. Glen "Next time I have the inverter off ( unlikely to happen soon, been running for weeks now without missing a beat )," It's been running for a bit now, has it performed as you expected, better than you expected, worse than you had hoped. Is there anything it can't do/run with your single transformer, and have you resorted to some current bypass to make huge loads run/turn on easily ( big 225amp welders, big air compressors etc) I still pinchmyself at how perfectly this little PJ thing works. Dumping 8kw on it makes no difference to 50 watts.... it is just rock steady stable.... still amazed. Grogster, it costs about 160AUD to send a transformer across the ditch. I noted one seller on ebay is happy to pull the transformers out and send them without the case. So <20kg parcel ...............oztules Village idiot...or... just another hack out of his depth |
||||
| Gizmo Admin Group Joined: 05/06/2004 Location: AustraliaPosts: 5078 |
Tinker, what Oz said. Both circuits work the same way. Oz, I would say its performed better than I expected. I fired it up a few weeks ago expecting to run it for a couple of days as a final test, but with all the storms we've been having, I've left it running. No real big loads, maybe the kettle and microwave oven both at once a couple of times, so around 4 to 5 kw. Its seen a few hot days, 38c, where it was supplying 1000 watts most of the day ( Air conditioner, PC, etc. ). No plans to change anything, its proven itself. I did take the lid off the Latronics 7kw inverter for a look. Not much in there, but well made. I changed the fan air inlet to hopefully reduce noise, but yet to test it. Glenn The best time to plant a tree was twenty years ago, the second best time is right now. JAQ |
||||
| Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
So I have another question about that power jack inverter. I should preface that by saying that solid state inverters were unknown when I did my electrical training and I never had to deal with them during my working life - meaning all I know about an inverter is what I'd read on this and other forums. So the question is: can I connect another independent (stand alone) inverter to where the power jack usually connects to the grid? thus making my own mini grid tie system without connecting anything to the mains grid. From what I've read the power jack would synchronize to the *grid tie* connection and then try to charge the batteries. But if that other inverter is just a 1.8Kw Latronics LS inverter its hardly powerful enough to push any energy into the power jack. At the moment I have two small inverters (one of which is the Latronics), both are fed from separate battery banks but charged from the same 2Kw solar set up. Each inverter powers a separate part of my house. What I'm hoping to do is to keep the two separate battery banks, one is LA, the other LYP, both are 400Ah @24V presently. I plan to reconnect the LYP batteries to 48V and use the modified power jack as my main inverter. With the smaller Latronics inverter synchronised to that it could work as a backup and I would not have out of sync voltages putting 500V stresses on my mains wiring. Please feel free to shoot this idea down in flames if this is just one of my famous magic smoke ideas. Klaus |
||||
| Page 5 of 8 |
|||||
 Print this page |
