 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 : Warpspeed’s low frequency stepped sine
  Page 2 of 6   |
|||||
| Author | Message | ||||
| Warpspeed Guru  Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
Toroidal cores are all made of very good grain oriented silicon steel, which will have a higher inductance, and thus a lower no load inverter idling current. The main disadvantage of a toroid, is the difficulty of winding on the turns. E and I transformers are becoming more difficult to find these days, most things now use switching power supplies. The iron in these E and I laminations is generally of a lower and inferior quality for low cost manufacture. You can buy new E and I laminations with the same good grain oriented silicon steel that the toriods use, but its going to be an expensive exercise. (That is how I did my largest and most recent inverter). E and I laminations are pretty straightforward to dismantle and reassemble, the windings easy to remove and rewind, especially if it uses a plastic coil bobbin. And its also dead easy to add an electrostatic screen between primary and secondary. You can get away without that, but its good practice to add it if you can. Without the screen you may see some voltage spikes across the secondary due to primary/secondary capacitive coupling. If you are planning on building something really large 3Kw+, winding a transformer from scratch is really the only option available. Stacking two large recycled toroids is the most logical way to go about it. That is how the Oz inverters are all being built, and its a very practical and popular approach that works out very well. Something a fair bit smaller, say up to about 2Kw for the largest transformer, the best approach might be to try and find a large toroid that already has a 220/240 volt winding already on it. Either a 230v step down transformer out of something, or a reincarnated grid tie inverter transformer. That is going to save a huge amount of work. The higher voltage winding is usually the inner winding, but not always ! Then its a case of removing or adding turns to the primary, or in an extreme case totally replacing the primary. That is not going to too be difficult, because the primary usually requires fewer turns of thicker wire. I would adopt a similar approach for the second transformer, but look for one that has a 110/120 volt winding already on it. Pretty easy in the US, but even in other parts of the world you can sometimes find toroids wound with two primaries that can either be connected in series or parallel for 230/115v for example. The smaller transformers are going to have to be wound from scratch, and the easiest way might be to strip old E and I lamination transformers. The transformers will not be that large, and it should not be too difficult to find something suitable to work with. Transformer no load losses become a much smaller part of the total. Its really the largest transformer that needs to be the most efficient regarding idling power loss. The smaller transformers add only slightly to that. Microwave oven transformers are welded together into a solid lump for low cost and very rapid manufacture, and cannot be dismantled. As they use the poorer steel anyway, its just not worth the trouble to thread each individual turn around through the two holes. Might as well use a toroid which will be easier to wind, and have the better core material anyway. I think the very first step might be to find a suitable larger transformer that already has a high voltage winding already on it, and begin with that. Second step would be to think very seriously about what minimum dc input voltage the inverter is ever going to see, and where its power is eventually going to come from. That will probably the discharge end point voltage from some kind of battery. Third step is to work out (with some testing) how to arrange a suitable primary to get from our minimum dc input voltage up to 225v across the secondary winding. Once that is completed, pretty much the same procedure should get a suitable second transformer working. Cheers, Tony. |
||||
renewableMark Guru Joined: 09/12/2017 Location: AustraliaPosts: 1678 |
Tony, I am warming to this idea, I have enough toroids to do it and the winding really doesn't worry me. I have seen your unit but never had enough spare time to go into detail of how it works. I get the concept of the different inverters in series, but what is the actual brains of it? Cheers Caveman Mark Off grid eastern Melb |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
Its rather like extreme slow motion PWM. The duty cycle of all four inverters changes with the dc input voltage to create a (fairly) constant sine wave ac output voltage of very low distortion. There are a great many different lookup tables to suit a 2:1 input voltage range, selected by the instantaneous dc input voltage. No microprocessor, no voltage feedback required. Its all extremely simple in concept. But the biggest advantage is that the very slow switching speeds, around 450 times slower than high frequency PWM, is much more forgiving. Anyone can make a wheelbarrow that works. Building a formula one car is a bit more difficult. That is a bit extreme, but trying to switch 5Kw at 23 Khz is a LOT more difficult to do, than switching 5Kw at 50Hz. Its not a popular idea commercially, because its more expensive to build because of the amount of steel and copper in the transformers. But for home brew it has the big advantage that its almost guaranteed to work first go, and keep on working. Cheers, Tony. |
||||
| renewableMark Guru Joined: 09/12/2017 Location: AustraliaPosts: 1678 |
OK lets talk nuts and bolts. Major parts are transformers x 4, that's the easy part. What else is in the parts list? Does it use normal mosfets like the 4008? Caps req'd?? Drivers? Brains? ala 8010 equivalent Programme? Cheers Caveman Mark Off grid eastern Melb |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
You could use one of the transformers you already have for the large inverter, but the primary would need about roughly 50% more turns. Not too difficult. The second transformer would only need to be one third the size. And the third transformer one ninth the size. So although there are a bunch of four transformers, only two of them are quite large. The other two rather small. Yes HY4008,s would be fine for all four inverters at 48v. Some really large high ripple current caps for 48v, need to be four times the size as for 100vdc, at a similar power level. That is the disadvantage of using lower voltages at very high power levels. Drivers are dead easy, just simple opto isolated one amp mosfet gate drivers will work fine at these very low switching speeds. Each one can easily drive a big bunch of parallel HY4008 mosfets direct, without anything else. Simple brain. No microprocessor required, no voltage feedback required, just fairly simple hardware on a small board with the only connections required being dc input power to the board, and twisted pair drive outputs to all the opto isolated gate drivers, and nothing else. No software as such, but just a big EEPROM to hold all the necessary lookup tables to provide voltage regulation. As the dc input voltage goes up and down, it jumps between different lookup tables in the EEPROM to adjust the waveform to produce a constant 230v sine wave output. Its super fast acting, much faster than voltage feedback. Cheers, Tony. |
||||
poida Guru Joined: 02/02/2017 Location: AustraliaPosts: 1419 |
"Its super fast acting, much faster than voltage feedback." And simpler too. And it has a completely determinable behavior. (that is, no closed loop control that depends upon correct or stable operating parameters and enough phase margin, la de da...) I'd love to make one of these, the hard part for me is sourcing the transformers needed. wronger than a phone book full of wrong phone numbers |
||||
LadyN Guru Joined: 26/01/2019 Location: United StatesPosts: 408 |
What would you suggest and where could we source them from? |
||||
| renewableMark Guru Joined: 09/12/2017 Location: AustraliaPosts: 1678 |
I'd help you out mate but down to the last 4 toroids now, just enough to build one of these. Aerosharps are very hard to come by now. Bryan might help here Cheers Caveman Mark Off grid eastern Melb |
||||
| LadyN Guru Joined: 26/01/2019 Location: United StatesPosts: 408 |
I am struggling with these myself. Thankfully Ben will help me out with some 300VA toroids and if I can't find a 1.2kWh toroid, I might have to build a smaller inverter that I would have to disassemble (or rework) when I DO find a 1.2kWh toroid. Another option is if we got all our heads together and learned how to make these toroids from scratch - in that case, we can scavenge toroidal cores that are easier to come by. I won't be able to pull that off myself as I completely lack the understanding of toroid physics. I have noticed toroidal cores that can be scavenged to be more plenty on eBay and the like but they rarely come with specifications and without those I would be lost |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
Looks like the free stuff has about dried up. There is always the option of getting some new toroids wound. Only source of these in oZ are AEM Cores in South Australia. They carry no stock, and there are no standard sizes to choose from. They have some rather special CNC machinery that can wind any size toroid with any dimensions you choose. There is probably a set up charge, so ordering one might be pretty expensive. But quite possibly a group buy might make rather a large difference, don't really know. No need to stack multiple cores, just order one as tall as you want it. If you guys can come up with some requirements, I am sure the happy smiling sales rep at AEM can give you a quotation. Cheers, Tony. |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
 Probably the best low ESR high ripple current capacitors would be the Evox Rifa type. These are horribly expensive new, but turn up occasionally on e-bay. They always have a white can and heavy screw terminals. I am using three 12,000uF 250v capacitors at 100v and 5Kw. These each have a ripple current rating of 18 amps. A 48v inverter might require three (or four) 47,000uF 100v capacitors with 30 amp ripple rating for the same 5Kw power level and energy storage. These go for typically around $70 each on e-bay secondhand, probably about a third of the price new. https://content.kemet.com/datasheets/KEM_A4034_PEH200.pdf Cheers, Tony. |
||||
Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
I still manage to source 3KW aerosharps occasionally, at $50.- a unit but these are in Perth, WA. Tony, thanks for your Eprom offer in another thread, might take you up on that when I got sufficient parts to start one of these monsters. Just to re cap some basic details, this might help others too. 4 transformers with a power rating 1 : 1/3 : 1/9 : 1/27 are required. For the first two toroids preferable, the smaller ones could be E I type. For 48V do we still use the 9:1 secondary / primary turns ratio for 230VAC? Am I correct in assuming each transformer has its own HY4008 mosfet bridge driving the primary, in other words do I build 4 ozinverter type power stages, one for each transformer? Those big caps pictured above are nice, would the 100V/ 10,000uF ones we use in the ozinverter also do if sufficient are used? If so, how many caps would you suggest? Would it be better to spread the caps to each power stage? Here the layout gets a bit involved, will have to look at your pictures again to see how you solved that. While using 100V batteries makes sense its not what I have here so nominal 48v it has to be. I'm sure you get more questions  Klaus |
||||
mackoffgrid Guru Joined: 13/03/2017 Location: AustraliaPosts: 460 |
Natasha, My email wouldn't allow me to send zip files. I'm getting bombarded by spam, but nooo... I can't send zip files. Here is a link to my github with the zip. zipped eprom maker This file is the test eprom I'm using atm. The software will go to 27C4001, as does my clock board. The software is setup to split the rom in 2, half for standard inverting and half for Hot water mode. The 4 meg eproms are plenty big enough for this. 2019-03-09_233247_rom_27C512_20181209.zip cheers Andrew |
||||
| mackoffgrid Guru Joined: 13/03/2017 Location: AustraliaPosts: 460 |
my transformer spread sheet |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
|
||||
| Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Thanks, that cleared some up but now you got me wondering about the 40 Volts. My battery bank is never going to be that low, an absolute minimum of 48 volts would be more realistic. So, do I just replace your 40 v with 48 and recalculate the ratios or is there some (to me) unforeseen reason why 40V? The square waveform difference is something I was unaware of, thanks for explaining. Klaus |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
Why 40 volts ? Well I argued for a very long time that anyone winding a transformer should first determine what the lowest dc input voltage was likely to be, and fit an appropriate number of turns to suit. I was shouted at and howled down by the "inverter mafia" here, and TOLD that 8:1 is the correct turns ratio for a 48v inverter and anything else is dead wrong. So I just gave up in the end, and decided to "go with the flow". Just could not be bothered arguing any more. But you are quite right to bring up this point. Historically the discharge end point for a 12v lead acid battery is 10v. And for a 48v lead acid battery, obviously 40v. But these days there are much better battery chemistries around, and the old lead acid standard voltages no longer always apply. Cheers, Tony. |
||||
| Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Would that info be accessible without having to sign up to some group or other as one is asked when clicking on that link? All the info about stepped sine wave transformers given in this forum has me confused not a little since replies by those who know are tailored for different battery minimum voltages, different AC grid voltages etc. etc. So, some sort of easy to read table would help a lot. Perhaps one for Australia and another for USA foumites .Also, a basic schematic drawing how the 4 transformers connect together on the AC side. As warpspeed says, best to tackle this from the transformer side first, then the power driving stages and last the step producing electronics. Klaus |
||||
| Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Sorry Tony if I hit a sensitive spot here... If you had said 'battery minimum voltage' instead of the value the "inverter mafia" suggested  then I would not have asked in the first place .As you say, not all batteries are the same. Now something different, more mechanical. I see you use just one common heat sink since you have isolated IGBT's for the switching job. So, using HY4008's would it be ok to use one big heat sink for all the mosfets that have their drains at + battery voltage (using the heat sink as a conductor)? For the HY4008's that have their drains connected to the primaries one would need a separate heat sink for each half bridge when using non isolated mosfet mounting? The above would mainly apply for the first & second transformer so, one long positive rail heat sink and 4 off 1/4 size primary rail heat sink. Transformer 3 & 4 'could' be similar but on a much smaller heat sink for the positive, etc. Klaus |
||||
| mackoffgrid Guru Joined: 13/03/2017 Location: AustraliaPosts: 460 |
you mean like this   Cheers Andrew |
||||
| Page 2 of 6 |
|||||
 Print this page |
