 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 : What inverter style would you build if starting a new system in 2020
Page 1 of 3   |
|||||
| Author | Message | ||||
| rogerdw Guru  Joined: 22/10/2019 Location: AustraliaPosts: 792 |
Hi guys, As a newbie to this forum and to the endeavour of building a ~6kW solar inverter … I am left wondering exactly what direction to take. If I have an incorrect understanding, please help me out. I understand the original inverter was an Ozinverter instigated by oztules … which evolved from using commercial Power Jack control and output boards … to eventually having a new control board using an 8010 inverter control chip and IR2110S mosfet drivers … … along with different variations of FET output boards. Probably most importantly … a new toroidal transformer, built by sandwiching cores from a couple of Aspire or AeroSharp 3kW GTI’s and rewinding from scratch. I’m assuming that … because Ozinverters needed a fairly specific start-up procedure to prevent blowing up … numerous other variations have evolved in an attempt to solve that issue. In that group I understand there are Warp/Mad/Nanoinverters and … ?? Of course that leaves me as a newcomer wondering exactly what I should build for myself. My main aim is to install maybe 10kW+ of second hand solar panels … a 48V ~600Ah forklift battery … and avoid the massive power bills we regularly get … though I am very happy to remain connected to the grid as my back-up. Thanks for any correction/clarification etc Cheers, Roger Cheers, Roger |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
That is a pretty good summation Roger, and you are on the right track. The whole thing has become an evolving art, and while its realistic to expect reasonable chances of initial success at the lower end of the power spectrum, the problems and frequency of blow ups seems to rise exponentially with the power level. There are many reasons for that, its not just one problem, but a huge nest of interrelated problems that are still not fully understood, and probably never will be. The original Oz inverters and many subsequent versions relied on Chinese microprocessors with unknown software, and possibly some bugs that led to a great deal of suspicion and mistrust. The Forum member developed nanoverters with open source software are a huge step forward in that direction. A lot of progress has been made with the magnetics design. Its now possible to build inverters with vastly less idling power, and much better output waveforms than was possible a few years ago. A great deal has been learned in that direction. But power levels keep rising, and the blow ups with high power pwm continue. Trying to switch currents at over a hundred amps at 20Khz+ with multiple parallel mosfets is far from easy to do. Layout is critical and noise, grounding, and uncertin conditions at power up and power down can lead to unsuspected destructive transients. Anyhow, before retirement I was a power electronics design engineer, and have a little bit of insight into all of this. I wanted a simple reliable bullet proof inverter for my own use at home. I simply bypassed all the potential problems with high frequency pwm and combined the outputs of four simple low frequency square wave inverters to generate a low distortion (<1% THD) sine wave inverter that could easily be scaled up to any power level without all of the above problems. That is the infamous Warpverter. There are advantages and disadvantages with this approach. Chief among the disadvantages is that you need to wind four transformers and build four invertes, not just one. Its much less of a disadvantage at really high power levels. There are a lot more individual parts, especially mosfets and heatsinks. But its all repetition, so similar circuit boards can be used in each inverter. The big advantage is that the really big inverter switches at only 50Hz not 23Khz. The speed reduction of around x450 is dramatic, and takes care of just about all of the problems that plague high frequency pwm. The big transformer now has to handle a 50Hz square wave instead of a 50Hz sine wave, and that has introduced an unexpected new problem with winding capacitance down to ground in the secondary. But its the ONLY problem yet encountered. There are three different approaches to solving that problem. One is to fit a series choke into the primary, and Andrew solved all his problems that way. Mark is winding a new toroid with a low capacitance technique, and Klaus is yet to start serious testing on his Warpverter, but also has developed a unique swinging choke design that looks promising. Slowing down the switching speed might also work, but has yet to even be tried. Anyhow, objectively pwm is perfect for power levels arguably up to about 4Kw to 5Kw, and there is really no other practical alternative. Above 5Kw to say 50Kw or more, pwm is just too difficult IMHO. That is where low frequency switching of multiple inverters to generate a sine wave will be a much easier thing to build and get going. Cheers, Tony. |
||||
Clockmanfr Guru Joined: 23/10/2015 Location: FrancePosts: 427 |
If I personally was starting from scratch then Tony's Warpspeed step inverter would be the way. The only slight drawback is winding 3 toroid's instead of just one big one. If he was to produce a set of PCB's and a build manual then I would have ago tomorrow. For now I will stick with the present OzInverter design, as I build and put them together in my sleep, according to my family. The present OzInverter has a set of PCB's available and a build manual, so its now relatively straight forward to build a working powerful 6Kw-15Kw 48vdc to 240ac pure sine wave 50HZ & a 60HZ 120vac inverter. My Mantra, Keep it 'SIMPLE', make it 'ROBUST', and importantly keep it 'COST EFFECTIVE'. 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. |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
I can supply a bare circuit board and a pre programmed read only memory for the control board. That will generate all of the required waveforms for the four inverters. Alternatively, Andrew can supply a Nano based control board that is 100% plug in compatible with my own purely hardware based control board, and has greater functionality. The mosfet power boards for the inverters are going to vary a lot depending on how many mosfets or IGBTs you plan to use, and their mounting method onto the heatsink. Its really more of a mechanical decision than anything else. But the inverter power boards themselves are pretty basic and very simple. At the moment two Warpverters have been up and running for many months without any problems, and two more are in the very final stages of completion. All four are very different in regards to packaging and physical layout, but all are in the 5Kw to 7.5Kw class and fairly similar electrically. Cheers, Tony. |
||||
mackoffgrid Guru Joined: 13/03/2017 Location: AustraliaPosts: 460 |
I was the first to build Tony's Warpverter (apart from Tony). From Tony's schematics and teachings I wrote all the software and drew all the PCBs to build a Warpverter. It went together, surprisingly well and without a hitch. I did hit a problem at the end and while I took my time investigating it, it was solved rather easily using a large choke. I have not built an Ozinverter but I do own a Latronics SPWM inverter. I am biased towards the Warpverter being the better inverter but I think it's just a little early to make proclamations. I think when Mark finishes his Warpverter and runs it for some time that his impressions will be interesting. If you need small size then build a Ozinverter. If you need a simple build then build a Ozinverter. The Warpverter provides a robust design, robust output, simple design, slow speed, but with a lot more parts (low cost) and four transformers to wind. Have a look at my thread on my build of a warpverter, Mack-OffGrids 26v Warpverter I've put most of the PCBs on github, including source files and a zip of gerbers ready to send to a PCB fabricator. The links to these are in my thread. I'm a big believer in open source and sharing knowledge, there is no such thing as truly unique idea, all knowledge has been built on a preceding idea from somewhere else, so if you can't find a file then ask me for it. Cheers Andrew |
||||
| rogerdw Guru Joined: 22/10/2019 Location: AustraliaPosts: 792 |
Thanks very much for your detailed explanation Tony Cheers, Roger |
||||
| rogerdw Guru Joined: 22/10/2019 Location: AustraliaPosts: 792 |
Thanks Clockmanfr. Before starting this thread, I had just about made up my mind to build an ozinverter ... as much because of the simplicity ... and also because of the numbers out there. (In fact, I sent you an email through your website several days ago, but hadn't heard back) At the same time, I was still trying to work out why so much effort was being put into building far more sophisticated systems ... if the ozinverter worked so well in lots of cases. Been reading and reading and exploring 'in between the lines' to try and make sense of it all ... but figured I should just ask point blank ... hence this thread.  Cheers, Roger |
||||
| rogerdw Guru Joined: 22/10/2019 Location: AustraliaPosts: 792 |
That's a very generous offer thanks Tony ... though I do start to get a bit nervous reading the rest Designing boards and developing equipment is not one of my strong points ... fixing stuff and rebuilding boards is what I'm good at ... though it gets old real quick if I have to fix the same fault over and over again if I don't know what is causing a failure. Cheers, Roger |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
I agree totally. Its very early days yet. Both Mark and Klaus are right at the very final stages of completion. Klaus has actually had his running at low power, but is awaiting some final parts and is currently a bit distracted by other priorities. When we eventually have four rather different Warpverters that have all been put together quite independently by four different people, the results and impressions gained are going to be very interesting indeed. That is an interesting story. My own Warpverter is rather different that the other three, in that it runs from 100v dc (not 48v dc) and it uses large and more expensive individual IGBT power blocks (not multiple low cost Chinese HY4008 mosfets) and it uses commercially wound E and I transformers and not toroids. More through good luck than anything else, the above three factors minimised the capacitace problem to the point that I was completely unaware that it existed. Anyhow, Andrew built a 24v Warpverter using a toroid for the largest transformer, and multiple mosfets, and immediately ran into trouble. He was seeing some very fast 80 amp (I think it was) current spikes in the primary of the largest transformer every time the inverter switched at 50Hz, even under zero load. The problem is that trying to make the secondary winding go instantly from zero to 225v and back again to zero requires a lot of current to charge up and discharge any stray capacitance. The faster you try to make it switch, the higher the charging and discharging current. Its rather like driving the gate of a mosfet, except 225v is a lot more than 15v, and the stored energy in a capacitor goes up square law with voltage. None of this is a problem with a 50Hz sine wave because charging and discharging follows the voltage sine curve and is slow. A fast edged square wave is far more violent in action. So hence the monster current spikes. There is a second factor involved that takes into account the turns ratio of the transformer. The higher the voltage step up ratio, the larger these current spikes are going to be in the primary winding. The capacitance effect between each turn and an adjacent turn is negligible, because there is only about one volt per turn, which is not difficult to charge and discharge very fast, and all the turns are in series anyway. The real killer is capacitance between the ends of each winding and the steel core, and to any winding placed over the top of the secondary. My E and I transformer uses a thick fiberglass bobbin and only the first narrow layer of a multi layer secondary winding, and the very last layer to the electrostatic screen produced comparatively low capacitance. My inverter operates at 100v so my transformer ratio (90v to 225v) is only 2.5 to one reflects almost nothing back. Andrew's toroid has a single very wide secondary layer with correspondingly more capacitance to ground. Not only that his 24v inverter has about a 9.0 to one voltage step up ratio. So all that combined to cause some rather impressive current spikes in the primary. All that was cured by placing an 80uH choke in the primary of just the largest transformer. The other three inverters did not have the problem because the transformers are smaller, and the turns ratios more favorable. Mark and Klaus will be better off turns ratio wise, but their largest transformers are both about twice the physical size of Andrews. So we sort of expected problems. Klaus did some initial testing and was popping mosfets and tripping a circuit breaker. The spikes are of unknown amplitude at this stage, but they are certainly there. Mark is winding a transformer with the primary on first, a massive thickness of insulation, then secondary on top. That is going to make a very big difference, but the scope of the problem is still not yet known. Klaus has a suitable choke ready to go, but hopes to resume testing fairly soon. So that is where we are right now. I had no problems, and Andrew found a solution to his current spikes. We will soon learn a lot more about this as Mark and Klaus can both add to the ongoing story. Cheers, Tony. |
||||
| rogerdw Guru Joined: 22/10/2019 Location: AustraliaPosts: 792 |
Thanks Andrew, this is all very insightful and very very impressive. Awesome work. I can certainly see the advantages of the Warpverter, though your build shows how much more detailed and complex it is. Your comments on size and simplicity make sense. I'm not really worried about size ... my initial research was because I wanted to take our pool pumps off grid ... and maybe a few other power hungry appliances ... so I don't really need megawatts. Would be nice though.  I definitely do lean towards simplicity ... at least for my first build. Maybe once I've built one I'll get the bug like the rest of you and then build something more sophisticated. I appreciate you and Tony sharing all your work, that is certainly a big plus if I were to go down the warpverter track. Thanks again ... Cheers, Roger Cheers, Roger |
||||
| rogerdw Guru Joined: 22/10/2019 Location: AustraliaPosts: 792 |
Thanks for the explanation, I think I even understood it pretty well ... and helps makes sense of all the threads I've read that didn't always convey the whole story. Will be very interested to see how Mark and Klaus get on. While I like to think I have the ability, I'm more concerned with my stickability if I have problems. Always very helpful to have people to ask questions of, but painful if there are constant equipment blow-ups. Too easy to get disheartened and give up. I do have a fairly heavy power load I would like to tackle in time for next winter ... but I reckon I could get away without a 'proper' inverter for that. Possibly using your circuit from the simplified solar hot water ... http://www.thebackshed.com/forum/ViewTopic.php?FID=10&TID=11546#137942 ... to run a couple of under floor heating elements ... aprox 2kW each. Maybe I'll start a new thread once I'm ready to tackle that. Cheers, Roger |
||||
| BenandAmber Guru Joined: 16/02/2019 Location: United StatesPosts: 961 |
You could build one from the pre-made China board They need a few modifications also but someone new to electronics that's what I would suggest be warned i am good parrot but Dumber than a box of rocks |
||||
renewableMark Guru Joined: 09/12/2017 Location: AustraliaPosts: 1678 |
If you want a really simple one to cut your teeth on Mad's little 4kw board is still quite a capable unit but very easy to construct. I made one for my caravan but did it in 24v so it ended up only being capable of 2kw constant. Here is the build for it. Anything bigger than 3 fets per 1/4 and you need the totems to do the switching properly, but this little board doesn't need them. Cheers Caveman Mark Off grid eastern Melb |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
Now complexity is really in the eye of the beholder. A brick wall, or a chain link fence has a very great number of individual parts, but "complex" is probably not the right word. Lots of mosfets, and heatsinks, and transformers, certainly adds up to a great many parts, but its just multiples of something fairly basic. Yes its a lot more work to fabricate. More wire to wind, more metal to cut and drill, more holes to tap. But with luck you only have to go through the ordeal once. And once its working, highly likely first time, it will keep on working. On the other hand, something with far fewer parts that has blown up a dozen times, and takes maybe a year of tears, anger, and frustration, to finally get going. Is that going to be less complicated in the end ? Cheers, Tony. |
||||
| rogerdw Guru Joined: 22/10/2019 Location: AustraliaPosts: 792 |
I didn't know that was still an option and they were still available. Having said that, I have 45 years as an electronics tech so happy to do some construction ... it's just that I am not a design engineer. Far out ... 51 pages ... took me 6 hours to read all that ... what a journey. You're a trooper Mark ... kudos for hanging in there. Mad's 4kW inverter ... do these still use the double stacked Aspire or AeroSharp 3kW cores? Where do I find prices for boards? Yep, very good question. You did mention luck up there too ... dunno how lucky I'm likely to be. Decisions, decisions. All very helpful guys ... still weighing it up. Thank you. Cheers, Roger Cheers, Roger |
||||
| djull Newbie  Joined: 28/10/2019 Location: FrancePosts: 9 |
What is a good price for X1 refurbished aerosharp unit ? something like 120 € is good ? |
||||
| BenandAmber Guru Joined: 16/02/2019 Location: United StatesPosts: 961 |
With all that experience and the people you have talking to you right now I would go for a building a warp inverter The people on here are very helpful and want nothing in return be warned i am good parrot but Dumber than a box of rocks |
||||
| renewableMark Guru Joined: 09/12/2017 Location: AustraliaPosts: 1678 |
I'm a straight shooter, no between the lines innuendos from me. Some had success with the Ozinverter, some didn't. It appears the circuit is pushing the limits of the driver chips to directly turn on the mosfets. Oz himself told me not to fit all the mosfets on the board. He designed it so yeah, I'll listen to him on that. Later on the totem pole idea came along from a commercial board and Madness incorporated that into his power board. Those totems do the switching close to the mosfets, not a overstretched driver chip via a ribbon cable 15cm away. After Mad made those boards available they became the default board to go to. You can run Mad's power board from the Oz control board, Mad's control board or the Nano control board. I have done all three and it works just fine off any of them. Cheers Caveman Mark Off grid eastern Melb |
||||
| renewableMark Guru Joined: 09/12/2017 Location: AustraliaPosts: 1678 |
Nah mate, that was't 51 pages, look here for the 4kw board If you want it to do 4kw you'll need a transformer that is capable of that, I didn't need that much myself(for the little caravan inverter), so it just got 1 torroid. You buy the boards directly from Madness. From memory the 4kw was $20 and the 8kw was $40. Here is the PCB thread You can also buy dubious assembled power boards from Ali express, there are many designs available, but most like to build their own here. The Warpinverter is proven to be reliable @100v, but that doesn't suit many people. It is yet to be a proven workhorse @48v, I think it's quite likely though or I wouldn't go to the effort of building one myself. . Edited 2019-11-03 07:53 by renewableMark Cheers Caveman Mark Off grid eastern Melb |
||||
| djull Newbie Joined: 28/10/2019 Location: FrancePosts: 9 |
Hi everyone, I have found many times during my reading here this totem driver mention for the Madness power board which seems the only one here equipped this way, but, even after some google research, it's not much more clear. Theoretically, with a good control near the mosfet (no long cable between control & power board), it should work better and less subjected to fail, as my readings. 1 : Is it really important ? I think Chinese power boards with a lot (24) mosfets are not equipped with totem driver. right ? Ozinverter which seems reliable too. 2 : I found here some fails and burning mosfets using this totem board for some complicated reasons. Is it more reliable now ? Thanks a lot in advance |
||||
| Page 1 of 3 |
|||||
 Print this page |
