 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 : Windmills : Transients &Switched Mode Power Supplies
| Author | Message | ||||
Megawatt Man Senior Member  Joined: 03/05/2006 Location: AustraliaPosts: 119 |
G'day KiwiJohn and anybody else who may be interested. Maybe this should be on the elctronics page, but I reckoned that nobody would know to look for it. Gizmo may want to transfer it. Well the main problem with switched mode power supplies is that right after their rectifiers is a big fat electrolytic capacitor, used as the main energy store. It operates at the peak mains supply voltage, ie about 300 volts on a 240 volt system. If you feed in a fast rising wave front, it gets rectified and presented to the capacitor along with the rectified mains voltage wave. The fast rising wave front looks like the start of a high frequency wave, so the big fat electro looks like a very low impedance, so it absorbs the energy in the fast rising front. Right? Not by a long shot. You see, these fat electros are built out of coils of aluminium, to get the large surface area needed to achieve the large capacitance. And as you know, a coil has the property of inductance, that presents a high impedance to high frequencies, or fast rising wave fronts. Now if you want to say, yes, but the manufacturers know that, so they cut the coils in the caps, to stop them being completed turns. But that is still not enough. At very high frequencies, even a straight piece of conductor on a printed circuit board has enough inductance to cause designers to consider it. So the half turns in the caps each present their own inductance and there are a lot of half turns. So what? Well, what it means is that when a fast rising wave front appears on the scene, the high impedance offered by the self inductance of the capacitor causes the wave front to shoot straight through past the capacitor into the rest of the circuit. What is that then? First of all there's the switcher, a silicon chip that bleeds off just enough energy from the big fat cap to supply the circuit of the electronic gadget. Now silicon chips like low voltages, they are designed to operate that way. And the fast rising wave front can be and very often is very much higher than the voltage rating of the chip. If you are relatively lucky, the voltage only stresses the junctions in the chip, but this doesn't help all that much. These junctions can only stand a certain number of voltage stress events before they fail. What happens then depends on the mode of failure. If it goes short circuit, the 300 volts on the big fat cap up the front discharges into all the circuits that follow, such a sound and picture processing circuits that operate at +/-5V, +/-12V, +/-15V etc. So that can kill the whole appliance. Small colour TVs are a case in point. When this type of failure occurs, you just chuck out the set and buy another. OK, you may say, what's that got to do with modified sin wave inverters? Well, they all too often don't produce anything like a smooth sin wave. They do produce a voltage that will read say, 240 on an AC voltmeter. But often their voltage output shape comprises two rectangular blocks, one above the zero axis and one below, and separated one from the other on the zero line. And guess what each side of such a wave looks like? Yes, a fast rising wave front. Another hazard and you can call this one a transient, is the voltage disturbance that comes into your electrical installation when the power company switches its power system off and on. People call this a "power surge", but that is not a good description, it's just a fast voltage transient. If the power system is well loaded at the time of the interruption, there is a lot of damping available and lots of the energy of the voltage transient is dissipated. But late at night, when there is not much load on the system, transient voltages can get very high. And they are characterised by having fast rising fronts. The only thing the power companies can do economically to keep this to a minimum is to do their maintenance, to limit the effect of trees in mains, broken crossarms etc causing interruptions to supply. They could equip every transformer with super surge diverters,(they already have ordinary ones), or they could purchase all new transformers with an earthed screen between high and low voltage windings, but what about the millions they already have on the system. Its just not economical. And they can't stop transients. Electric current just cannot flow without electromagnetic fields and the presence of EM Fields means that when you turn off the power there will be a transient. A bloke called Maxwell wrote about at all a long time ago, and the four equations that describe it mathematically are called Maxwell's equations. Back to those caps. Have a look at a lot of circuits using electros in the power supplies. Notice that many of them have a tiny capacitor in parallel with the elctro. Some have up to four in parallel. Why do that? The little cap(s) just can't make any difference to the capacitance in the circuit! No, it's just to offer a low impedance path away from the sensitive parts of the circuit for the fast rising fronts. You see, the little caps aren't all made of coiled up aluminium foil, so they don't have a lot of self inductance. The same goes for cheap generators. They too can generate sharply rising wave sides. I once really stuffed up and it was this event thet caused me to start studying the field of "Quality of Supply". I was working with a power company and we had to turn off power where there was a service station and they really wanted the power on. So we hired a genny for the day. We had to replace all the electronics in all the bowsers and they couldn't sell any fuel either. The square wave produced by the genny took out the elctronics for the reason described above. Some insurers are on to it, they won't replace a burned out TV if they know you took it camping with you and powered it from an inverter - unless it's a sine wave inverter. The cure for the problem is for manufacturers to put in a few more components in their appliances that safeguard them from transients. But that would make them cost a couple of dollars more, so forget it. That's it. Megawatt Man |
||||
| dwyer Guru  Joined: 19/09/2005 Location: AustraliaPosts: 574 |
Megawatt Man l just read your stories few mins ago however something is missing about "Quality of Supply". I was working with a power company " l would like to know about power supply to houses and some have off peak hotwater system often turn on automatic cause by change in difference signal frequencies is that correct? dwyer the bushman |
||||
Gizmo Admin Group  Joined: 05/06/2004 Location: AustraliaPosts: 5078 |
Good read Megawatt man, and very true. I used to repair computer monitors and power supplied for a living, and the odd power outage at night would mean a weeks worth of repairs to follow. From what I can tell the modified sine wave inverters are ok for devices that use a mains transformer, as the transformer can help to filter out the high speed rise time on the wave, but it would still have a reduced life. Some inductive loads, like cheap pedestal fans, will go open circuit on a modified sine wave inverter. My mothers place ran a solar/batter/inverter system, and she said a fan would last 2 months at most. My fan makes funny noises late at night, which relates to what dwyer is talking about. You can hear these little tones and beeps, its the power board sending out the modulation signals on the 50Hz mains to tell the hot water systems to turn on and off. Glenn The best time to plant a tree was twenty years ago, the second best time is right now. JAQ |
||||
| Megawatt Man Senior Member Joined: 03/05/2006 Location: AustraliaPosts: 119 |
Hello dwyer, Yes, it is tried and true technology. The company I was with used 1050Hz signals, on a three phase basis. It was injected at the 11 kV level, although it can be done right up to 132kV. The aim is to start off with about 6 volts of 1050 Hz signal on the 240 volt mains, but along the lines the voltage drops but the relays still operate at about 3 volts. The signal is injected in bursts, and each burst is counted in just the same way as with digital signalling. Presence of the signal at a particular instant means a "1", absence means a "0". The signal always starts with an "announce" bit, so all the detecting relays around the power system all start monioring for their particular combination of Os and 1s. These days they are programmed with EEPROMS, but in the old days they had a synchronous motor that started with the "announcement" and they had machanical switches that allowed a signal through or not. Again in the old days the signal was generated by a three phase 1050Hz alternator, but today they are generated by bloody big solid state devices. The frequency can vary from around 185 Hz (I forget the exact number at the low frequency end), 750 is common also. You have to pick a frequency far enough away from one of the common harmonic frequencies on the power system, third, fifth and seventh are there in sufficient quantities to worry about, but higher frequencies can still cause problems with resonance. At night when you are tossing and turning, thinking about windmills, you'll hear the signal in fans mostly, although plenty of sound systems reproduce it in their outputs. It's not there for long and when you know what it is, it doesn't seem such a bother. But there is a case where it does get serious and it is related to resonance. Power system operators often push their assets for as long as they can, to defer large expenditures. An example is to use part of an old 33kV or 66kV line as a "cleanskin" feeder into a new development area. These lines are generally bigger and can ship a fair bit of power in for a few years, and deferral of the capital expenditure on a new zone substation can save a tidy sum. But the development areas most commonly these days use underground cables, which have more capacitance that an overhead line. So we can have a resonance between the inductance of the overhead line and the capacitance of the cable. When enough underground cable is connected to cause resonance near 1050 Hz, the signal grows from a max of 6 volts to maybe 20 volts. The relays all operate properly, but because the 1050 Hz voltage is relatively high, as the 50 Hz wave gets to the point where it crosses the zero line (imagine we are graphing it), there are several other zero crossings because of the 1050 Hz riding the 240 volt wave. That causes digital clocks to advance in time, because they work by counting the number of zero crossings and rely on the power system frequency that over a day is very highly accurate. THat's enough on that subject. Yes Gizmo, appliances that use a mains transformer instead of a switched mode power supply use the leakage inductance of the transformer to smooth the fast rising wave fronts. And another yes, it happens hundreds of times a year that electronic gadgets fail on power system switching events. The power companies will not admit that, because lawyers don't understand that there's not much the power company can do to alleviate the problem and they are petrified of being held liable. Of course, apart from the circumstances I mentioned before, it really ain't their fault. And on your comment about monitors etc, I noticed over the years that many computers died, but not so many monitors. Yet they all experienced the same transients! It is because the monitors are better designed, because they have very high voltages inside them anyway, so the designers really have to make them robust, so they ride through more voltage disturbances than associated gear. Megawatt Man |
||||
Highlander Senior Member Joined: 03/10/2006 Location: AustraliaPosts: 266 |
Hi Megawatt, very interesting! Would this product help a modified sine wave inverter? http://www.altronics.com.au/index.asp?area=item&id=K6045 And what is a switch mode power supply? I have a 1000 watt modified sine in my vehicle and mainly run tools from it, I plugged in an old tv once, which I knew was risky and it blew after a few uses. I didn't care as it was old but I was never sure if it was age or the inverter. Central Victorian highlands |
||||
| Megawatt Man Senior Member Joined: 03/05/2006 Location: AustraliaPosts: 119 |
G'day Highlander, The answer is yes and no! Filters do a few things. They absorb some enrgy and divert some more away from wher you don't want it. But they have to be built for a price and that means they don't all have lumpy enough components to absorb enough of the energy from some sources, so that energy gets where you didn't want it. So far as the modified sin wave inverter is concerned, if it is purely a square wave type, I wouldn't rely on such a filter. If it is one that presents a more gradually changing voltage wave, it could well be OK. The solution to the problem is to try it, looking at the waveshape with a CRO with no load, then connecting the load via the filter and looking at the wave shape at the filter output terminals. I would still use a floodlight or radiator instead of the TV set for the first look and only try the TV if things seemed OK. Now, what's a Switched Mode Power Supply? Well in the good old days when we wanted a DC supply we worked out what transformer ratios we needed and started there. We took the secondary voltage from tne Tx and rectified it - that basically just tips all the negative going halves of the sine wave up side down, so the wave looks like a piece of corrugated iron, except the bottoms are all sharp rather than round. So if we wanted to listen to music using that sort of power we would be driven mad because that shape wave would just put too much noise into the sound processing circuits. Hum, they call it. So we filter the wave, basically using capacitors to absorb the energy and discharge it uniformly into the load. We would also have used series inductances to further smooth the output, until we had a pretty straight line for our graph, that being a picture of near enough pure DC. When transistors came out, we started to use them in circuits called "active filters" that made things a bit more economical. Three terminal regulators are also used. But transformers are expensive, so a few years ago a bloke came up with the idea of a transformerless power supply that turned into the swtched mode power supply. How that works follows. The mains voltage is rectified and the output stored in a large capacitor, electrolytic universally, because they can be made large in capacity relatively cheaply. On a 240 volt system, the voltage of the energy stored in the cap is 1./414 times the 240 , or about 300 volts. That's far too much for transistors and other silicon devices. So we use a very special silicon chip, called in the vernacular, a switcher. It bleeds off just enough energy from the capacitor in short vertical slices or pulses and these are then smoothed by filtering circuits associated with the switcher. The switcher determines how much to bleed off by measuring the voltage it is supplying to the load and adding or subtracting a bit to keep things in balance. One of the internal circuits of the switcher creates a non-changing reference voltage that is used to compare with the load voltage at any instant. That's one type there are others. But they all switch little bits of energy through to their circuits. Having said all that, most of them these days do use a transformer, but it's a very different animal to those we used to use in power supplies . They are high frequency transformers and they use ferrite and more advanced material cores. They are very small for the power they handle so are cheap. About the old TV, it might have been a bit sick, after years of copping switching transients that cumulatively cause damage. But the inverter would have finished it off, quicker if it was bit crook in the first instance. Whew! Megawatt Man |
||||
| KiwiJohn Guru Joined: 01/12/2005 Location: New ZealandPosts: 691 |
Thanks Megawatt Man that make good sense but I think we can guard against most of the dangers of using a modified sine wave inverter. What you are telling me is that the modified sine wave has high frequency components that we dont want in the SMPS so surely the protection is some series inductance? But what to use? Some sort of high current choke, how about the primary of a mains transformer? Is there enough inductance in, say a 500 watt, transformer to knock the corners off that square wave? But does the peak volts stay at only 240? |
||||
| Megawatt Man Senior Member Joined: 03/05/2006 Location: AustraliaPosts: 119 |
Well, that's a pretty elementary filter. It's hard to generalise, ecause there are transformers and transformers. One of the newer toroidal core types will have less inductance than a laminated iron one. And it's not just the top corners you need to knock off, the bottom ones need the same, otherwise you still have some component of fast rising wave. So you need to store some energy probably in a large capacitor to deliver some energy when the input wave provides none. That's why I have said in the above that it is hard to make filters for appliances and I didn't say that the reason is that some use amps rather than milliamps. The answer to your question is to go pragmatic. Measure it with a CRO and see what the difference is. Megawatt Man |
||||
| KiwiJohn Guru Joined: 01/12/2005 Location: New ZealandPosts: 691 |
Yea, a suitable capacitor in an LC filter but I think they would be rather big lumps of L and C. Might be easier to just avoid the square wave inverters!  |
||||
| Highlander Senior Member Joined: 03/10/2006 Location: AustraliaPosts: 266 |
How about this one? Good surge on it click here Central Victorian highlands |
||||
| Megawatt Man Senior Member Joined: 03/05/2006 Location: AustraliaPosts: 119 |
Should be right, unless the advert isn't. Megawatt Man |
||||
brett Newbie  Joined: 08/02/2006 Location: AustraliaPosts: 39 |
I have just organised the 3000w 9000w surge version of that inverter from those people of Egay, I will be installing it this weekend into My father in-laws lifesytle cabin in the sticks this weekend. I will also be installing a home made F&P mill and a PL20 charge controler. I will let you know my thoughts on the inverter early next week. Brett |
||||
| Megawatt Man Senior Member Joined: 03/05/2006 Location: AustraliaPosts: 119 |
Sounds good, Brett, if the sine wave is nice we won't hear from you about bad stuff happening to devices connected to the inverters. I guess tha charge controller looks after a storage battery, in which case, even better, the inverter is not going to be pounded with transients. Megawatt Man |
||||
| Chipboy Newbie Joined: 13/12/2006 Location: AustraliaPosts: 16 |
Rectification need not be so bad, I measured a Schottky half wave rectifier out of a computer PSU (its the big one on the 5 V rail) and its junction drop was 0.127V which is a lot less than the traditional 0.6V for a silicon diode. Its also quite capable of a few amps too. Matt Wind wannabe |
||||
Bryan1 Guru Joined: 22/02/2006 Location: AustraliaPosts: 1344 |
Hiya Brett, Eh Mate I got one of 3kw sinewave inverters and customs stung me over $400 for it. so if you haven't got yours yet be prepared to get stung by little johhny. Anyway I couldn't wait for a fuse with mine so I hooked it up to my batteries and ran my 2hp lathe for about 20 minutes off it, the unit didn't even seem warm after that and the batteries held up well so I reckon you'll have a good inverter there. Cheers Bryan |
||||
| brett Newbie Joined: 08/02/2006 Location: AustraliaPosts: 39 |
Got it, no customs yet. It was a risk I was prepared to take, and new it would cost extra but thought it would still be cheep even with another $400 on Top. Good to know that others are using this gear and quaility is OK. Brett |
||||
 Print this page |
