Which topology is good for 12V to 300V?

[MrAl]

Dear Mral thanks for your sympathy.yap I saw the Sandy on the TV and net . I was awful. I hope everything goes good. I got the point and I have an idea :

Can I use a special feedback for IC driver ? For example we can change the FB of TL494 and add a sine wave to it with Vpp=2.5v and 50Hz. so we wil provide pure sine wave. ( I am not thinking about negative stage ...)

You should be able to rig something like that up, using instead of a sine a sine that is full wave rectified, then use every other half cycle as the positive half cycle and every other other half cycle as the negative half cycle, changing the switching technique of the output.

We did a commercial AC converter design that went into many many commercial products way back a long time ago that used a sine generator ROM to generate the sine pattern. The trick is to use the sine pattern (full wave rectified pattern) as the reference rather than a constant DC reference we find in DC to DC converters. It's a little tricky but it works. You can then employ sub cycle voltage regulation if you choose to do so, if you think you need that good of a sine wave.

Did you mention yet what you really need this for, such as medical equipment or just the telephone service or maybe computer, etc. ? The end use plays a part in how clean of a sine wave you need. The less clean you need the easier it is to build, more or less.

 

[()blivion]

We did a commercial AC converter design that went into many many commercial products way back a long time ago that used a sine generator ROM to generate the sine pattern. The trick is to use the sine pattern (full wave rectified pattern) as the reference rather than a constant DC reference we find in DC to DC converters. It's a little tricky but it works. You can then employ sub cycle voltage regulation if you choose to do so, if you think you need that good of a sine wave

In μC designs, one trick is to only use one quarter of a full sine wave and essentially reverse and/or invert that data to reproduce the whole thing. This avoids doing any kind of costly maths and lets you either store a more precisely calculated waveform in ROM, or use 1/4 less space. The end application ultimately deciding which you do.

 

[MrAl]

Hello,

Yes and sometimes it works out better to do that in the software generation phase and program the ROM for the whole half cycle. This avoids having to count down as well as up when generating the pattern. Counting down as well as up gets messy so the double size ROM simplifies the hardware design.

We also did a design that was to be used to study the effects of the interplay of the harmonics from many converters connected to the grid. That ROM was designed by a professor who programmed it to generate a given known set of harmonics as well as the usual sine pattern (sine patterns all have some harmonics, but this one was designed with a specific set in mind). It's interesting to find out what thousands of converters pumping these non ideal sine waves into the power grid can do

 

[()blivion]

That's very interesting because today you can put just 1/4 or 1/2 a sine wave in ROM and synthesize virtually any waveform you could ever want by knowing it's Fourier series. Tough I could easily see someone precomputing a waveform that already had the fundamental and some key harmonics on it. It certainly would be a lot easier to reproduce the intended waveform with the end hardware that way.

 

[MrAl]

Hi,

Not sure what you are talking about when you say the 'end hardware'.
What i meant was that if you store only 30 deg, 60 deg, and 90 degrees, instead of 0,30,60,90,60,30, then the address has to run up and down. If you're using a uC this isnt a problem, but in discrete hardware it's harder to do, and back then uC's where not as popular and we didnt want to use them for this anyway.

 

[()blivion]

Oh, I know what you were saying. I just didn't know that they did it at all in discrete hardware while also using a ROM look up table. It's news to me. I figured it would be done in pure analog long before anyone resorted to such things. I mean, it's trivial to do with a μC. But If I didn't have one of those then I don't think I would try to use like TTL or something. I think I would try hard for an analog solution unless I absolutely couldn't. But I'm sure you guys had your reasons for doing it the way you did. Namely the high power probably.

Not sure what you are talking about when you say the 'end hardware'.

What I meant was exactly what you pointed out. If you store only part of the wave, you have to do at least some processing on it if you want a full wave. If you store the whole thing however, then there is no processing required, just output the data to an ADC.

But that this can also be extended to other waveforms if you simply precompute them is interesting. It's like a microscopic PCM recording. Only computed, not recorded.

 

[MrAl]

Hi again,


Oh ok sure.

Well think about the implementation, just a binary counter and a ROM, with a clock generator, period. Later in the next version we used an ASIC for all the logic functions.

With more modern hardware (keep in mind that was 30 years ago) we dont have to store anything. We can compute the sine during the initialization phase using simple algorithms and then use the results for the generation of the sine for as long as the device is turned on

Also interesting is feedback. During the cycle there could be unequal loading during different parts of the sine which would of course cause harmonic distortion. Using sub cycle feedback leads to the strengthening of the wave where it needs it most leading to less distortion due to sub cycle transistor loading and resistances. This is how we were able to obtain less than 1 percent distortion even on higher powered converters.

 

[()blivion]

With more modern hardware (keep in mind that was 30 years ago) we dont have to store anything. We can compute the sine during the initialization phase using simple algorithms and then use the results for the generation of the sine for as long as the device is turned on

That's a good idea. Normally I avoid sine calculations like the plague, CPU expensive. But I never thought of doing it once, storing it to RAM, then reusing the result as long as the machine is on. There is always plenty of ROM and not enough RAM in the μC's I work with. But that is still a neat trick to kick around. It's definitely in my notebook now .

Well... we went way off topic. I think anymore will be seen as thread jacking, so we should probably stop here. Oh well, I'm sure it will come up in another thread though. Until then, good luck to the OP, and good bye.

-()b

 

[arvinfx]

Did you mention yet what you really need this for, such as medical equipment or just the telephone service or maybe computer, etc. ? The end use plays a part in how clean of a sine wave you need. The less clean you need the easier it is to build, more or less.

I am serious about it, what about when μC halt? I could be happened ! , What is your idea, which wave form is good for a LED TV, A laptop and a PC ( all of them are SMPS), but may be I need to turn on a fridge.

 

[MrAl]

Hi again,

Ob:
Well hopefully the OP will gain some insight as a result of our discussion. He may want to go with a more pure sine wave, im not sure yet.

arvinfx:
Yes i realize you are serious about building this. What do you mean though about uC halt? The uC should not be allowed to halt ever or else the thing would be shut down.
I am not entirely sure what waveform is good for all those devices, although i know they use a single pulse sine for PC equipment. I dont know about a fridge either, it has a motor, that's about it, unless it has onboard computer too.
Maybe you could start another thread about asking about experience running various devices like these off of a modified sine which is usually just a single pulse not a sine at all. Sound good? That way you would get a better idea what you really need and if you can get away with the single pulse type it would be easier to design. Not hugely easier, but still easier.
From what i read on the web however it sounds like a guessing game. Some quote battery chargers as not working well with 'modified sine'. Others say motors run harder or something like that. So maybe you are better off trying to make a converter that mimics the sine wave after all, maybe at least with some purity but i would bet if it is close it will work well, as long as it has all the required safety features like current limit. Also remember motors take more current to start up, much more sometimes.

Another idea is to try to convert most things to DC. Perhaps a small DC fridge, DC lighting maybe 12v, etc. This is the most efficient way to do it actually as to run time per battery. If you can get everything to run on 12vDC you can use a battery. RV's do this too. I already have some things that run on 12vDC because i have been preparing too. I dont have fridge though yet. They make small ones enough to keep a few items in that dont use a motor but a Peltier device. Not sure how well they work in practice but maybe that's enough for a few hours or so. You'd have to move the perishables into the new fridge during a power outage. Not too much trouble though if you only need to move a few things.
Lighting is easy. Use LEDs, choosing the power you need (1 watt, 2 watt, etc.). Heating is very hard to do, and im not sure about your area if you need it, but a sleeping bag helps here. As far as cooking, i picked up a Coleman stove and some propane tanks. It beats using the battery power and running them down.

BTW have you built any electronic projects in the past and if so what what the most complicated thing you build (ie soldered etc.)?

 

[arvinfx]

I thing pure sine wave is the best for me because I decided to run all devices without any change. and about battery I will use a 100HP car ( switch on car) for input 12V ,so it won't be problem.

And about experience: I have a small electronic lab in my bedroom and I have made an 350W amp ( rms) for car . and some PIC projects for my home and car. the most complicated project in analog was 12V to +-35V. and digital was a heat and humidity , night , day control for my Iguana terrarium.

 

[tcmtech]

and about battery I will use a 100HP car ( switch on car) for input 12V ,so it won't be problem.

Well it will be if your 100 HP engine only has a 50 - 70 amp alternator. 70 x 12 = 840 watts and that is assuming it has absolutely no extra electrical systems loads on it.

Realistically most modern vehicles are lucky to have an extra 20 - 40 amps peak reserve available with a stock alternator once system loads are factored in.

If you need a 1000 watts or so of sine wave power and you plan to run your car to get it you may want to consider going with a small belt driven AC generator that runs off your engine instead.

Just some more things to think about.

 

[CmdrRickHunter]

One detail I've learned from engineering: ask yourself "how pure" of a sine wave.

Making sine waves pure is an excruciatingly expensive art, especially at high power. A better approach would be to ask "how close to a sine wave is good enough for me." If you're using a lot of commercial grade equipment, they are often quite robust in terms of the inputs they'll handle (hence why most UPSs get away with generating very poor sine waves). If you're using scientific equipment, you may have to study what sorts of noise terms they'll handle. If you're doing high end audio recording, you might want to consider other forms of UPSs, like flywheel technologies that use real generators generate the energy in the same way the power stations generate it for power station grade sine waves.

In the end, you may find it's cheaper and easier to build a simple DC->AC converter and then filter the result to meet your line noise demands.

 

[MrAl]

arvinfx:
Oh ok great, it sounds like you've had enough experience to at least tackle this thing with some confindence.
Iquana? That's cool.

CmdrRickHunter:
Welcome to the forums here at ETO.
It used to be an expensive art, but not in today's world. With the modern microcontrollers you can do a sine
converter *almost* as simply as a standard H bridge to single pulse AC output (they call modified sine).
It just takes a little more calculation time to figure out the pattern you want and implementing it
properly. Almost the same control exactly, just with more pulses. Maybe a little more expensive if
you want a inductor/capacitor output filter, but he's only talking 1000 watts output so that's not too high.
At 120v that's around 8 amps and at 220v that's even less.
I agree that we have to ask the question how pure it has to be, that's why i asked him a few posts back and
he answered that question about three posts ago. He seems to want somewhat pure sine but doesnt need
super perfect sine. He just wants to make sure that everything works the same as it always did, and without
being able to answer the question about how bad some appliance would perform without a pure sine, he wants
to get at least something that resembles a sine to make sure there are no problems that come up. I would bet
that he would be happy with a 10 or 20 pulse synthsized pattern without any (or much) output AC filtering.
Im almost tempted to try this myself because as of late we've had a lot of storm problems knocking out
the electric. The last storm took out our power for 4 days and nights.
Any other ideas you have would be interesting to hear too.

 

[arvinfx]

Well it will be if your 100 HP engine only has a 50 - 70 amp alternator. 70 x 12 = 840 watts and that is assuming it has absolutely no extra electrical systems loads on it.

Realistically most modern vehicles are lucky to have an extra 20 - 40 amps peak reserve available with a stock alternator once system loads are factored in.

If you need a 1000 watts or so of sine wave power and you plan to run your car to get it you may want to consider going with a small belt driven AC generator that runs off your engine instead.

Just some more things to think about.

You right, But I will use 500W continuous not more. and 1Kw is for a small period of time maybe 1 min and I thing it will provide from battery. My alternator is 80A.( car is 405 peugeot)

If I would build a good inverter I built an alternative petrol generator.