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.
Joined: 26/01/2019 Location: United StatesPosts: 408
Posted: 06:37pm 13 May 2019
Copy link to clipboard
Print this post
Since my panels generate ~300V DC unloaded, I am thinking of making a solar direct modified sine wave inverter.
I don't plan to have much mangetics, if at all, so this should be pretty much solid state.
PV array through diode OR on the input and modified sine wave on the output. If I can get the control loop fast enough, the diode OR could probably replaced with switches but don't want to get ahead of myself too much.
I have thought about it to some detail and this approach is fraught with issues: regulation is going to be extremely poor but I will attempt at addressing it by monitoring the array voltage level and changing the step output accordingly.
Even if it does not work out, I will walk away with some added understanding of a modified sine wave inverter.
Which brings me to the topic of the post: Modified sine wave fundamentals!
1. There seems to be some ambiguity w/ what people mean by modified sine wave i. One school teaches that instead of square wave at 50/60 Hz with almost no dead time, output square waves with dead time. 3 steps are used: Vpeak, 0, -Vpeak ii. Another school teaches to use 5 steps instead: Vpeak, Vpeak/2, 0, -Vpeak/2, -Vpeak
2. I would like to learn more about this - what the pulse widths should be, at what amplitudes, why choose Vpeak/2 and not, say, Vpeak/5 and so on
I have been looking for these details and the closest I have found are lecture notes about Pulse Code Modulation, but that does not look like what I need to learn to make this?
I have looked at projects here and was unable to find MSW details either - if I missed any, please let me know and I'll read them!
The closest output wave I came to was looking at the output of a 3 phase star square wave generator where the output was measured between two phases directly but I don't think it needs to be that complicated if I can switch the input voltage itself between Vpeak and Vpeak/2
For now, any resources to help me understand the pulse widths and amplitude calculations would be very very helpful. I am not afraid of math although I find the Taylor series to be more manageable than the Fourier which I am struggling to grasp and if possible, would like to do without Edited by LadyN 2019-05-15
Boppa Guru
Joined: 08/11/2016 Location: AustraliaPosts: 814
Posted: 03:30am 14 May 2019
Copy link to clipboard
Print this post
I used several back when they were more properly known as modified square wave...
Even back then, 4 step/bit squarewaves were the more common, I went to a 14 bit/step modified square wave (ie 14 discrete steps from o to Vsup) which was far 'kinder' to the old linear powersupplies then in use (my old CRT TV had 'static lines' spaced evenly down the screen on the 4 step I first started using, but they disappeared completely on the 'newer' 14 bit one- as did the audio hum, and the fridge motor stopped buzzing too A 24v 2400w one sold by Jaycar, a steal at only $899!!! (mid/late 80's)
How they did it internally I never bothered to find out, but I know it had a HUGE transformer inside- it needed two hands to carry it- so I suspect the magnetics did a lot of the 'smoothing'
LadyN
Guru
Joined: 26/01/2019 Location: United StatesPosts: 408
Posted: 07:47pm 14 May 2019
Copy link to clipboard
Print this post
Boppa, thank you for the input!
You're right - I should start calling these modified square wave.
I am now really curious about how that 14 bit square wave output was designed!
With my thoughts about 5 steps, I perhaps was thinking of a ceil(log2(5)) = 3 bit design.
My stumbling block is I don't know the amplitudes and widths to select to output the proper modified square wave pattern. I have been looking around for some design files and most of them are either square wave or pure since wave designs.
Nothing in between.
I think the primary function of that transformer was to step the voltage up from 24v to 240v. In my case, my PV array already provides 300V DC, so in theory, if I was not interested in load regulation, I can just switch the HVDC to output this modified square wave
Boppa Guru
Joined: 08/11/2016 Location: AustraliaPosts: 814
Posted: 09:43pm 14 May 2019
Copy link to clipboard
Print this post
Oh the transformer was for the voltage step up, but it was massive in comparison to the previous inverter, which was the same rating, that one could be carried in one hand easily, the multistep one required two hands and a great deal of effort to carry it That transformer although outputting 'only' 2400w was similar in size to what the shed guys are using for 5kw or more- hence my musings that it had a part to play in 'smoothing' the raw output
By modern standards it was both huge- and rather primitive- it had a manual multivoltage output selector- if you were charging the batts, you had to select the 'high' value- otherwise it rose to 260v, then once the batteries got flat enough you changed the position to keep the voltage around 240v
I can't confirm it, but it did have a LOT of switching transistors/mosfets on the heatsink, and a lot of wires going into that transformer from them- maybe it was literally feeding different points on a multitap transformer?
One source of ideas for your system might be looking at a switchmode power supply circuit
One thing to look for too is that pure square wave inverters are like rocking horse droppings these days, even the cheapest $20 'inverter in a cup' is a 'modified sine wave' ie actually a modified square wave... Pure sine wave inverters are of course pure sine waves, but anything marked as a 'modified sine wave'is basically what you want, a multistep square wave inverterEdited by Boppa 2019-05-16