 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 : Simplifying Solar Power Systems
| Author | Message | ||||
| azhaque Senior Member  Joined: 21/02/2017 Location: PakistanPosts: 117 |
While discussing the requirements of an MPPT CC with SolarMike in another thread, a thought came to my mind regarding the complex systems that Solar power systems have become. Starting from a concern for efficiency the designs have reached a point where exotic ferrites, uControllers with complicated programs and low ESR components so on and so forth, have made things fairly difficult. So the point of departure is SIMPLIFY. Taking De-Bono's advice I did some lateral thinking. I present my line of thought to this august forum for their considered views. Firstly lets have a simple CRGO (not ferrite) based toroidal transformer connected to a half bridge array of MOSFETS driven by a simple IC like the TL494. MOSFETs can be 200 VDC or more to reduce I-square losses. The transformer ratio would be 1:1. Frequency could be between 400-1000 Hz, depending upon CRGO specs. This would not be complex at all as there is no need to have very substantial regulation at this end. Can be mounted in a water-proof box right under the solar panels, that are connected in 200 VDC strings. No need to have a sine wave here. A square wave will do nicely with its associated advantages for the MOSFETs. To top it all, all this won't need drawing up complicated PCB patterns in EAGLE or whatever pcb software you happen to use. The generated AC (around 200 VAC @ 400-1000 Hz) can easily be transported using slightly heavy house wiring cables from the rooftop to the battery room or where ever required with substantive savings in terms of I-square losses. At the business end we would need another toroidal step-down transformer with a ratio that is commensurate with the battery/system voltage (48, 72, 96 etc.). A simple tap changer arrangement to control o/p voltage or a more complex SCR based based topology could be used without having to resort to a COMPLEX uC based system. Whilst a larger capacitor bank would be needed for filtering yet these simple off the shelf capacitors would be without tight ESR specs (and thus cost effective). Normal high current diodes or rectifier stacks will be usable. Your views pls. azhaque |
||||
LadyN Guru  Joined: 26/01/2019 Location: United StatesPosts: 408 |
If you had a chance to look at my thread here , you would realize that I considered this option before and then considered using a DC boost (because this AC route was a no go) for exactly the conduction losses you are worried about. The primary side works well, I have tried it using a MOSFET full bridge drive controlled by an ESP32 that I was able to monitor remotely. The issue is the secondary side. The square wave ripple on the secondary is absolutely brutal. We would need EXTREMELY LOW ESR (== EXTREMELY EXPENSIVE) capacitors to handle the 10+ A ripple for 10 hours a day in high ambient heat. I tried at 120Hz and the transformer, capacitor (10x 3300uF in parallel, 200V KMH (3A ripple rating)) bank heated up enough that I could not continue. I still don't understand why the transformer heated up but the capacitor bank was not warm, but hot to the touch. As I am the least educated member on this board, it's extremely likely I made colossal mistakes, in which case, please educate me. Salaams, |
||||
| LadyN Guru Joined: 26/01/2019 Location: United StatesPosts: 408 |
Wait a minute, using active rectification, the short to ground could be avoided which would make the square waves better than half sine! Unfortunately I can't think about this until the weekend but I will think about it and get back to you unless you solve it already. But why did the transformer heat up so bad? It was rated for 50Hz use, and I ran it at 120Hz |
||||
| azhaque Senior Member Joined: 21/02/2017 Location: PakistanPosts: 117 |
Slow down Natasha. I will need some more detail before I can analyze your design better. Did you discuss it with someone or on a forum before you dumped it? Your thread does not touch on the AC path that you took. The thread is about the DC path using the boost converter. This is also not a bad option if we USE THE I^2R LOSS USEFULLY. What I mean is that this loss is unavoidable if you are going the DC route. So let this loss be the means to bring the DC voltage level down to your working voltage- of course within a limited band. azhaque |
||||
| azhaque Senior Member Joined: 21/02/2017 Location: PakistanPosts: 117 |
By 120 Hz I divine that you were using half wave rectified wave. That has a DC component that you injected into the primary of the transformer. This would have led to core saturation and serious rise in iron losses. Thus the temp. rise. azhaque |
||||
| LadyN Guru Joined: 26/01/2019 Location: United StatesPosts: 408 |
Salaam Mr. Haque, ordinarily you would be right, but I had Tony educate me about the volt second balancing (although that my education about that matter has a looong way to go as practise has revealed to me) so I designed a basic full wave H Bridge that generated both phases. Obviously I still did something wrong (transformer still heated up) - so if you are interested, we can revisit this experiment. I did not. If you are interested, we can revisit this experiment with more vigor but be careful what you ask for - I ask a lot of questions! Don't say I didn't warn you! so this is where I was going with this. In theory, "biphase" (is that a valid term) square waves upon rectification should result in a high duty cycle square wave on output:  We would love to have that now, won't we? However, in practise, a transformer, the power lines etc are not perfect and have capacitance, inductance, stray or otherwise that affects all the nice harmonics that make a square wave, a square wave. So, a square wave on the input:  does not result in a square wave on the output:  Source of images: https://forum.polkaudio.com/discussion/92655/home-cookin-the-audiodharma-cable-cooker So as you can see, the transformer emphasizes the edge transitions of the square wave and attenuates the DC component at the output. Where does the DC component go? I think that could explain the heating in the transformer that I noticed |
||||
| azhaque Senior Member Joined: 21/02/2017 Location: PakistanPosts: 117 |
Salaams to you and the family Natasha. Well, first about revisiting the experiment. I am certainly interested because this is the poor man's way forward. I retire in mid June and I intend to take up all these things thereafter. But as I said I am very interested too. The o/scope screen shots are I think OK for an illustrative purpose but infact over exaggerate the effects of strays. This is because the frequency used is 40KHz where the strays will have significant effects. However at 120 Hz the effects would be almost negligible. So without giving explicit consent of joining your experiment I would still wish to join it in the way that we are working on it right now. I am pretty sure others will also add their inputs enroute. What were you driving the H-bridge with. If it was an ESP32 you would probably have used the PWM outputs. Since I know that you don't have a scope, are you dead sure about frequency and wave overlap etc. One can measure frequency on the common DVMs available in the market in the 100-400 Hz range. I guess you used that to set the frequency at 120 Hz. Then comes the transformer. Did you do a no load test by connecting it to the mains. 50 Hz or 60 Hz would hardly make a difference. If it is 230 VAC primary and you do not have 230 VAC just connect it to 110 VAC without any load and see if it heats up or has shorted turns on the primary side. At least we will know if there is something wrong with it if it draws too much no-load current or the core heats up unduly after an overnight soak test. Salaams again all round. azhaque |
||||
 Print this page |
