 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 : MAD/OZ AC/DC Charge Controller
  Page 8 of 9   |
|||||
| Author | Message | ||||
Tinker Guru  Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Mad, have you ever measured between each negative input of the GTI if *nothing* was connected to it? If you do not see a short circuit between those connections then the GTI input is perhaps not happy having the negative inputs switched rapidly on and off if there might be different power levels at the positive inputs. Just a theory... Klaus |
||||
Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
The only time the GTI has had an issue is when I had the 300 odd volts dc arc to ground, the PV needs to be isolated. THe Aerosharp GTI's don't care if they are grounded but these HF type GTI's test resistance to ground at startup. Anything less than 1 megaohm and they refuse to operate. When it did arc to ground it resulted in shorted IGBTs in the GTI's output. Which I am told would have put HVDC on my AC wiring, this stopped the off-grid Inverter dead with lots of smoke released. What I have had happening with this PCB is that the GTI runs happily however I get no regulation as at least one MOSFET short between drain and source when it starts regulating with PWM. I have not checked this Zeversolar GTI to see if the 2 negative inputs are in common but I suspect not. I have a couple Aurora GTI's as well and it has the facility to parallel the separate inputs both physically and with a dip switch setting. I have learnt a lot through this process, the results have been worth it though. The Off-Grid inverter spends most of the morning running backwards now charging the batteries with the excess power produced by the GTI. The regulator I have built (thanks to Oztules help) then kicks in when Absorb voltage is reached. This reduces the GTI's output as required to maintain the battery Absorb voltage until it then changes to Float mode. Once this occurs the GTI is powering the house load and the OGI is not doing much more than providing the 50HZ Sinewave for the GTI to follow. So while there is sufficient sunlight the GTI is doing most of the work, I could theoretically draw 5KW from the GTI and 8KW from the OGI. Secondhand GTI's are readily available at cheap prices, it gives you a number of advantages including less and much lighter wiring from the panels and MPPT. Also higher efficiency as the PV voltage is high there is less loss due to resistance and no big step voltage step up. Here is the results from the GTI for yesterday, the no output at 8:40 was due to me turning it off to make some changes. The spike at 12 was from my wife turning on something while preparing lunch from then on it was just running lighter loads and maintaining float voltage.  There are only 10 types of people in the world: those who understand binary, and those who don't. |
||||
| Mulver Senior Member  Joined: 27/02/2017 Location: AustraliaPosts: 160 |
So glad to see you've solved your issues!! Do you think with this method you could control multiple strings of panels into the sample gti input? I was thinking on overcast days.. say you have 2 x 2500w strings in parallel into each gti input, total of 10kw. Then as they start to generate too much power for the gti, throttle the extra strings back?? |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
Hi Mulver, That probably could be done however there may be risks of over powering the GTI. There are other ways that it could be done with the way I have designed this. You could have a second GTI and control it the same way, I plan to add current sensing with a shunt so as to be able to set a maximum charge current to your battery bank. My battery specifications say 10% of the 20 hour AH so for a 800AH battery at C20 maximum charge should be 160A. There is also a limit to how much power you can shove back through the off-grid inverter. The extra panels could also be setup to directly charge the battery, the way I have made it you can control up to 4 seperate arrays either via GTI's or DC to DC. There are only 10 types of people in the world: those who understand binary, and those who don't. |
||||
| Mulver Senior Member Joined: 27/02/2017 Location: AustraliaPosts: 160 |
Thanks Madness I'm thinking towards more the maximum utilisation of the gti during low sun days, when you need stacks of panels and only really 1 gti. Like overclocking / virtual tracker to the extreme. I agree you would need to measure and control the extra strings to not allow excessive power go to the gti. |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
I won't be attempting the double up of panels on one GTI. I have enough GTI's now, they can be bought very cheaply. Even on very cloudy days my 5 KW GTI goes very close to running the background loads like fridge, modem and that kind of stuff that is always on. But there have been times that I have seen less than 30W from that GTI in the middle of the day just before a storm on top of an already very cloudy day. There are only 10 types of people in the world: those who understand binary, and those who don't. |
||||
| Mulver Senior Member Joined: 27/02/2017 Location: AustraliaPosts: 160 |
Are the IGBT's in the AeroSharp's any good for PWM of the PV panels? Just deciding what to keep and what to junk! https://www.mitsubishielectric-mesh.com/products/pdf/PM50B4LA060_n.pdf  |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
They could be but they are setup as a H-Bridge internally, they will be good for something, don't know what though. There are only 10 types of people in the world: those who understand binary, and those who don't. |
||||
| Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Gary, looking at your sketch while trying to suss out a simple solar panel string on/off switch for my planned single input GTI I notice you just switch the negative side? That good enough? I presume you use a similar gate drive as with your inverter to turn the mosfet on/off? Sadly, no sketch for that  .What type mosfet did you end up using? I know, you like the all bell & whistle system here but I prefer to start with a simple on/off method, triggered by battery voltage sensing & timed delay off to stop it going bananas on cloudy days. thanks. Klaus |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
Hi Klaus, I am using 600V 20A MOSFETs 20N60S5 and yes they only switch the negative side you only need to switch one to stop the current. It is much the same as the inverter gate drive but this switches much slower and does not need the diodes to turn off so quickly, nor is there the risk of playing Russian Roulette with the opposite side of the bridge. There is an image of the PCB circuit here it is not complicated. https://www.thebackshed.com/forum/forum_posts.asp?TID=9734&PN=2&TPN=7 There is further details in this thread of Oztules version which is simpler, I now have temperature compensation and a number of other changes/additions. You must have the MOSFETs completely isolated from ground and have an isolated power supply to the gate driver chip for most GTI's. I had a couple failures with the old PCB I was using where the MOSFETs arced to ground. It did not harm the MOSFETs but it did kill the GTI output IGBT's. Since I found the problem and made the modification in the diagram above it has worked absolutely flawlessly. I do not like the idea of just turning off the GTI by cutting the PV power. It is a very hard on off situation if my GTI turned off completely I loose up to 4.5 KW of charge. Then the GTI would shut down thinking it is night, it then takes it at least 60 seconds to start making power again, some take 180 seconds. It is not a good way to regulate charge to your batteries. Oztules previously regulated his by turning off the AC connection to the GTI. With what I am using it smoothly reduces or increases the charge as required by the batteries and whatever is drawing power at the time. If you go with the BANG-BANG method I think you will quickly want to change it to a PWM solution. There are only 10 types of people in the world: those who understand binary, and those who don't. |
||||
| Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Yes, I would agree, its not the best method. I did not know it was possible to turn off the AC side, that would be a lot simpler to do with a SSR. Not being familiar with GTI's, would it not loose the grid sync if one cuts the AC side? And how long does it take to re sync after reconnect? You see, its just experimenting here. I have enough solar power (3.1KW) to run my house more than 90% of the time. The GTI input would be only for cooling/ heating with a reverse cycle heat pump A/C - if there is sufficient solar available. Klaus |
||||
oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
I found switching the AC hard, works, but the inverter ( aerosharp 3kw)got system errors frequently, so did not like lower failures like that... which it is supposed to handle as anti islanding.... so then sent for smooth pwm switching on the DC side. It is working perfectly, as is Mads, so it seems to be a solutuion worht pursuing. Unlike Mad, I am happy to turn the GTI off completely, rather than having a minimum level...as the system the current one is in, has pl60 and another DC controller as well 30m closer to the batt bank. I let them take over, and let the GTI turn off when they have enough power to hold float etc. So the GTI may not run all day, only when it is needed. Mad runs a minimum pulse width to keep the GTI alive, which over charges the battery if left in holiday conditions. On one island here, the owner has no been back for 6 months, and the system had been running autonomously for that time... freezer, hot water and not much more. It is checked bi monthly by a fisherman from here. It is comforting that the system runs itself nicely, but a minimum pulse width in this circumstance would be detrimental to the battery... so mine shut down if not required, and wakes up if the PL60 can't keep u[p again. May try to get a pic of one of the installations today. 30 km away. .........oztules Village idiot...or... just another hack out of his depth |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
With a big fridge and big chest freezer my minimum load never drops below 300W so the minimum pulse width has never caused me any issues. Without this I had trouble with my HWS being turned on by a separate charge controller, this triggered the GTI to startup but by the time it started producing power the load had been turned off, this would then keep cycling over and over. I could possibly minimize the effect of the minimum pulse width by dropping it to zero if it stayed there for a period of say 10 minutes. If you are using a fixed speed AC there may be similar issues as the GTI could turn off when the compressor stops, then when it restarts you have the delay before the GTI starts making power. The 60-second delay for the GTI to restart is for it to do self-tests and resync to the grid (inverter). When I first connected the GTI it had no regulation and I controlled the excess power with dump loads ie HW and a heater. This was done by my 2 Midnite charge controllers. I have made provision to turn on loads with my charge controller, one way this could be done is once it starts to reduce the pulse width by a set amount it could turn on an SSR for the HWS. I also plan to use it to control the AC where my inverters and battery is based on temperature and battery voltage. What is really good about Microcontrollers is you can put in a set of conditions and it will just do what you want it to do. There are only 10 types of people in the world: those who understand binary, and those who don't. |
||||
| oztules Guru Joined: 26/07/2007 Location: AustraliaPosts: 1686 |
Yes I have the 750ltr freezer and the twin door ice making fridge.... but I had 9kw of solar on it. When I went down to the 20 you/we set for the minimum, there was still too much power... and I didn't want to go lower, as the pulse width was getting to small for my liking, If very skinny, remember it was switching 100 amps or more, things could go pear shaped very quickly. I expect switching losses to be severe, as the fet would spend too much of it's time in transition, rather than full on.... may be wrong, but thats one of the reasons I relented and went back to zero. As I said, holiday mode would be a bit tragic also, so another reason to keep it to the proper float. With 20 mine kept drifting back up to 59v in the afternoon. My house idle draw can be around the 150 watts for long periods once the HW is over... you can see my problem. I liked the idea though, and it works for you.... oh that was on the"normal controller", not the grid tie. The grid tie installation has the PL60x2, as well as the grid tie, and this would suffer the same fate, as the PL60 would hold 54v or so, and the grid tie would probably drift up from there, the PL60 would probably shut off, but the house that one is on has about 50w idle draw. ( they were used to living on the bare minimum of power)... so you can see where this would have gone  It is as always, horses for courses I guess. .......oztules Village idiot...or... just another hack out of his depth |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
I plan to change the code so there is the best of both worlds, if it stays at minimum pulse width setting for 10 minutes it will then allow it to go to 0. There are only 10 types of people in the world: those who understand binary, and those who don't. |
||||
| Mulver Senior Member Joined: 27/02/2017 Location: AustraliaPosts: 160 |
Mad. How have you found the Aurora GTI? I just picked up a 5kw one today with 5kw of 250 watt panels. |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
I have one working one and 2 that are showing a false ground leakage fault. I have tried the good one for a little while to do some testing of the regulator but that is all I have done with them. I am planning to attempt fixing the 2 faulty ones and sell all of them. Not that I think there is anything wrong with them. I have had a good run with the 5KW Zeversolar that I bought at the tip shop for $5. I now have an interface for it for online monitoring and I have bought another 5 KW Eversolar (now called Zeversolar) so I plan to stick with them. I am waiting on a power meter that uses modbus to supply power data to the Zevercom device, then I will be able to see what power is being used and what is going to the battery from the GTI(s). There are only 10 types of people in the world: those who understand binary, and those who don't. |
||||
| kanchana Regular Member  Joined: 08/05/2018 Location: Sri LankaPosts: 56 |
Interesting, any progress sending the signal to the remote unit? , So Can we can house the battery and the inverter in a separate place connected by the grid to the GTI which controls the battery charging Regards kanchana |
||||
| noneyabussiness Guru Joined: 31/07/2017 Location: AustraliaPosts: 513 |
Mad, just a thought. . A lot of gti's have 2 seperate mppt controllers on the inputs. I know the auroras i have you need to switch them to ether parallel or series on their inputs. Maybe something there why it blowing up mosfets when connected drains. By the way, i love the auroras, they are adjustable via software. Mine ive made to reconnect after 5 secs (still takes a little longer than that) and voltages/frequency limits to max. You can even adjust there max output and a bunch of other stuff i don't fully understand... I think it works !! |
||||
| noneyabussiness Guru Joined: 31/07/2017 Location: AustraliaPosts: 513 |
Mad, also to be a pain, can you please upload latest code for your boards. Re-designing my power wall and going to include your boards. Just want to make sure i have latest. .. Thank you in advance. Glen I think it works !! |
||||
| Page 8 of 9 |
|||||
 Print this page |
