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.
Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498
Posted: 11:17am 04 May 2023
Copy link to clipboard
Print this post
The worst thing for upsetting the wave form is things like hair dryers that take part of the sine wave. Also using PWM to drive a solid state relay to a HWS makes a mess of things also. I don't do that anymore, it would cause an EG8010 to shut down and restart again. I don't think a filter is going to stop that stuff. Edited 2023-05-06 11:17 by MadnessThere are only 10 types of people in the world: those who understand binary, and those who don't.
Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498
Posted: 05:44am 07 May 2023
Copy link to clipboard
Print this post
The inverter is working again after fitting a new power PCB. It has been running for more than 24 hours powering the house and back feeding from the GTIs. There have been no hard start events that let out lots of smoke.
I was able to get the cable I used on the chokes to do another turn so each choke now has 4 turns, I do not know the Henries ATM.
It has been running fairly well but it is not as stable as the EG8010. At times today there was at least 6KW being back fed into the battery but the sine waves a little distorted and the inverter makes a little noise but certainly not loud. The voltage though gets pushed up, it is set to 228VAC however I had the GTI's disconnect several times that I am aware of due to VAC above 250. In comparison to the EG8010 the AC voltage never changed.
If I had nothing to compare it to I would probably be happy with it but I think more work is needed on it.There are only 10 types of people in the world: those who understand binary, and those who don't.
Murphy's friend Guru Joined: 04/10/2019 Location: AustraliaPosts: 648
Posted: 08:15am 07 May 2023
Copy link to clipboard
Print this post
Gary, those symptom's are similar to what I experienced but I do not have as much solar power to back feed as you seem to have.
Yes the voltage gets pushed up, happens even with my warpverter which I usually have running my system.
Sine wave distortions seem to be par of doing backfeeding with the nano inverter here so far (more experimenting to do) BUT the warpinverter does not suffer that .
I can't remember what happened when back feeding when I had EG8010 based inverters, it's been some time ago.
What you could try is to use your old EG8010 inverter exclusively for back feeding. Connect your GTI inverters to make a mini grid with your EG8010 inverter, the only common part with your inverter powering the house is the battery connection.
This would let you experiment without losing house power if something goes bang.
pd-- Senior Member Joined: 11/12/2020 Location: AustraliaPosts: 122
Posted: 08:40am 07 May 2023
Copy link to clipboard
Print this post
Is your battery voltage rising as your VAC heads towards 250V
Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498
Posted: 08:41am 07 May 2023
Copy link to clipboard
Print this post
Back feeding with the EG8010 is completely drama-free, with no noises, voltage rise, or disconnects. It just works, I wonder if Poida has EG8010 style modulation code for the nanoverter? If not I will see if I can modify the code myself for it.
Running an inverter just for the GTI's is not a good idea IMHO. That means inefficiencies going from AC to DC, then back to AC for some of the power and losses just from having a second inverter running. The way it works now is the AC power from the GTI's are used as much as possible when the sun shining. Running things like Air Conditioners, clothes washing and drying, dishwasher and I have got my wife to to use the oven as much as practical while the sun is shining. The reason I am doing this is that it is a stepping stone towards synching with a generator and back feeding from it. If I can backfeed successfully then I might as well just stick with the EG8010.
GTIs are the most efficient way to get from PV to AC power. This is most beneficial when there is heavy cloud and not much power available.
I am building another Nanoveter setup to experiment with, I remember reading somewhere how the EG8010 has open loop voltage regulation and Poida made it closed loop for the Nanoverter. Not sure if that would make any difference but if I can get a Nanoverter to emulate an EG8010 then I can get a better idea about what is causing the undesirable effects. One thing I noticed that seems a little odd is that the positive heat sink runs much warmer than the split heat sinks on the output side of the inverter. Edited 2023-05-07 18:49 by MadnessThere are only 10 types of people in the world: those who understand binary, and those who don't.
Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498
Posted: 09:05am 07 May 2023
Copy link to clipboard
Print this post
Battery voltage should not have anything to do with it. Yes, battery voltage is rising slightly as the charge current increases, it might go up 0.5V from 10A charge to 100A.
As the GTIs push more power back through the inverter the VFB should keep AC voltage the same by reducing the amplitude of the modulation. With the EG8010 AC voltage never changes.There are only 10 types of people in the world: those who understand binary, and those who don't.
pd-- Senior Member Joined: 11/12/2020 Location: AustraliaPosts: 122
Posted: 10:26am 07 May 2023
Copy link to clipboard
Print this post
The only time you get back feed is when the GTI's are supplying more power than you have load " other than the battery inverter"
so in effect the fets are turned off " very short on duration "
and the transformer just works as a step down transformer to a bridge rectifier " the diodes in the fets " and the batteries get charged. when the batteries are full there voltage rises and that means the voltage on the primary rises witch leads to the voltage on the secondary rising.
but that doesn't sound like the problem.
it sounds like the voltage control loop in the code is trying to match the output voltage of the GTI's and the GTI's just keep ramping up the voltage to max volts trip
could be totally wrong just a thought
Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498
Posted: 11:11am 07 May 2023
Copy link to clipboard
Print this post
The MOSFETs are being switched on and off at 20khz, what changes is the duration that they are on. That is how the sine wave pulse width modulation SPWM works. If they were not turning on at all then there would be no sine wave and the GTIs would shut down and say there is no power grid.
The code is supposed to keep the AC output voltage at the setting it is given, there is no code that tells it to match the GTIs. A GTI acts as a generator and pushes power into the grid at a slightly higher voltage. The inverter regulates the voltage by reducing the amplitude of the sine wave by reducing the pulse width.
It may be that the modulation method used is also not really compatible with GTIs like the Unipolar modulation normally used in the Nanoverter.There are only 10 types of people in the world: those who understand binary, and those who don't.
pd-- Senior Member Joined: 11/12/2020 Location: AustraliaPosts: 122
Posted: 11:43am 07 May 2023
Copy link to clipboard
Print this post
Lets say the voltage is at 228 the battery inverter is outputting 0W , the GTI's are supplying all the load. the sun pops out from sum cloud and the GTI's start pushing out a few more Kw The battery inverter is still producing 0Kw but the Ac voltage is now 229 and the extra power is charging the batteries " backfeed"
The code should be doing nothing it should se that the AC voltage is > set point and continue to output nothing.
but lets say there is a bug, and the code matches the current voltage of 229 by trying to output Kw from the battery inverter at witch point the GTI's will increase there output voltage to 230 in an effort to supply there available power. until the system hits 250v and the GTI's shut down.
Its tricky sh*t i know when ever anyone cums to my place and i try to explane how 3 GTI's a battery inverter and diesel generator all share a common AC buss for supplying the house and charging the batteries i get sum funny looks
Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498
Posted: 11:51am 07 May 2023
Copy link to clipboard
Print this post
If the code did nothing when the GTIs are making power the voltage would rise a lot.
You can download the code back a few pages and see if you can find a bug in it.There are only 10 types of people in the world: those who understand binary, and those who don't.
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 01:49pm 07 May 2023
Copy link to clipboard
Print this post
Open loop implies no feedback signal from the output for stability or control whomever made that "open loop" statement is wrong unless it was to do with how they had implemented it in a project (probably with rather poor output voltage regulation).
The EG8010 definitely has a closed loop control and is intended to work that way, not wanting an argument but you may as well spend your time looking for the real culprit.
Yes the nano also uses closed loop control. The Warpverters that have been made using a feed forward control based on the battery voltage are tending towards open loop but still have a feedback of sorts from the battery voltage.
I am not backfeeding so dont have a need or motivation at this stage to help you investigate why as it would be 5 odd months before I could get to doing anything anyway.
I suspect the back feeding issue culprit probably has more to do with the nano implementation of my advocated scheme of alternating unipolar PWM which is quite different to the EG8010 SPWM operation. I think sticking with the 8010 modulation scheme and finding another solution of adding the generators output to the system might yield more favourable results.
With regard to synching a generator, nothing would make me happier than to see a generator AC coupled and synchronised to a nano system. However the current scheme of closing the loop which melds the two synchronising signals into one, at the time when they are both in synch and the relay connects the two sources together is flawed. Now as they drift out of synch you are in for a bad time as the waveforms try to drift apart but they can't as they are tied together so heavy currents circulate and bad things may happen.
A couple of methods that may work, one is to convert the generator from AC to DC and feed it via a GTI by pretending to be a solar feed, or synchronising could work if the synch signal comes from the flywheel of a true AC generator, or if the synch signal can be bought out from the inverter stage of an inverter generator and connected as the nano synch input you are in with a chance. I think you will probably ignore my "help" and continue on, so kudos to you and seriously I wish you luck and look forward to being proved wrong - it could happen?
I know big AC generators can be coupled together rather successfully but I am very skeptical of melding a generator to a SPWM inverter unless resorting to two separate synch signals that always remain separated.
This is probably bad news for KeepIS as the 8010 he is currently building which has been cobbled to run similar to the Nano's code probably may not like backfeeding either. But they work great as a plain old inverter....If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
Murphy's friend Guru Joined: 04/10/2019 Location: AustraliaPosts: 648
Posted: 02:34pm 07 May 2023
Copy link to clipboard
Print this post
Mike, you may not know that Tony came up with a current sensing idea that, via a Hall sensor, altered the warpverter's reference voltage, thus adding more AC voltage control. I have experimented with that and it appears to work, however, over time, the delicate trim pot settings needed re adjusting so now have given that idea a miss (and gained a nice 100A hall sensor ).
My back feeding is rarely over 2KW (from 3KW of panels in 3 directions E,N,W) so the AC never got pushed up much beyond 240V and my GTI is happy with that.
Is it possible to use opto isolated drivers (aka Warps idea) with the EG8010 chip?
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 02:52pm 07 May 2023
Copy link to clipboard
Print this post
Yes I do remember discussing that current sense idea with Tony and had forgotten that it was also used.
In an experimental control board I made that Mike (KeepIS) has also just built up, we implemented the EG8010 to produce similar code to the one Gaspo made for you? But it does use the inverse opto/Warps method of connecting to the power H bridge. However that is very different to the EG8010 normal operation, and actually behaves like the nano code. Not to be confused with the special Nano codes that Poida did for Madness recently which more closely emulates the EG8010 code.
BTW we may not be coming West on our holiday due to all the National parks and non allowance of a dog in most of the places we wanted to see. So we may head up to Qld and soak up some heat up there instead
If you are really keen to try the Tony opto method with an EG8010 running in the regular fashion, I would be happy to knock one up for you or just create a schematic that you can have the PCB fun with ? It would remove the glitches of the EG8010 or at least render them completely harmless. Edited 2023-05-08 01:00 by wiseguyIf at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
analog8484 Senior Member Joined: 11/11/2021 Location: United StatesPosts: 108
Posted: 04:38pm 07 May 2023
Copy link to clipboard
Print this post
That's great progress. So, choke winding is the only change?
I suspect the AC voltage regulation issue is more due to the control loop speed and PID gain parameters than the modulation scheme.
BTW, do you know why the nanoverter does not use the eg8010 modulation scheme?
Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498
Posted: 08:55pm 07 May 2023
Copy link to clipboard
Print this post
Klaus, please keep on topic. I know you are not happy when others go off on different tangents on your threads.There are only 10 types of people in the world: those who understand binary, and those who don't.
Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498
Posted: 09:26pm 07 May 2023
Copy link to clipboard
Print this post
I will have to find where but I am sure Poida said that the EG8010 was open loop for some reason, but I may be wrong. I agree that the modulation is most likely the cause, it may have nothing to do with the VFB.
What I plan to do is work on the Nano1 code to get EG8010 style modulation. I have seen you write several times that you thought the EG8010's modulation is inferior but in my experience it has worked flawlessly. The unipolar modulation you suggested to Poida that is in the Nanoverters standard code works beautifully, I like it a lot. However it is not compatible with backfeeding. Perhaps it may (or may not) be possible to switch the type of modulation when the inverter current approaches zero when backfeeding occurs. So that at night it runs in unipolar mode and then as the sun comes out or a generator starts it switches to EG8010 mode. Whether this switching can be done in real time at zero crossing I do not know ATM.
I did try running the EG8010 in what they call Unipolar modulation, I had to use 2 VFB circuits as shown in their diagram, it ran but current draw at idle was over 20A. I gave up at that point.There are only 10 types of people in the world: those who understand binary, and those who don't.
Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498
Posted: 09:30pm 07 May 2023
Copy link to clipboard
Print this post
Yes, the only change apart from a new but identical power board and MOSFET driver ICs is the extra turn on each of the chokes. The previous failure I am 100% sure was due to the brutal instant hard start.
I think but may be wrong that the initial reason for going away from the EG8010 modulation was due to the wobbles in the sinewave that occurred at zero crossing. I think this though is a product of incorrect choke values. Another reason for using different modulation is that some deem the EG8010's modulation to be inferior.
Due to a completely blue sky and a maximum forecast temperature of only 22 degrees, I think it is best to not let this new setup run today. Otherwise, the GTIs are going to be wanting to produce a lot more power than yesterday's partially cloudy conditions. Edited 2023-05-08 08:40 by MadnessThere are only 10 types of people in the world: those who understand binary, and those who don't.
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 10:44pm 07 May 2023
Copy link to clipboard
Print this post
Maybe clarification is needed here, I am confident inferior is not a word I ever used re the EG8010. I believe I said it had some foibles I did not want to deal with at that time. I may have once used the term "dreaded" in a derogatory fashion referring to the EGS002 and surgery required to stop it exploding FETs.
With the EG8010 system, at the zero crossing all 4 drives to the Mosfets were changing at the same time, a small glitch could bring trouble. The primary/secondary windings were subject to a 50Hz x 48V change of polarity every half cycle which was also undesirable. TinyT mentioned the 3 soldier glitches which I also identified along with a shoot through glitch that Poida also confirmed, all undesirable.
In the alternating mode I initially advocated for there were no changes occurring at the zero crossing. Using the inverse opto coupled drive also makes it impossible for a shoot through event even with an EG8010 supplying the drive signals.
The fact that I have used the EG8010 in a recent design albeit making it work the way I wanted it to work should also indicate that I believe it may have promise if used correctly. The makers of the EG8010 admitted to inferior code and operation by adding the 4 external extra transistors to remove the possibility of shoot through in the EGS002s that were being produced by the EG8010. Edited 2023-05-08 08:49 by wiseguyIf at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498
Posted: 11:09pm 07 May 2023
Copy link to clipboard
Print this post
Sorry I was generalising when I used the word inferior, The Nanoinverter modulation is superior for most who are not back feeding. As for shoot through I have never had my inverter fail when running normally. Yes I had lots of failures in the early days, these always occurred during initial testing or if there was an overload. I did have one spectacular failure after I had a short to ground in my GTI regulator PCB, but that was not caused by the EG8010. My house has now been running for over 6.5 years off an EG8010 driven inverter, There have been periods of more than 1 year that the inverter has run non-stop, it could have even been more than 2 years. As for the issue Tinyt raised that is easily avoided by keeping the MOSFET driver ICs shut down until ready to start the inverter. I have added this to the start-stop switch, so the MOSFETs can never turn on until I tell the Inverter to start inverting. The last time my inverter stopped working was due to a connection in the DC breaker working loose over time and overheating until the breaker failed. I replaced the breaker and the Inverter went straight back to work.
The EG8010 has been very useful but it has its limitations and we can't change that. So that is why I have started playing with the nanoverter. It works so well and I am not wanting to take anything away from it and the work Poida has done. What I do want though is to get it back feeding first then to move onto generator syncing, I keep being told I am better off just DC charging from the generator but the more people tell me I can't or shouldn't do something the more determined I will become to do the opposite of what they say.
I don't want to get into a debate about what is right or wrong about the EG8010, I am just saying it has some good points and that it has worked well for me. The Nanoverter works exceptionally well, it's just the back feeding that needs development.There are only 10 types of people in the world: those who understand binary, and those who don't.
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 1156
Posted: 12:18am 08 May 2023
Copy link to clipboard
Print this post
Not wishing to be argumentative but to quote; "I have seen you write several times that you thought the EG8010's modulation is inferior" came across more like quoting me than generalising, that was why I went for the clarification.
I also predicted your mindset already re continuing, I also don't like to be told it can't be done - it just focusses me more - good luck with it. Maybe I misunderstood you but I also liked the original concept of switching a nano between 8010 mode for backfeeding and normal mode when not, I reckon that could work. Edited 2023-05-08 10:27 by wiseguyIf at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike
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.
Print this post
.
).