 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 : Inverter synchronising
  Page 2 of 3   |
|||||
| Author | Message | ||||
Tinker Guru  Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Wiseguy, please be informed that there is only *one* opto in play with my syncing experiment. The opto in the other is *not* connected. What happens is, the nano with the opto is sensing the other nano's frequency & phase. The first nano will adjust to match that, though I found the program needs to be adjusted as well initially, to get an exact phase match. All this happens before the switch allows the relays (one in each AC line) to close and link the two inverters. From that point on both must run in unison at one frequency - if I can get them to do that  .Remember, both will have the same power source (battery) so there should be no voltage "pushing" as was suggested. Now the loads on each inverter might be different but if their individual VFB works as expected both output voltages remain the same, they are paralleled after all. I look at it like a cart being pulled by two horses, where the draw bar forces them to match their pace despite them not being a matched pair. So, (like in big & little inverter synced), they both pull the cart successfully. Klaus |
||||
wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 995 |
What if the little horse has a shorter stride - he gets dragged by the big horse. I am going to take more convincing Klaus, I believe having only one units opto sinking to the other still has an issue. Your inverter had 1 output 50Hz derived from a different clock source to the nano. Nano1 adjusted itself to a point in time where they were both in synch. A relay closed and nano1 and Inverter1 are joined together. Nano1 is happy to sync to the merged voltages even though Inverter 1's clock is now slowly drifting away from the perfect phase agreement they had when the relay closed - it has no idea it is wrong now, same for the nano. The result is leading lagging power / energy shift between the two inverters uncontrolled and will get worse as time goes on. I am really referring to current that is - the voltages should be the same - they are joined at the hip...) Steps off soap box lol If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
| noneyabussiness Guru Joined: 31/07/2017 Location: AustraliaPosts: 506 |
Correct me if im wrong but the sync is updated every 50hz clock cycle (in software on nano). . So as long as voltages are the same, there should be negligible current moving between the two.. I think it works !! |
||||
| Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
No, he gets lashed by the driver's whip to keep up the pace.  See nonyabusiness's reply above. Klaus |
||||
| wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 995 |
Ok youre wrong lol The sync is updated to the merging of 2 power sources that were in sync as the relay closed. But their clocks are not locked together, the merged resultant sinewave that is now running from 2 clocks that are not in step and out of sync, they are groaning and grunting and the nano is trying to sync to the resultant mess. As the non nano inverter has slowly drifted from "in phase" the outputs are still tied together and currents are now circulating. I am not going to comment further - I am getting worried that I seem to be the only one that has this opinion. Maybe one last effort. The non nano inverter is free running and can move away from being in sync. the nano cant update to a difference between the 2 sources because - THEY ARE TIED TOGETHER and it is trying vainly to sync to the one resultant combined mess. If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
| Mulver Senior Member  Joined: 27/02/2017 Location: AustraliaPosts: 160 |
I totally agree!! Once they are tied together, the sync function is useless, there is either another way to stay in sync like a GTI does or you need to disconnect, check sync, reconnect. I am positive its possible, i am just unsure how. How does a GTI stay in sync once its connected and pushing its own power out at the same point it checks and finds sync?? I know they push out frequency/voltage changes to confirm the grid is still active, is this part of the staying in sync function? (read this last night trying to find out how they stay in sync frequency wise) |
||||
| Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Thanks wiseguy, I now see that lot is deeper than it looks from the surface. Always good to read others ideas about something I do not know much about. That would explain why my two inverters, once being synced, only remained so for a few seconds... Another thought, if both nanos try to sync to each other but each corrects only in one direction, one towards higher freq, the other towards lower, would they not remain stuck somewhere in the middle? Its getting late, perhaps just a crazy idea Klaus |
||||
| wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 995 |
It seems there is a big problem with synching cause once you synch you are sunk ! I agree there must be a logical way to approach this problem but I also sense a second problem. If the amplitudes are not essentially identical (remember they are individually measured and controlled by each nano) one nano sees a slightly smaller voltage and tries to lift it slightly. The other nano senses this and tries to reduce its output accordingly, I am not sure how this story ends either. Maybe we have a 4KW multivibrator? The good news is that most of the energy sucked by one inverter is returned by the other one. The supply leads are glowing and we dont even have a load connected....... Current signal sensing and comparisons between the 2 inverters and attempting to balance them feels like it could work, the same current signals could perhaps be used to determine synching them together ? Gut feel is that below say 25% output or some arbitrary level the inverters dont attempt to join forces to share the load and only the master operates. If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
| noneyabussiness Guru Joined: 31/07/2017 Location: AustraliaPosts: 506 |
sync It does actually check for difference and corrects accordingly. . However i think you are right that AFTER its synced and relay latched it will pick up only the peaks and as Tinker observed it will drift out of phase until it goes over the set limits and drop out... as a suggestion is it possible to check phasing of 50hz cycle against its own clock and correct as necessary after link has been made... A hardware implementation could be a opto isolated , cleaned up, sinewave from ONLY the nano output and check differences with that.. or im sure you could check at zero crossing via software I think it works !! |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
There is one possible suggestion to the sync problem. Build a nice high accuracy 50 Hz frequency reference, and sync both of your inverters to that. If both inverters are identical, and have similar phase locking dynamics, you might even have a faint chance of it working. Connecting two inverters directly in parallel is tempting fate a bit. But it might have more of a chance of survival if you connect an efficient balancing choke between the two inverter outputs. A suitable choke would have a centre tapped winding. Each inverter feeds one end, and the load is connected to the middle centre tap. If each inverter runs in phase the choke will have no effect except having some minimal dc resistance. If there is a slight voltage or phase difference between the outputs of the inverters, the choke should be able to absorb some of that difference without drama. Its similar in principle to connecting two dc supplies together through load sharing resistors, although the choke will be much more efficient at doing a similar thing with ac. Cheers, Tony. |
||||
| noneyabussiness Guru Joined: 31/07/2017 Location: AustraliaPosts: 506 |
I think the output inductor for sync is a given. .. if you look at any OGI they will have at least a 2 pole output filter arrangement ... well all ive ever seen anyways. I still think its doable with a sync reference from the source / slave inverter compared with actual output wave.. Poida, what you think ?? I think it works !! |
||||
poida Guru Joined: 02/02/2017 Location: AustraliaPosts: 1388 |
I gather we want to run two inverters from an external clean 5V 50Hz clock pulse. This to ensure both are running very close to same phase. Sure, I can do that easily. I would use the pin2 interrupt function to restart the 50Hz output at 0 degrees each time the interrupt occurs. It will be very fast and so not upset the main PWM production. wronger than a phone book full of wrong phone numbers |
||||
| wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 995 |
Poida, the query was for Tinker trying to synch and sum the two mains outputs from a Nano inverter to a non nano inverter. I understand the events were as follows I assume the nano looked at the non nano mains waveform and synched to it. Then the relay closed connecting nano 1 mains to non nano mains. The non nano internal clock slowly drifted from the ideal phase match causing unintended power transfer between the two inverters. Because nano 1 is looking at the resultant sum of the combined wave forms it doesnt actually know the true phase difference between it and the non nano inverter, as it only has the sum of the two as a single composite waveform. I also cannot follow how we can successfully synch and then combine two nano inverters outputs together. If one nano adjusts its output say a small amount higher the other Nano would back off causing the first nano to add even more etc - if this assumption is wrong what is the magic applied to stop this ? If you have a simple answer to either or both I would love to understand it. If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
| Mulver Senior Member Joined: 27/02/2017 Location: AustraliaPosts: 160 |
Poida, could this not be simulated on one of your micro controllers? Does the measurement of the 2 connected wave forms have a measurable difference that can determine the direction and amount of adjustment that is required? What happens the moment you connect the 2 waveform together? does the voltage double as they are both idling but using enough power to sustain the 230v? Then they both get slapped together and instantly you have twice the power required to sustain the 230v, then they try and correct, to come back down to 230? Also how about syncing the wave form before allowing the soft start? raising the power slowly like an GTI? |
||||
| Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Having observed the "soft start" many times on my oscilloscope I can say this, the wave starts up from a straight line at zero volts. So, there is *no* sine wave to sync at this stage, the syncing would need to start after soft start has commenced. I do like the slowly raising power suggestion but have no idea how that is done in a GTI. I'm putting the whole syncing idea on hold for the time until I get some more definite suggestions that I could try out with my hardware. Wiseguy's theory about two clocks chasing each other after syncronisation has occurred got me wondering what happens when a GTI syncs to the grid frequency. They are obviously not chasing one another or using this method for selling solar power to the grid would just not happen. All I was trying to do is something similar. I'm sure the mains frequency drifts a little up and down, the GTI senses that and adjusts to remain phase locked. I do have a GTI that is synced (stand alone) to my home built inverter. And yes, it pumps charge into the battery, according to how much load is connected to my inverter. With a load on it the charging current from the GTI to the battery reduces as the load sucks some away now. And with a still bigger load on my inverter there is no battery charging current observed at my Amp meter, with all coming from the GTI now. More load still and both, GTI & battery supply the current for it, my Amp meter showing current flowing out of the battery. This all works seamlessly and with no dramas of destructive pushing or pulling Volts. Strangely, my inverter's volt meter shows a slight AC voltage drop when a lot of current from the GTI flows into it. Sudden loads have no effect either. The GTI disconnects (bang bang regulation)when the battery reaches a pre set voltage level. When there is no current from my solar panels, like at night, the GTI disconnects but I have observed it to supply as little as 10W of power at low light. So, perhaps we need some way to disconnect the sync link when there is no power drawn from the two inverters. Perhaps this is why I had problems, my testing was done while both inverters were idling. As there is already IFB (current sensing) incorporated with our inverters perhaps this can be used to establish synching only when the power drawn from the master inverter exceeds a set limit. This would justify the whole idea of the sync feature, to increase the power available from one inverter, when it is taxed close to its limits, by connecting another inverter in parallel. Klaus |
||||
| wiseguy Guru Joined: 21/06/2018 Location: AustraliaPosts: 995 |
The only grid tie inverter I ever looked at internally used the following scheme. It was impossible for it to generate anything without having a live grid that was within parameters. I understand it isolated and recreated a mains waveform to analyse by the uC and conditioned it further to use as a reference sinusoidal signal that was used as a reference for current control of the power stage. The uC must have used the mains sine reference (via the A/D) instead of a sine generator/look up table or whatever you are using and created the PWM switching required to generate a copy of the AC signal in current not voltage. I assume it must have used a similar concept to the soft start of Poida's waveforem and monitored the DC supply current until the set power level was achieved and then it stayed at that point, similar to voltage feedback stablisation. If the grid died so did the output from the Grid inverter. if the grid was distorted the sample reference was also distorted which avoided transients etc to a large degree. Automatic voltage matching happens automatically as the AC output ramps up (within parameters) until sine current flows and then matches the set point. Zaphod (or anyone who) has done some grid feed work, perhaps could enlighten us further. If at first you dont succeed, I suggest you avoid sky diving.... Cheers Mike |
||||
Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
I might get shot down in flames but I am going put my 2 cents worth in anyway. First of all the whole point of this is not to sync 2 inverters together but to get the NI (NanoInverter) to sync to another AC source, either mains or a generator(or another inverter for test purposes). So the NI must follow the external source and adjust itself accordingly as long as the external source stays within set limits. Myself and a number of others are currently using Grid Tie Inverters that are quite happy to sync to an Inverter based on the EG8010 chip. There is nothing telling the EG8010 what to do, it just takes the extra power and deals with it on its own. This also is very stable, there is no shutting down and restarting or anything, I have had my inverter running non-stop for over a year with two grid-tie inverters connecting to it every day, there are times that there is only a few watts being contributed. When the sun is shining bright and the battery is not yet up to absorb voltage it is not unusual to have 6KW being converted from AC to DC to charge the battery. So my thoughts are that the NI (NanoInverter) first needs to sync and lock to the source then adjust its voltage to match the source. In this state, the 2 should sync together and run happily ever after with no substantial current flowing between them and no growling. Once this has been achieved the NI should monitor the current being drawn from the source and gradually lower its voltage till a preset limit of current from the source has been achieved. From this point on the NI would regulate how much power it should draw from the generator to charge the battery or add power from the battery if the load exceeds the generators current limit. There would also need to be parameters set for the battery to not overcharge etc. 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 |
No flames to shoot with I can see there Gary, we are on the same page. Its just I did not express it the same way as you. So far in my testing, the NI *did* follow exactly (match the frequency) of the external source. No problems with that, that external source's frequency is sampled via a small transformer. If you study the schematic of this 'sync box', the growling only occurred when I threw the switch to allow power from the NI *and* the master source to meet. Well, I did the test without a load connected, perhaps I should have done that. I too use a GTI to charge my battery bank from 3KW of solar panels - works well. Something the GTI does differently than my syncing trial did. Did you read wiseguy's comments about that? Klaus |
||||
| Madness Guru Joined: 08/10/2011 Location: AustraliaPosts: 2498 |
Klaus did you have the voltages identical when you threw the switch, also do you know how much current was flowing? I get it that the was no issue while it was just syncing, connecting them together is very different thing. It may be worth trying with something like a stove element in series initially, that should prevent any smoke escaping. There are only 10 types of people in the world: those who understand binary, and those who don't. |
||||
renewableMark Guru Joined: 09/12/2017 Location: AustraliaPosts: 1678 |
I have recently been testing a latronics GTI hooked up to my "house grid" So far it's only got two panels hooked up to it. This latronics is a low DC voltage input type (up to 100v DC input). That has worked fine, so the nano doesn't seem to mind incoming power. But that was with only two panels. I'll have to take down some 30voc panels and swap them with 45voc panels to set it up properly. Cheers Caveman Mark Off grid eastern Melb |
||||
| Page 2 of 3 |
|||||
 Print this page |
