1 minute ago, Manie said:

Open a new topic and then we can maybe come up with a solution for the cluster mode

Sent from my SM-G920F using Tapatalk
 

OK

If you are on utility and you have pv then it will not add by grid watts. I mean if you are on utility mode it is grid watts. What else is it then.

Also when you on utility and no pv then the battery watts will be added to the grid. Because were does it come from. If you chare via n generator. Well then i cant do nothing about it.

Grid watts will be 0 if you are on battery mode/pv. It cant be more than 0

Maybe i dont understand you correcly
I did test this with a single inverter.

Cluster if one charge from grid and one from solar i cant help you. Why would you want to run like that. Does not make sence to me. If your pv is not anougth to carry your load and charge the batteries. Go to grid and and let the pv fill the batteries.

Sent from my SM-G920F using Tapatalk

Totalwatts is apparent power that is calculated as a totalizer.

Sent from my SM-G920F using Tapatalk

Total watts does not include pv at all. Thats only from the inverter load watts

Sent from my SM-G920F using Tapatalk

1 hour ago, ebrsa said:

odd occasion when ICC-Pi hung up.

There is a watchdog on board but I have yet to figure out if and how it works. The CCGX uses a watchdog, so if it hangs it reboots itself, and I MAY have inherited the functionality on Venus_rpi, in which case you are of course welcome to borrow the idea :-) That part of the implementation is open source after all.

You could of course just google "raspberry pi watchdog" :-)

@Manie for my money you can take the graphic display out on the ICC-Pi. I have been using these programs since ICC-Pi became available and never bothered with the Pi graphic display. MQTT works much better for me. On the Pi I really only look at the second tab displaying the information in text mode. I find MQTT infinitely more informative and since the power graphs are of a shorter timespan, it has helped me detect a faulty rainwater pump system where a one way valve got stuck and caused water to flow back and repeatedly start the pump which has a float valve. 

Could you please advise if the changes you showed on the Raspberry Pi ICC thread, that was done to the Inverter/Battery Values tab of ICC-Pi, have been incorporated in version 1.1.2 of the program. If so I guess I have to download and re-install the image. Is Ver 1.1.2 the latest version as that is what my program shows.

I will skype you to get a clear understanding

Sent from my SM-G920F using Tapatalk

There is a watchdog on board but I have yet to figure out if and how it works. The CCGX uses a watchdog, so if it hangs it reboots itself, and I MAY have inherited the functionality on Venus_rpi, in which case you are of course welcome to borrow the idea :-) That part of the implementation is open source after all.
You could of course just google "raspberry pi watchdog" :-)

We did it basicly already. If the app fel it will lock it and restart the app

Sent from my SM-G920F using Tapatalk

The way I see it, there are two things causing major confusion on Axperts running in parallel and almost impossible to solve, unless you add additional power meters or readers:

1. Although you have 2 inverters, including 2 MPPTp's (1 in each inverter), when running in parallel mode, the data comes through as one inverter. You do not get the data for each inverter individually. When you look at the inverter data tab, the data values will change between the 2 inverters about once every second, i.e. if you look at the "MPPT PV Volts 1" data block, it shows you data for inverter 1 and then inverter 2, continuously in a loop scrolling from 1-2-1-2-1-2. The first thing someone will ask is "Why can't I see MPPT 2 Volts for my second inverter in the MPPT PV Volt 2 block? The simple answer is it is shown in MPPT PV Volts 1 block. The only time you will see a value in the MPPT PV 2 block is when you have an inverter with Dual MPPT's, built in a single inverter. 

2. There is no sensor measuring Grid Watts. The inverter only measures Load Watts only. Load Watts excludes Battery Charging Watts. The sun sets, you change over to Utility mode and look at load watts sitting at 2000 Watts and you are happy, unaware your batteries are charging from Grid at 3200 Watts in the background, which is not accounted for anywhere. Your Grid Watts is actually 5200 Watts. Then someone has different charging priorities on his 2 x inverters running in parallel. One set to only to charge from grid and the other set to only charge from Solar panels or a combination of the two, to split the charging load between Grid and Solar. That makes it very difficult for someone like @Manie trying to write a program to manage your system. He can assign any known values to wherever you want. First of all, the charge priority is not shown anywhere per inverter. So how does he assign a certain value or a portion of that value if those values are not available and shown individually per inverter? 

@Don your explanation should clarify the operation and the complexities for @Manie for everyone. I have just one observation, why would anyone be so uninformed to charge batteries from grid and then use batteries to power the inverter. The losses are just too much to make this a rational decision. I set my 2 x Axperts at only 2A from grid as some battery power seems to be drawn from the batteries when the inverters are running off the grid during part of the night. At present I am running on a time basis instead of SOC after experimenting with how long I can run off batteries and not go below 75% SOC before the panels produce enough to supply the house and charge the batteries. That worked out to running from midnight to 17:30 off batteries/solar and off grid for the rest of the time. Takes care of occasional large loads from the kitchen for preparing dinner.

So all I can say is that ICC-Pi and the MQTT Remote Dashboard is worth every $ I paid for it and more. Saves me a pile every month in electricity charges.

1 hour ago, ebrsa said:

I have just one observation, why would anyone be so uninformed to charge batteries from grid and then use batteries to power the inverter.

I have come across some weird setups, lol. I am not sure how you set that inverter priority. As far as I know you can only set the charge priority.

When I leave the house or over weekends when I am not home, I will select the SOC at the same time. Murphy's law - When I leave the house it becomes overcast and starts to rain. I don't want to stay on solar until 18h00 because of the time setting and run my batteries into the ground because there is no sun. If you select SOC setting at the same time, it will go to grid, depending on your SOC setting, irrespective of the time setting. Whichever comes first. 

1 hour ago, ebrsa said:

So all I can say is that ICC-Pi and the MQTT Remote Dashboard is worth every $ I paid for it and more.

I agree. No other way to have complete control over your system. I VNC to my PI from wherever I am. Check the graphs and log out. If I think I am wasting my time running from PV, I change to grid and disconnect. 

2 hours ago, ebrsa said:

@Don your explanation should clarify the operation and the complexities for @Manie for everyone. I have just one observation, why would anyone be so uninformed to charge batteries from grid and then use batteries to power the inverter.

Not sure who or exactly what you are referring to here as I haven't seen the post that mentions this. I presume you mean "use batteries to power the load" because the Axperts only use 30 watts of power each so it is not about powering the inverter.

I imagine that is it very rare for anyone to use the grid to charge their battery bank, but it certainly could happen as a one off. For example, if an installation has a large battery bank, say of 1500 Ah capacity and it has been unexpectedly unattended and the batteries have discharged 900 Ah over the course of 6 days because of poor weather or unexpected loads in the installation, then in order to keep the battery bank in good working order it may be necessary to use grid to help the PV array with charging as 900 Ah takes a LONG time to recharge, especially if there are still loads and the weather continues to be cloudy. If batteries are not fully recharged for a period of more than 3 days they can become damaged. The system will then almost certainly be used again to power loads, including from the battery bank, but this is not done because it makes economic sense but because it is the only way to keep the batteries in good working order.

28 minutes ago, incagarcilaso said:

I imagine that is it very rare for anyone to use the grid to charge their battery bank,

For me it is not rare, I do it every single day. I start the day off with a fully charged battery bank. My load during the day ranges between 2000 and 5000 watts peaking at 9000 watts for a few seconds at times. My 3780 watt panels cannot cope with the daily load, especially when overcast and raining like the last couple of days, so I end up with a SOC between 85 and 95% when the sun sets. I then switch to grid to charge my batteries so they are fully charged by 7h00 tomorrow when I change over to solar again. 

1 minute ago, Don said:

For me it is not rare, I do it every single day. I start the day off with a fully charged battery bank. My load during the day ranges between 2000 and 5000 watts peaking at 9000 watts for a few seconds. My 3780 watt panels cannot cope with the daily load, so I end up with a SOC between 85 and 95% when the sun sets. I then switch to grid to charge my batteries so they are fully charged by 7h00 tomorrow when I change over to solar again. 

Makes sense. Grid charging batteries could also make economic sense when cheap rates can be used to charge batteries that can then carry loads during more expensive periods. In some cases KW hour costs at cheap times can be one third the price of the peak times. Battery banks that can hold and carry KW hour loads through 24-48 hours, coupled with grid charging means that you can effectively run all loads in any 24-hour period "inside" the cheap rate period.

I think batteries are the most inefficient, overpriced item you can think of. They use words like "Deep Cycle", "Renewable Energy", "Low Maintenance". Then you read the small print, (as long as you only use 20% of the battery bank capacity, otherwise you start to reduce the life of the batteries and the number of cycles it is supposed to do.) My base load is around 800 watts at night watching TV. When I go to sleep and shut down everything I can think of, it goes down to 600 watts (2 x fridges,security system with 16 cameras, ADSL Modem, DSTV,  and electric fence. Not even 1 security light). I paid about R40 000 for 8 x 75 kg, 250 Ah batteries, they cannot handle the load, unless I now start to switch off security items not to cause any damage to my batteries and walk around with a candle. I am not prepared to do that and nor will I spend another R40 000 just to get through the night. I pay Eskom the R20 a night for electricity and run the high power guzzlers during the day when solar power is available for "Free", what a joke. When I think what I paid for my system, it is not for "Free".

I do the same, charge by Utility by night 18h45. My base load till bedtime is 350-550W, and 250-450W after lights out. I then sometimes get a switch to Solar @ +-22h45, which lasts till +-02h15, when the system goes back to grid again till 06h30, when ICC auto switches back to solar. How do I prevent this behaviour? Is it because SOC is 100%, that it switches, or is it Axpert setting that does this?SOC switching is set for 88 - 100 %.

23 minutes ago, Sidewinder said:

How do I prevent this behaviour?

Your SOC hits the 100% mark at 22h45. Then it changes to batteries and run until 2h15 when it hits the 88% mark and goes back to grid uttil the next morning. If you want to go right through the night, you would need to reduce the 88% SOC setting to probably about 60-70%. Or remove the SOC tick mark and only run on time control. This will prevent the system from going to batteries during the night for a few hours. 

Sorry read it wrong the first time and had to edit it.

1 hour ago, Don said:

I think batteries are the most inefficient, overpriced item you can think of. 

Yes, they double the price of an installation (at least) and so make it nearly impossible to recoup the costs of installation before real savings start. The idea is right but their price and inefficiency thwarts this. If you calculate the KW hours just the batteries save you it is frightning to see how long it will take before they pay for themselves. It is the same with generators.

Having said all that, I hate having to "waste" PV array generation and batteries are the only way to compensate for that. Having run my installation now for just over two years I know by rule of thumb for all the four seasons roughly how many amp hours my system can recover during the day as well as supply loads fully. With this in mind I always allow my batteries to discharge to about that number of amp hours, knowing that that way all my PV generation will be used. For me that is what makes the battery bank worth having and why I am happy to pay the price for it.

@incagarcilaso sure your batteries may at times not be fully charged and I agree that they will be damaged in time. My point is that given the lack of efficiency in using the grid to charge your batteries is not a smart idea. In Swartland municipality where I live, we have brackets of 0-50, 51-350, 351-600 and >600 Kwh. with escalating tariffs for each bracket so there is no cheap period. I am sure you will agree that one needs to ensure a balance between the solar panel capacity, battery capacity and the loads that you draw. If the weather is such that the panels and batteries cannot meet your demand, taking into account that ideally you want to charge the batteries with solar, then the only alternative is to use the grid directly which would be a great deal more efficient to my mind. I am happy to say that at present with much sunshine my batteries are at 100% every afternoon having been recharged by solar. The system runs on grid mainly at a time of larger loads.

@Don you sure seem to have some substantial loads at times and high currents only serve to reduce the time the batteries will supply power given that it makes sense not to go below about 75% SOC. That is why I have proposed a feature in ICC that would switch to grid should the load exceed a settable value and then revert to solar/batteries once it falls below the limit, unless it was already on grid. I have promised to define exactly how I see this should work to @Manie and hope to get to it soon. It would seem to me that one would benefit by starting the day with the batteries discharged to your desired SOC so that you can optimise on solar charging of the batteries. The question really is whether your panel and battery capacity is sufficient to meet your demands. I guess we all have our own ways of trying to optimise on solar power usage with our systems and each one must work it out for themselves. I am happy with the way my system runs at present after several weeks of experimenting.

@Sidewinder it would help with possible suggestions if you post your panel watts, battery AH and if you have a BMV connected to your Pi. What we all have to consider are the losses when charging from the grid and then the losses when you drain the charge from the battery. I believe it would be substantial in which case it is more efficient to run from the grid directly.

Judging from your past posts, you folks are all familiar with the technicalities of the matter and I am just posting my sixpence worth of opinions in the hope that it may help.

@ebrsa " the only alternative is to use the grid directly which would be a great deal more efficient to my mind". Yes, I follow now. It's not efficient to charge batteries from the grid, if only because recharging requires more power than the discharge. In the 28 months that I have been running my system I've charged the batteries from the grid (simultaneously with solar) twice only, in exceptional circumstances when the discharge got out of control and I couldn't recharge them with solar alone - and I was reluctant to do it even then.

@incagarcilaso I am glad that we agree on the issue of charging with grid power. It is true that there are times when things go awry. On a recent occasion I did not notice that my Pi's realtime clock was wrong and it did not pick up the correct time from the Internet either. Since I run on a time schedule, switching to batteries took place at the wrong time resulting in an SOC of 69%. Fortunately by the time I noticed it, solar had already started charging. What we all mistakenly tend to do is to expect anything technical to work flawlessly and nothing does. I don't mean just forum members, it is  flaw of all mankind. There are times when things just go wrong. I would suggest that the answer lies in trying to identify such possibilities and within reason try to prevent it.

@ebrsa, Signature updated! Yes, also noticed this clock thing this morning, after last night's thunderstorm, my input power tripped, and ran the batteries down to 65% (normally 88% on ICC setting), and then switched the output off. When I woke at 5:10, I reset everything, and ICC showed the time as 4:10. I think the pi boots faster than my LTE router and AP, so it misses the NTP request. I need to put my router (& pi) on Battery power. !! 

10 hours ago, ebrsa said:

@incagarcilaso I am glad that we agree on the issue of charging with grid power. It is true that there are times when things go awry. On a recent occasion I did not notice that my Pi's realtime clock was wrong and it did not pick up the correct time from the Internet either. Since I run on a time schedule, switching to batteries took place at the wrong time resulting in an SOC of 69%. Fortunately by the time I noticed it, solar had already started charging. What we all mistakenly tend to do is to expect anything technical to work flawlessly and nothing does. I don't mean just forum members, it is  flaw of all mankind. There are times when things just go wrong. I would suggest that the answer lies in trying to identify such possibilities and within reason try to prevent it.

In IT, hardly anything run as expected. 

On 2/10/2017 at 10:07 AM, Sidewinder said:

When I woke at 5:10, I reset everything, and ICC showed the time as 4:10.

I wonder how it does that, it comes up with some kind of time estimate. I'll google it when I'm done writing this :-)

I've told this story before, but we once bricked a piece of hardware due to this lack of an RTC. Our Rpi based "gateway" device was upgrading the firmware on an embedded device when in the middle of that NTP decided to reset the clock. Now the way PySerial handles timeouts, is that it records the time when it makes the request, and periodically checks if enough time has passed. So the moment NTP bumped the time forward it turned into a ten minute timeout event and the firmware update was aborted. Now thankfully our devices remain in bootloader mode, so there was a way to fix it... BUT... we had no code in place to do it automatically, so we had to improvise.

The next version of the software had a small workaround: We stop NTP while updating firmware :-)

Aaaah, got it! It saves the time on shutdown in /etc/fake-hwclock.data. So next time it boots it sets the time to at least something this decade instead of going back to 1 January 1970 :-)