Elbow

Hi, Like most here I'm on the Cape Town home user tariff. I made two enquiries to Cape Town - one was about upgrading my supply from 60A to more - partly because I'm considering an electric car and partly in order to be able to have a larger PV install. They basically said I have to apply to find out how much that might cost. I then also asked about time of use tariffs. They didn't exactly answer that question but sent various tariffs where there is some reference to time of use. The spreadsheet isn't clear but there is a commercial small user tariff which maybe has a lower off-peak rate. Did anyone look at this to see if there is some opportunity in here somewhere? Thanks, @Elbow

Hi, I had no trouble with my Pylon’s, but couldn’t come right with the Carlo Gavazzi ET112 until I got the Victron USB/RS485 adapter and then it was fine. Usually 115200 bps for the Pylon. You can try your A/B the other way around? RS485 adapters often have a FTDI chip in them and cheap ones have clones and they don’t always work, especially because the maker of the original FTDI put something in the windows driver that can “kill” the clone chips. Hope you can come right.

Hi all, I’m getting used to my system which is more of less halving my grid consumption. We were using somewhere around 22-30+ kWh/d, now we are getting 12 or so off the panels. I have 2 US2000Bs and I wonder how others use their batteries. Right now my setup is grid-tied without feed-in. So I have set it up so the panels first supply load then charge the batteries. I also allow the batteries to be used to supply load, but limited draw from the batteries to 1000W. My logic is that if I have power available - whether directly from the panels or from the batteries - then I should supply any load that there is there and now. “Spare” charges the batteries. The nett effect is that the batteries get charged during the day depending on how sunny it is and how much daytime consumption there is. Then from early evening the baterries discharge and usually they reach the threshold (I’ve set 48.5v) before we go to bed and then we are on the grid till the sun gets going in the morning. It looks like this over the last week or so: (Yellow is the power flow in or out of the batteries - above 0 is power going in to the batteries, below 0 leaving them. The blue line is the state of charge as reported by the Pylons). So sometimes there is enough spare to charge the batteries up to 100%, sometimes not. Today was washing day and you can see that the batteries didn’t get much of a look in. Got them to about 70% charged and then they were discharged back down to 30% or so by 9pm or so. So conclusion is that I could use more panels. Only the CoCT won’t let me. There’s no point to add more batteries since I don’t have spare power to charge them from the panels. Of course if/when shedding comes back I’ll have to change my approach. But in the meantime is what I’m doing reasonable? On a really good solar day (for the time of year) the chart looks like so: So I used everything the panels would make - either for daytime consumption or to charge the batteries. And in the evening the batteries lasted till midnight.

I have Pylon batteries and a BMV. It’s nice to look at the BMV but I don’t think it’s really necessary with Pylons. The leds show the state of the batteries and the built in BMS protects the cells.

Hi, The "minutes to run" only works when the battery is discharging, otherwise it says 9999. The others should work, i’ll look at the code when i’m home tomorrow and maybe we can see which output is different and fix it.

Hi, My electrician was quite confident that it should be permanently tied. So that is how it is now. This will trip your earth leakage if your inverter is fed after the E/L - mine is not. I think it's wrong so sometime I will get a contactor and make it dependent one whether the inverter has isolated the input or not.

We shall see, right? Perhaps closer to possibl than the million cycle ultra-capacitors.

This one has 27000kms and 17.7kWh available on the battery - down from the 18.8 new capacity. That is probably "normal", but I have to admit it scares me to imagine the battery capacity shrinking over the years and the utility decreasing. does the 94Ah battery have a better lifespan? You’ve looked after yours better?

To be fair, that is a lot simpler...

Hi, I'm busy with a fun little project to make a little display of available solar power to put in the house to try to encourage running the high draw stuff when there is spare solar power. I used a little ESP8266 board and some led strips driven over i2c. For me the ESP8266 runs a program that listens to MQTT events from my monitoring and drives the display based on the worked out numbers. So on the right you see the current power usage in the house. (Geyser was on so we are full scale) On the left the idea is to show the "available" solar power - at that time it was just a few watts. I might refine to use colours to show more info. But a challenge is that if there is no load there is no power from the panels. So I want to show an estimate of the available power. The first part of making that estimate is the weather - yr.no has an API which should allow me to get an estimate of cloudiness. The second part would be the solar insolation at the address on the date. Those two with a fudge-factor should provide an estimate of what is available. Has anyone looked at this before - can anyone point me at a public API or dataset that can work our insolation for a lat/long and date? My idea was to mash together the insolation and the weather and come up with a "factor". I can make this available on a public API. User would query the API and multiply the factor by their system size and a fudge factor that gets the estimate to be approximately right for their results.

It’ll be nice to compare how you do with my setup. Though in CT this week weather is a bit up and down. The install looks very nice.

I’m watching with interest, especially curious about the "supercap".

My impression from my Pylons is that the controlling BMS takes care of things - mine stay in lock-step (though I only have 2).

Hi, As y’all know by know - I have Infinisolar 3KW hybrid inverter and a pair of US2000Bs. Its in grid-tied mode, and at the moment I’m prioritising supplying loads over charging batteries - my reasoning being that if there is demand why not supply it there and now. I have 9 350W panels on my inverter which gets me close to the VNE of 500V But I have three extra 350W panels and space to put them. And I wired through cabling for a second string all the way to where my inverter and batteries stand. So I consider to add a second MPPT solar charger connected in parallel to the batteries. Since the panel voltage is quite high it would need to be 150/35. 3 panels is about 1kWp and a Voc at 138V. My inverter can only charge up to 25A and the batteries will take (want?) 50A so there is plenty of room to push in more power. 1kW at 50V about 20A so should work OK. Is it OK to do this? What are the pitfalls? Thanks!

Pylontech suggest you pick up + and - at both ends of your stack ala a ring main. Doesn’t that get the same kind of effect?

Musical @Coulombs? 😆

From my side i’ve had no more trouble - probably since I’ve been cycling them after my panels went in. Set the float down to 52v. The 1% imbalance is a lot but for me imbalance tended to go away as soon as the cell voltages went down a bit.

I was lead to understand that the ET112 would work, didnt understand the principle by which the modbus card worked, and assumed that they all worked to a standard. IE I flashed the plastic before doing proper research... and I’m stubborn. OTOH, the SDM630 wouldn’t have fitted in my DB and is much more expensive.

I read this and then looked at your .sig and you are falling over solar panels. Actually you've got big PV systems at home and work. I presume that helps.

Today in Cape Town it started misty but cleared up nicely. With @plonkster's help and after fixing a config problem on my inverter my "Grid-tie without feed-in" is working properly. This is where I used a python program on a Raspberry Pi to convert the measurements from my Carlo Gavazzi ET112 meter so it looks like the Eastron SDM630 that the inverter modbus card expects. I'll share this code sometime since the idea could be used to feed the modbus card data from anywhere - like EmonCMS for example. We had a high consumption day. My batteries also started the day quite empty and were fully charged about 16h30. Today the main consumers so far Geyser: 7.5Kwh Pool filter pump: 4.7kWh Stove: 3.7kWh "Essentials" (computer stuff, tv, lights): 6.3kWh About 8Kwh of "other" (Tumble dryer, Washing machine and dishwasher were all run too but don't have separate CTs for those) As of now in total: 15.9kWh drawn from the grid, 14.5 from the panels. You can see, we used everything the panels could do until 16:30: Yellow line is the nett flow at the grid connection Blue is what is coming out of (positive) or going into (negative) the batteries Red is the grid interface on the inverter - positive if power is flowing into the inverter, negative if flowing out Green is power taken from the PV panels Purple is power on the ac out of the inverter - the "essentials". The control of feed in worked well - 0.3kWh fed in in total. Flows on grid connection to CoCT looks like this: I guess I could offset the zero point a little to reduce the feed in. But 0.3 kW in a day is an average of only 12.5w. So since my meter doesn't trip off or anything I don't see its worth trying to change it. I'm starting to understand better how this all goes down in practice. Thank you SO much to everyone on Powerforum who has helped so much.

Hi Gerald, and thanks. I will take a little look around to see if I can find anything with 94Ah - I think that extra would give the confidence. Might be that it still a matter of waiting for another few years.

Heh. My office Park is a "green building" which means they must do lots of things and I bet car charging would get them points. They have 1000s of solar panels on a huge grid-tie system with sell back to Coct so it wouldn’t cost them so very much.

Yes - they are expensive - probably by any standard but especially at first glance. 2015/2016 models have a 60Ah battery - so theoretically 130km range but I suppose that might be 100km in the real world? So this is firmly a round-town car. I think it will be tough to find one at 250k, but of course in a couple of years time maybe. Wonder what is involved with buying one in the US or EU and importing it. Since their used prices are MUCH lower. Up front cost is offset with lower running costs - but you are proposing very low annual mileage. A calculation attempt : its a 360v battery; 60Ah so 21.6kWh. I think you can use 19 or so. So 100km on 19kWh. If 90% charging efficiency then 21kWh to charge or 210Wh per kilometer. So worst case R2.56 * 0.21 = R0.54 per kilometer. Well that's cheaper than petrol or diesel by about a third? (say 6l/100km @ R14/l = R0.84) Best case you could charge off solar if you are patient and it suits to charge in the day. Maybe get half your charge from your panels. So over 5000km fuel saving between R1500 and R2850. Free charging at work, convenient to use other public chargers can change the results. Our office park has a huge grid-tied installation so talking to them about charging there. There is some sort of hacker culture around the car with "coding" via the canbus - it would be interesting to integrate it with the house solar install since a 20kWh battery is not to be sneezed at. If I can sell my office park on free or low rate power I could bring that power home 😉

@Travis, @Gerald_db, Can I ping you about your i3. I’m toying with purchase of a 2016 model with 60ah battery and 27000kms. No REX. I live in Cape Town and it’s mostly the daily commute maybe a customer visit - 50km a day would be a lot, Stellenbosch and back is less than 100. its very well optioned - must be close to full house. What do you reckon - @Travis is yours lasting well? Would you recommend? The latest has a much bigger battery but oof the price (well this used one ain’t cheap either)

Hi @riogrande75 id be interested to look at the data and attempt to understand the beast. On my side I want to make a little change in my distribution board so I can put a CT probe to measure the actual house consumption - with that and the energy meter reading I can know exactly how much power the inverter is feeding in. My graphs seem to show that the machine ramps the power up and down quite slowly. So some feed-in seems inevitable. Luckily for me my meter just registers feed in as consumption. Other South Africans are so lucky and their meters have a very low theshold and go into "tamper mode".