Hi,
Looking for info to start upgrade my project I have found this great post; I will read and follow it. Thanks for share your job.
Regards from spain.
Love this thread, amazing!
I just finished an Off Grid build in the cabin with 16 EVE 280AH cells (48v total), a JK BMS and a Multiplus II 5000 VA. The battery is working amazingly well, and I would definitely recommend using those Lifepo4 cells with a JK BMS for anyone wanting to save some money on the battery side. Would also recommend installing Venus OS on a RPI instead of a Cerbo GX if you are on a budget, works spotlessly and is of course cheaper.
Only thing that is left is getting the JK BMS into HA, which should be doable with an existing ESPhome component. Waiting for some cables.
My next project is to set up a solar system on my House. It’s been snoozing since we will rebuild the roof a bit, but hopefully it will start in the fall.
Will look into combining Victron and Fronius. Got to say lol that I can’t figure out the benefits, but will check the seminar linked above. Does the Victron invert the 48v from the battery while the Fronius invert the PV DC to AC instantly or what?
Edit: Read post 17 and now I get it, thanks! 
One thing: As I understand it the power goes form DC to AC and then back to DC again to charge the batteries if there is spare power. Is the losses there significant?
@drcoolzic Saw that you had a screen dump from the Off grid garage, been my bible when building my off grid system.
But I was meaning to ask, do you also use the JK BMS with your LiFePo4 cells? If so, have you been able to integrate it to HA?
1 Like
As I already mentioned, I am in the early stage of building a house. Therefore, I am trying to define the best possible solution. I already changed my configuration by replacing the MultiPlus II by a Multi-RS solar 6000 VA and other modifications … but probably it will change again before the construction is finished. Just a quick reply about the advantage of AC vs. DC inverters: it is only a matter of efficiency. In case you want to use to power during the day the AC inverter convert directly the PV energy in AC energy that you can use directly on your loads (PV—> AC) so it is slightly more efficient than a DC converter PV—> DC—> AC, but on the other end a DC coupled system is more efficient for storing the energy in batteries (PV—> DC versus PV—> AC—> DC). But looking at several seminars on the subject it seems that we are talking about only a few percent differences! Combining Victron Multi and Fronius Primo is a good compromise (best of both worlds). If interested I can publish my latest configuration.
Yes, I am a fan of Andy on the off-grid garage channel. Very nice guy and very instructive channel.
If you want to integrate JK-BMS to HA, there is already a fully working solution developed by Sebastian Muszynski. This is an ESPHome component so you just need to connect an ESP32 or ESP8266 board directly to the JK-BMS. A friend of mine is using it and it works like a charm to get/set all the parameters.
You can find the project on GitHub: https://github.com/syssi/esphome-jk-bms
If you speak French, you can look at the usage of the integration on the channel of e-nomy for example Vlog #27 en route vers le photovoltaïque : premier bilan au bout d'un mois (avril) - YouTube
1 Like
By the way I have a question for you. You’re saying that you are using an RPi instead of the Cerbo GX and from what I understand it works fine with VE.Direct (e.g. for Victron shunt) and VE.Bus (e.g. for MultiPlus II) but the Multi RS Solar 6000 VA uses the VE.Can bus. From what I understand, someone has developed the Venus OS drivers for VE.Can bus but I am not sure it is easy to use. Do you have any Victron device connected via a VE.Can bus to your RPi like for example a Victron MPPT?
1 Like
Sounds great, would love to see your configuration! We are basically at the same stage on the grid connected build.
The RS looks cool, built in MPPT right? I am a real noob on these things, but can it charge the batteries with AC?
None of my devices have VE.CAN. My Victron SmartShunt and MPPT both have VE.direct and the MultiPlus II have the ethernet looking cable (VE.Bus if remember correctly).
Besides the mess with different connections, all Victron gear has just been plug’n’play once you got the the right connection. But I have no experience of the can-version. I know that Will Prowse did a short video on Victrons connections but I didn’t really learn much watching it.
This is the data on my consumption that I have, first per day and secondly the real time consumption, which dimensions would you guys recommend?
I have a roof with 70m2 facing South East which is good in Sweden, but there will be some lost spaces due to chimneys etc.Processing: 1CBA9FCE-04C5-4346-88B3-C0203F59A894.png…
Yes the Multi RS can charge from AC. The big advantage of the RS over the MultiPlus is the fact that it works at higher frequency and therefore uses smaller core the weight is only 11 Kg compared to the 30 Kg of the MultiPlus and it isapparently much quieter! The only thing it is missing at the moment is the ESS capability but according to Victron this should come soon.
1 Like
VE.Can is usually used for connection to BMS. You can add a CAN interface to RPi and it works well. I did it sometime ago but opted for an out-of-the-box CCGX as my system is mission-critical and the RPi/CAN had its quirks like for instance having to re-install and re-configure the CAN interface after a Venus upgrade. After all I intend to use this for 20+ years so the extra cost of a CCGX is quite small over that period. I keep my RPi for prototypes.
Take a look at this page for info on the subject
1 Like
The Multi-RS are new, didn’t have them available 2 years ago when I set-up my system. From the documentation it appears it can use the PV output directly to the DC/AC converter without going through the batteries; that is dope as it saves you the Fronius-type inverters and may end up being more efficient. Well done.
1 Like
Your post just opened up a whole new world of opportunities. I checked out the EVE batteries and JK BMS and want to try them! The cost per kWh is much lower than with integrated batteries like Pylontech. Thanks for sharing.
1 Like
To size your system I’d recommend using solcast.com
Create a user, a rooftop site with your planned installation, and check-out the forecasted production. My experience over the past year+ is the they’re very accurate. This would give you the production potencial which you can then balance against your expected consumption.
I see from your screen capture you’re in Sweden right? About 40 years ago I had business with Vatenfall up in Stockholm. Nice memories
1 Like
Great!
Getting them from AliExpress can be a bit of a jungle, but I got mine from the recommended vendors at this place:
I was afraid that I would struggle to get mine to work properly. I am a business lawyer and not good with technical things, but everything has worked great (after watching a ton of YT ;)). They are just very friendly to handle.
Ok so I read the documentation in more detail and the BUT is the RS Smart does not support parallel or 3 phase configuration, at least for the moment. So you’re limited to 6000VA max consumption which is fine with small off grid installations but not enough for a fully equipped house. Still this type of all-in-1 equipment is quite attractive if it fits your needs in terms of power capacity. Cheaper, less cabling and security devices, etc. Oh pity, I guess if I was to re-do my setup today I’d go the same Fronius-Victron route. No commercial interest here but the setup has 100% uptime since May 2019, worth the money.
What I would “possibly” change are the batteries. I’m looking hard into the EVE/JK setup as it talks to my geek DYI side and is quite cheaper per kWh capacity than the Pylontech. I still need to sort out the integration with Victron Venus and HA but I guess those are details.
Here is my current thought about my configuration.
First it is not simple to correctly size an installation but I have a big advantage over you as my installation will be connected to the grid. From what I understand, in theory, the total power that your inverters need to provide correspond to the maximum power of what you can draw at a certain time. However, as I said my installation will be connected to the grid and, in overload situation, the extra power can always come from it. So, my idea at the moment (will see what the architect thinks) is to go with a Muti RS 6000 VA and in parallel a Fronius Primo 3.6. During the day, when the sun provides the power, I can hope for about 8.6 kW and during the night for about 5 kW (6 kVA). So, based on this information I think I will install either 12 or 16 kWc of solar panels. This corresponds to 30 or 40 400 Wc panels. They will be organized in a series of 10 delivering a maximum voltage around 410 V with a current of 12A. One string will be handled by the Multi RS one string by the Fronius and the last two by a Victron MPPT RS 450/100. For the batteries I am thinking of using two sets of 16 EVE 304 AH cells with two JK-BMS 200A (should provide above 30 kW). For monitoring the inverters/charge controllers I will use a Cerbo GX and for monitoring the set of two batteries, I will be using a Victron SmartShunt 500A. The Multi RS and the MPPT RS will be communicating over a VE.Can bus and the SmartShunt over a VE. direct bus. The Cerbo and the Fronius will be connected over ethernet.
As you mentioned currently, there are a lot of limitations in the Multi RS firmware: No support for ESS, no support for parallel/3 phases, and apparently not all the information is returned correctly either through MQQT or Modbus. But apparently, they are working of fixing these problems and by the time I will buy the equipment I hope these problems will be fixed.
Below is a simplified diagram of my current thinking
1 Like
It’s an interesting setup combining several options. From what I understand you want to self-consume as much as possible and rely on the public grid just for grey or black days. Have you done a forecast of your consumption, both peak power, kW, and daily energy, kWh? My experience is that we underestimate real consumption and it is good to add some contingency. In parallel, you need to do a forecast of production on “normal” days. As mentioned above, I find solcast.com to be a very reliable source. Also in my experience, we tend to overestimate production so I’d take out some contingency too. Of course production will depend on your panel orientation and inclination and unless you have moving panels you need to decide what you privilege, winter, summer of in-between season production.
One thing about your setup is to check if it respects the Victron 1:0 rule as explained here AC-coupling and the Factor 1.0 rule [Victron Energy]
Because of your mixed setup I’d move this discussion to the Victron community to get advice there.
Another thing is that from the diagram it appears you are connecting the over voltage protectors to the common; I wouldn’t do that, I’d connect it directly to the earth/ground to avoid any spikes going through your system cabling but directly to ground.
1 Like
I’m sorry to monopolize this discussion with questions that get further and further away from the original topic but your experience is very useful.
For the moment I could not make a significant estimate of the consumption but this should be done by the technical office which studies the thermal, hydraulic and energy aspects of my future house. I can’t compare with my current house because the new house will be extremely different. The new house will be built in wood with a very strong insulation and will use a heating floor fed by a heat pump. I guess that during winter the main energy expenditure will be the heat pump and the various household appliances… From what I have seen, the annual consumption of a heat pump with a COP of 4 is estimated to about 50-60 kWh per m². As my house will be about 170 m², this represents an estimate of about 10 Mwh.
To get an estimate of power production I am using JRC Photovoltaic Geographical Information System (PVGIS) - European Commission site (integration in HA Forecast.Solar - Home Assistant). The solar panel will be facing South but unfortunately, they will be mounted on a roof with an angle of only 8° which is far from optimum during winter time. But overall, the loss of production due to this angle compare to the optimum is only estimated to about 4%. Here is the simulated performance for 16 KWc of panel with system lost of 14% an azimuth of 0 and an angle of 8°
The yearly production is equal to 16.3 MWh
In the solution I describe the Factor 1.0 rule is well respected as I only have 3.6 kW for the AC coupled PV inverter compared to the 6 kVA of the Multi inverter.
Actually, while talking about the inverter, I moved my solution from a MultiPlus II to a Multi RS mainly because I am concerned about the noise of the MultiPlus. Many people complain about the noise made by the MultiPlus (both ventilation and rumbling) some even report 70 dB!? In my case, the inverters will be located inside the house and all the comments I’ve seen on the Multi RS seem to indicate that it is very discreet. Are your three MultiPlus located in your house? and are they that noisy?
The drawback of using the Multi RS is that it is a new architecture compared to the MultiPlus/Qattro line that have a long history. It seems to use newer and potentially better technology but it is far from being mature and this explain why it has bug and still missing a lot of features…
Where in the Victron community would you move the discussion? I am not yet familiar with their forum.
I have updated my schematic above to take in account your advice (also added the batteries descriiption)
I can see that there is an integration for solcast.com in HA https://github.com/oziee/ha-solcast-solar
1 Like
No problem to use this thread/forum, glad if I can be of any help as I myself profited a lot form other user’s experiences. Now your post raises a number of questions/subjects.
First, the rule 1:0. My understanding of this rule (but that’s why I still recommend addressing the Victron user community directly - https://community.victronenergy.com/index.html) is that the Victron inverter/charger, in this case the Multi RS, needs to be able to absorb the extra power (current peak and eventually voltage peak) if, when the batteries are almost charged, the consumer loads being fed by the grid-tied-inverter, the Fronius, suddenly go from maximum to minimum such as for instance your heat pump which was on full power stops because target temperature was reached. The Victron senses that and increases the frequency on the output to “tell” the Fronius to stop producing but even if electronics are fast real-life is faster and the Fronius will still over produce for a small time during which the Victron needs to be able to absorb that extra load otherwise over-current and problem. Now, since you Multi RS has its own set of panels connected to it, it could be that its current load is already taken up by these leaving no overhead to absorb the overload from the Fronius. Maybe I got it all wrong and if that’s the case please excuse me but if I were you I’d just research this topic to be sure all is going to be ok. The 1:0 rule was there for the world of combined grid-tied inverters and inverter/charger like “traditional” set-ups. The Multi RS is new, so better safe than sorry…
Second, the noise. My Multipluses, all 3 of them, are installed in the technical room inside the house together with electrical, HVAC, domestic hot water, server, router and switches. There is some noise and the Multis are not the worst offenders but with the door closed barely audible. Just measured and Multis’ SPL is 58dB(A) whereas the server is 64 (it is a 1U with small fans, bad choice). After 1 year living in the house what I have learned is that the issue is not the noise but the heat. 3 inverter/chargers + 2 grid-tied inverters and the batteries do produce heat and whereas that is good in the winter it is not in the summer so I have a medium term plan to move the PV installation out to a shed by the still-to-be-constructed carport and get rid of the heat generation. Cabling is already in place as I figured I’d need something like that in the future when I laid out the cable runs.
But the real point is the Multi RS. I believe it is a very elegant solution as it combines everything under one roof but you may be right that it is still in its infancy. Nevertheless, Victron is a serious company and will not let you high and dry and you will be acquiring tomorrow’s tech rather than today’s. It’s all about choice and it’s good you have it.
Third you are absolutely right, it doesn’t make sense to talk about consumption without taking into account the type of construction. From what I understand your’s is planned to be energy efficient. Just for info, mine is located on the Spanish Mediterranean coast and is built to PassiveHaus standards and due to orientation, openings, and insulation it has zero need for heating energy in winter and requires about 10W/m2 of cooling power in summer plus about 1W/m2 dehumidification power in summer. The house is 360m2 floor so that means roughly short of 4kW cooling power on the hottest days. Of course this need varies during the day/months and the theoretical result for cooling energy requirements is in the order of 14kWh/m2 per year (from mid-May to early-September) which adds to about 5MWh/year. Comparatively the heating energy requirements is zero as said because of the orientation and building envelope characteristics. These numbers are normal when you follow PassivHaus standards and build in this area. Of course, if your winters are colder you’d need more heating but less cooling energy but one should balance the other. If you’re interested browse through Energy efficiency of the Passive House Standard: Expectations confirmed by measurements in practice [Passipedia EN]
As I guess you are in France and if you’re interested in a 3rd pair of eyes I can recommend a very competent architect that can advise you without being intrusive and respecting the relation you already have with your own architect. I did that myself and it was oh so useful! http://www.visionhabitats.fr
The best thing about PassivHaus standards (and mind you, it is not magic, some countries already have new construction standards which are close to PH but Spain definitely not) is not so much in the energy savings but above all the comfort. Constant temperature, equal throughout the house and through the year, no drafts, the first time you experience it you have a strange feeling of wellbeing difficult to explain. And then, when you go back to your old house that you thought was well insulated and had those good double-pane windows you notice how uncomfortable it was!
But I digressed here and the point is that you should be aiming for around 15kWh/m2 per year space heating energy requirement which in your case with 170m2 is less than 3MWh per year. Happy to discuss further if you’re interested.
Finally, a word on production. PVGIS is an absolute place to go for estimates and the numbers they give are a good starting point. The reality is that the sun insists on a 24h cycle and that’s where it gets complicated. Monthly or even weekly production figures are useful for PV power plants and a guideline for a private PV installation but they don’t take you from day to day. Below the daily forecast from the last 6 months
Yes, it goes all over the place and those troughs, black or grey days, that’s when the batteries and the generator are needed. Because of the panel orientation (full south, 35 degrees to horizontal) the highest daily energy production is expectedly in April, with winter months quite good, when the sun shines. And this is the Spanish coast where people will tell you there’s always sun and never rain! Trust me, climate perception is just that, perception and if you want to be self-sufficient and still cook that canard à l’orange in your oven in winter you either a) will thank the public grid or b) will need many more batteries than you’re planning.
Seriously, when I was planning for my installation about 4 years go I used climate records for the area going back all the way 25 years and statistically, there are never more than 3 black days in this area. As you’ll see from the graph, this winter we had a full 2 weeks where production was so low that the generator had to come on almost every day. And it doesn’t help to just add batteries because there’s no sun to charge them either or you have to add panels etc.
So my humble advice is:
- Do your research as well as you can, and you’re definitely already doing it, and come up with your baseline scenario,
- Increase your baseline consumption by 20 to 30% and decrease baseline production by another 20 to 30% and see what it means - how does your installation cope with that? Those are first-level buffers to dampen short term unexpected variations, which will happen for sure,
- Create a design where you have already planned for extension - technical room space, roof size or ground panel structures, batteries and battery chargers, cabling or at least cable runs,
- Decide on your design, get it up and running, kill any bugs and live with it for at least 1 year. Collect all data including statistical cost savings so you know the ROI of your installation,
- After 1 year, compare plan and results, and decide whether you’re happy, or if you need an upgrade,
- Carefully balance simplicity/cost/innovation. Not in your case but if you’re going to rely on your installation being up to have lights, open the door, hot water for the shower, basic stuff, you don’t want to cut corners or be at the cusp of newness.
I would like to give you some background information. Normally I don’t really need to do all this research about the photovoltaic system because I have a design office that is supposed to calculate and determine the different technical installations for me. But, because of my scientific background, I am interested in studying in detail certain fields, in particular everything that touches electricity and home automation. Even if my proposals will have to be validated, it allows me to orientate the objectives because the solutions proposed by the architects are often standards based on big brands which offer little flexibility of evolution and interconnection. That’s why I make sure that the solutions I select can be easily integrated into the Home Assistant that will be the heart of my home automation system.
As you have noticed the construction project will take place in France in the Paris region. Since January 1st of this year the building norms in France have evolved and become stricter concerning the ecological aspect and in particular for everything that concerns insulation, heating, and refrigeration. So, my house will have to respect a certain number of rules which do not go as far as the PassivHaus certification but which are close to it at least on the thermal level. I don’t have the possibility of reaching this level of certification because my house will be built in a city on a small lot that doesn’t allow much flexibility in the orientation of the house and the openings. In the main room of my house, there will be a closed fireplace with heat recovery that will be redistributed by the closed-circuit ventilation system of the house which should minimize the electrical consumption of the heat pump. I should have in a month or two a precise estimate of the energy expenses necessary for the heating and cooling of the house because these data are essential to respect the new standards in France and therefore will be calculated by the design office. This should allow me to calculate more precisely the photovoltaic needs.
In view of a number of installations described on the internet, I had initially assumed to have a photovoltaic power of 8 kWp (20 panels). And this is probably what I will install in the first instance but with the ability to easily expand it to 12 or 18 kWp (40 panels) …
Now to return to the technical point about the Factor 1.0 rule you explain very well the justification of this rule but what I retain in the end is that: “The maximum PV power must be equal to or less than the VA power rating of the inverter / charger”. In my case the PV power of the Fronius is 3.6 kW and thus largely lower than the 6000 VA of the Victron Multi RS. The Multi RS is what is commonly called a hybrid inverter but in practice it is simply the integration of a MPPT charge controller (MPPT RS 450/100) to a conventional inverter. So, this structure is equivalent to that of a MultiPlus inverter coupled in DC to an MPPT and therefore it should not change anything to the rule of the factor 1.0. In fact the coupling of the Multi RS to a Fronius is described in section 4.13 of the manual and refers to the 1:1 rule.
One of the major problems for me currently is that the Multi RS does not handle ESS and therefore I do not think it is yet possible to use the zero feed-in of the Fronius. See what Guy Stewart (community manager) says about ESS here Multi RS Solar Hybrid - Self Consumption Optimization - Victron Community . I will try the demo of the Multi RS in VictronConnect 2.70.
It is good to know that the MultiPlus is not that noisy as this gives me an alternative.
Again, thank you very much for all your advice. I hope this can be helpful to others who are trying to define their PV system.
1 Like
I would like to come back to the 1.0 factor story. I could not understand why the power of the Victron inverter had to be at least equal to that of the Fronius inverter? So I asked the question on the Victron community and the answer is obvious! When the Fronius delivers its maximum power to a load and if that load is stopped suddenly, the energy delivered by the Fronius must be absorbed somewhere before the regulation takes place. The only place where this can happen is into the battery through the Victron inverter, and therefore the inverter must obviously support a power level equal to or greater than that of the Fronius.
By the way, in your case your 2 Fronius have a power of 16.4 kW and your 3 Victron only 15 kW so they are slightly undersized.
1 Like