Here’s my burning question that keeps me awake at night.
First, this my two month old setup:
48V 3Kva Multiplus II
150 | 35 MPPT
Venus GX
BMV 712
4 x 200Ah AGM bank
4 x 410W Canadian PV (2 x paralleled in serie with another 2 paralleled)
Multiplus II is 30 meters away from the main db. Had to lay 4core 10mm2 armour cable for this
As a result of this, I have no external energy meter to talk to the inverter because the distance is too great.
Critical light consumers had been moved to a 2nd db but remains tied to the main db via 3-way bypass switch, for in case batteries are drained to much or the inverter blows etc
Multi is powered through AC1 In
No alternative source on AC2 In
The critical loads db is fed from AC1 Out
No loads connected to AC2 Out (which I understand is only live whilst the grid is available
I was advised not to make use of AC2 Out, as this would overwork my small 3 Kva
So here is my burning question:
On the VRM portal, I’m able to see an immediate spike in power when critical loads are activated. So it is clear the inverter knows about them.
But this is not the case with the non critical loads that remains on the main db.
So in the first instance, this makes me worry because my first thought was that AC1 In must get its power from directly at the main breaker in main db, so as to “see” ALL power consumed by ALL loads, critical and not critical. Don’t think mine was wired up as such.
But secondly and more importantly, " if my Multi doesn’t “see” all this power, am my non critical loads still being powered with PV and only supplementing the shortfall with grid power when there is ample PV available? Because I have no way of telling this currently.
I have this sickening feeling that for the Multi to be aware of the non critical loads, these would have to be connected to AC2 out - which will cause me tremendous heartache and pain and funds to rectify - remember the 30 meter distance…
Guys, in your explanations, please write slowly and clearly, or be prepared to be bombarded with clarifying questions, as I’m not fastest of graspers lol!
You can also use a Carlo Gavazzi and install it in the main DB. The inverter will then send power to its input side to try and supply those loads that was suppose to be on AC2
This is not entirely true. Using AC out 2 or a Carli will have exactly the same effect, the inverter can only contribute 2.4 kw or 3kva to the loads, irrespective of where the are connected.
No it isn’t. It’s friggin modbus-rtu at 9600 baud. That can go hundreds of meters…
At that distance you may want to terminate the cable with 120Ω resistors. There is a trick on the meter (look in the documentation, I can never remember) to activate termination, but on the other side you may have to add your own. Measure the resistance across the data wires (yellow and orange), it should be around 60Ω if you want to be 100% sure.
Importantly though, when you splice the cable, make sure you make a good splice, and use good cable (cat5 or cat6 is good even if you don’t use all the strands). The support people don’t act too kindly when they have to chase your issue for three days only to discover there’s a bad cable in the mix.
If cabling is not your thing, you can also install a pair of Zigbee repeaters.
I’d prefer not to route any additional cables for the moment, but if I absolutely have to, I suppose the data cable would be the lesser of the two evils, the other being laying another 30m armour cable…
But, I have not done cable splicing before and would not even know how to that 12o Ohms resistor things.
What is the distance of these frisbee things? How the crow flies, the actual distance would be closer to 15m with 3 walls between the main db and the inverter.
Does one of the go into the Venus I suppose?
What do they cost? maybe this is my saving grace if this is sure to sort out my issue.
I actually have no idea… Google says 130 feet, around 35 meters. Seems a tad close for comfort.
I suspect this will cause you to forget bout the previous question anyway A pair is going to set you back €215, which is around 4k. Probably not worth it.
You can get the DTK modules directly from China, and set them up yourself (the USB side is a coordinator, the RS485 side a router, all at 9k6 baud). A quick eBay search seems to indicate prices around R750. Just be sure to get a pair that works together. The DRF2618/2619 works together, and the DRF2658/2659 works together, but you can’t mix them up.
Or… you can get the Carlo Gavazzi EM24 ethernet model. Then the problem becomes a network one and you can bridge it using Wi-Fi. But the EM24 ethernet model is €245… also on the other side of 4k…
So it actually seems as if an ebay pair of Zigbee modules might be what you want to look at.
Comes from here: In contacting my supplier, as Swartkat got a Victron dealer in JHB to install for him, I asked them if one can go 30-100m from Carlo to Venus. They advised that in their experiences, they tended to pick up issues with data from the Carlo to the Venus, to keep it within the cables max 5m length.
The Zigbee’s was also suggested, the ones on EBay much cheaper.
Swartkat 's installer, being a “high profile” “highly recommended” Victron installer, I have gotten an idea to get a thought to form a feeling to maybe ponder on the idea that maybe his installers have not focused on his install, same as they would have on a larger client. Swartkat’s one is simple and straightforward, just the distance from the main DB and inverter, the challenge.
If I had a spare Carlo, I would have lent it to test on a long UTP cable.
AC_Out 1 and 2 in Swartkat 's case, is a rather big cost issue ito distance of 30m armored cable.
The perfect most cost effective solution would be a Carlo with a long UTP cable.
Oh I’ve had my share of issues too, but that is why I said do a good splice and use decent cable. By a good splice, I mean 1) solder the wires, don’t just twist them together, 2) heat-shrink the connection when you’re done, 3) then tape it up neatly or throw a larger bit of heatshrink over the whole thing.
I have seen issues with modbus cabling used in noisy environments, and I mean REALLY noisy environments. So either use good shielded cabling, or at least don’t run the cable in the same trunk as all your other cabling.
Finally, terminate the cable, that’s literally just a 120Ω resistor across the data lines. On the meter side it is easy (just screw it into the terminals with the wiring), on the other end it might be a bit more difficult to do neatly.
If you do that, there should be no reason why it would not work just fine.
My own house runs with a spliced cable, probably about 10 meter (so not that much). No issues at all.
Adding to this, I once also went 10m, I took 2 x UTP wires connected to each of the 5m RS485 to USB interface cable instead of just one UTP wire per interface wire.
Me thought that would be less resistance … if you agree Plonkster.
I think I might try the lan cable route then, though I will need to get assistance with cable splicing and adding resistors etc…
In my searches today, I also stumbled upon a Victron product. What are your thoughts on the ET112 Energy Meter? Better than the Carlo? Cheaper? More expensive?
Maybe they made bad connections, on the mine where I use to work for many many years, we would run RS485 and Modbus signals for up to 300 meters and sometimes even more with no issues, if the joint is made correctly @Swartkat will have no problems.
A piece of cable vs the cost of the Zigbees, well for me the cable wins hands down… Over 1.5k for the zigbees alone…
I have the exact same inverter and have connected my mains incomer through a double pole CB and change-over into the AC-In, then split the loads to AC Out1 and AC Out2. It works well and the family are well trained to only use 1 appliance at a time. It has only been a month, but no trips so far.
I have a 4600W array and the AC Out2 runs on solar during the day.
I did the install myself and now you have me worried if I did the correct thing by going all-in through the inverter.
A pair is going to set you back €215, which is around 4k. Probably not worth it.
