diy solar

diy solar

Wireless solution to feed CT clamp data to hybrid inverter

Hi Folks,

God bless i found this thread. I was trying to achieve the same thing. Unfortunately I already spent a lot of time trying to make it work with EBYTE E32 RF433 transceivers. I'm unable to make it work so far. Meter COM LED just doesn't come up, despite config on E32 being correct. Does anyone have success with with those?

Or i guess I'll go with already tester method via ethernet. I found that there is also Waveshare to Wifi DIN module available.



BTW pinout I'm using from RJ45 to unit are pins 3+7 and 6+8 as per wiring diagram. I've tested it with solis directly connected and it works:
 

Attachments

  • Screenshot 2023-07-18 at 20.42.04.png
    Screenshot 2023-07-18 at 20.42.04.png
    114.2 KB · Views: 10
Or i guess I'll go with already tester method via ethernet. I found that there is also Waveshare to Wifi DIN module available.
Sure. But more expensive and, personally, I'd choose wired rather than WiFi for critical systems any day. Also note that if you ever need G100 compliance, it must be wired, not wireless.
 
Sure. But more expensive and, personally, I'd choose wired rather than WiFi for critical systems any day. Also note that if you ever need G100 compliance, it must be wired, not wireless.
I guess for G100 ethernet cable would not work either.
I'll be doing G98. It is limited in inverter anyway. I'd have to drill a lot of walls to get cable, so no go for me.
I personally don't see a lot of issues with p2p wifi option, latency will be low and in terms of reliability it will be the same as ethernet. I've checked what meter is sending and it is like one or couple messages per second, so it will be fast enough. Latency for power output inside the inverter are like 10x higher than this.
 
Direction of AC power flow detection relies on AC phase measurement between AC voltage and AC current.

Putting extra delay in current sensor output signal path can screw up power flow direction sensing.

Inverters with ability to block grid back feed usually limit back feeding when grid incoming current drops below about 3 to 5 amps. This is to allow a margin of error to avoid measurement inaccuracy of CT sensor allowing inverter to produce some grid reverse feed current.

Just a long twisted pair wiring from CT sensor to inverter can cause enough phase error due to parasitic capacitance to impact power flow direction measurement. Since a twisted wire pair has a transmission line impedance of close to 100 ohms, there is often a 100 ohm 'burden' resistor placed in inverter to properly terminate the twisted pair line coming from CT sensor.
Topic subject line might be misleading here.

Discussion here is not putting stuff between CT clamp as such, it is to transfer already harvested readings for longer distances. There are no alterations between CT and meter, where it is sensetive.

CT=>Meter=>Inverter - in this route we're looking at meter=>inverter link, which even according to many meter manufacturers can be extended as long as required (RS485 Modbus over twisted pair).

What discussed here is introducing tcp/ip for transmitting modbus (either using UTP or wirelessly) CT=>Meter=>Ethernet=>inverter.

Hope this helps.
 
I was pointed here from someone on a Facebook group, and this setup ended up working perfectly, but I did have a few issues that I'll go through to hopefully help someone in the future...

The in laws had a Solis 3kW string inverter installed in an outbuilding, but decided they absolutely needed a battery (mother-in-laws sister got one)
I was tasked with updating, and saw the opportunity to offload my Solis 3kW hybrid and upgrade myself to a 6kW hybrid, as long as I could sort the meter issue out. I'd already established a powerline ethernet link for the data logging stick, so just needed some way of running RS485 over that. Enter this thread, and the Waveshares...

There was a spare row in the house consumer unit (former off peak half of the board) so the Acrel ACR10R meter, Waveshare and power supply were mounted there. Everything set up and configured per this thread, blue lights, packets being transmitted and received, everything looked good except the inverter said 'RS485 fail' and the 'Com' light on the meter wasn't illuminated or flashing.

After a good few hours of messing about, I found the issue. Despite the markings and the manual, the Acrel ACR10R only seems to want to communicate via the RJ45 socket on pins 3 and 6. I had a load of shielded patch leads around 30cm long left over from a SCADA upgrade at work, but the ends were made off non standard, so pins 3 and 6 were across different pairs on these weird leads. Consequently, I was using pins 1 and 2, which were a pair on my leads, and claimed to be OK on the diagram on the meter itself and the manual. I also tried using just a bit of twisted pair straight into the terminals at the bottom (21 and 22 off the top of my head). Still, no dice. I also tried using pins 3 and 6 despite them being on different pairs, just in case, that didn't work either.

Obviously many hours passed as I was sure it wasn't that, but it was only until I did a continuity test between all the pins that were all supposed to be the same thing I brought myself to cut up a standard patch lead, using pins 3 and 6 which were the same pair, and it all sprung into life. Everything's now working perfectly and I'm awaiting delivery of a Puredrive battery to complete the setup.
Could you please confirm which pins you've used for Acrel? I'm a bit struggling to make it work with the ethernet server. For inverter I've used 3+7 and 6+8 and it worked fine with Solis.
 

Attachments

  • Screenshot 2023-07-19 at 18.12.28.png
    Screenshot 2023-07-19 at 18.12.28.png
    486.6 KB · Views: 9
If you are in the Victron world a Shelly 1EM (Wifi) works well (with the added benefit of being able to use a second CT clamp on an EV circuit if you have one)

It does not have the terrible lag that some Victron meters have




Screenshot 2023-07-20 112424.png
 
If you are in the Victron world a Shelly 1EM (Wifi) works well (with the added benefit of being able to use a second CT clamp on an EV circuit if you have one)

It does not have the terrible lag that some Victron meters have




View attachment 158361
Shelly meter is very nice, unfortunately i don't think it will work with my Solis inverter.

I actually found this bridge:

benefit is it is 220v and you don't need additional PSU.

Unfortunately I still can't make stuff work with Acrel, it is just not sending anything to the bridge
 
Could you please confirm which pins you've used for Acrel? I'm a bit struggling to make it work with the ethernet server. For inverter I've used 3+7 and 6+8 and it worked fine with Solis.
From the RJ45 connection on the Acrel connect green to RS485A on the Waveshare adapter and green-white to the RS485B. Common/ground is not required.

BTW: For some reason the individually marked RS485 pins on the Acrel didn't work for me.
 
From the RJ45 connection on the Acrel connect green to RS485A on the Waveshare adapter and green-white to the RS485B. Common/ground is not required.

BTW: For some reason the individually marked RS485 pins on the Acrel didn't work for me.
I'm pretty sure I tried it, going to do it again.
Yeah pins 21/22 do not work at all.

PS. I got another meter - Eastron SDM120CT-Modbus, i hope it works, if so, it will also take less space on a din rail.
 
OK, Got Eastron today, however same issue... really not sure what I'm doing wrong...

it is just doesn't send any data on enternet bridge RS485 port

1689947410659.png


Interestingly if i connect bridge to inverter, inverter actually sending data:
1689947510390.png
there are not many settings to play with really...

1689947552727.png
 
Ok, I made it work. If i select protocol to NONE instead of Modbus it weirdly works....
Now I'm facing another issue, after it starting to communicate after few minutes I'm getting ileak-pro01 error message and inverter shuts down. It happens only if I connect this RS485 to Ethernet server, without it everything works just fine, not sure why it is causing DC leakage to trip, doesn't make sense to me.
 
This is a very interesting thread and a solution to a common problem when faced with a retro-fit, when existing system is remote from incoming electric supply.
I would like to install a similar system. I have one ethernet cable which was installed into shed but never used and, via a switch located in a home office, it has the potential of connecting to a router in the main house. I need 2 x connections from the Huawei SUN2000 Hybrid inverter - 1. WLAN (or Wifi but no signal) connection for comms and 2. RS485 to the remote DDSU666meter near incoming supply(at house). I therefore think I need -
INVERTER END
WLAN connection from inverter to an ethernet switch which is connected to the existing ethernet cable
RS485 connection converted to Ethernet using waveshare converter or similar and plugged into same switch
METER END
Another switch (to 'split' the ethernet?)
Ethernet patch lead from switch direct into router for comms to inverter
Ethernet patch lead into ethernet / RS485 converter and wired into connection on DDSU666 meter

Is this correct folks?

Also are the RS485/ethernet converters all bi-directional and can you forsee any issues with using PoE devices?
 
Ok, I made it work. If i select protocol to NONE instead of Modbus it weirdly works....
yes.. odd.
Now I'm facing another issue, after it starting to communicate after few minutes I'm getting ileak-pro01 error message and inverter shuts down. It happens only if I connect this RS485 to Ethernet server, without it everything works just fine, not sure why it is causing DC leakage to trip, doesn't make sense to me.
@cliffhanger had a similar error that was fixed with a Solis firmware update - see thread
 
yes.. odd.

@cliffhanger had a similar error that was fixed with a Solis firmware update - see thread
I've made some more progress. It works now, but signal periodically drops. Inverter shows communication OK for 5-10s and then it drops to fail, then back to OK. can't figure out what's causing it.
 
For the purity of experiment I got exactly same devices today - waveshare r485 to eth.
I did setup everything and it is working for few second, then it throws met slt fail error. I really do not understand what I'm doing wrong.
Can someone post screenshots from waveshare, one from meter side and one from inverter side, just to ensure I'm not missing anything.

PS. Also, it looks like Solis is RS485 master, not slave, so he is requesting and meter is responding, not other way around.
 
Hi Folks,

God bless i found this thread. I was trying to achieve the same thing. Unfortunately I already spent a lot of time trying to make it work with EBYTE E32 RF433 transceivers. I'm unable to make it work so far. Meter COM LED just doesn't come up, despite config on E32 being correct. Does anyone have success with with those?

Or i guess I'll go with already tester method via ethernet. I found that there is also Waveshare to Wifi DIN module available.



BTW pinout I'm using from RJ45 to unit are pins 3+7 and 6+8 as per wiring diagram. I've tested it with solis directly connected and it works:
0
 
Hi All,
I've read this thread several times and It's full of interesting stuff. So I have a Question.
I've just fitted a Growatt SPH3000 down the garden with 6.6kw storage and 3.6kw PV panels. This has been wired using the suggestions from this thread CT coil to smart meter, RS485 to waveshare, Direct connection to router-POE switch in shed, waveshare to RS485 to inverter.
All seems good? So my question: Has anybody had and problems with the amount of energy feed back to the grid?
I had one day when i didn't feed into the grid as i only want my system to feed my household and not the grid, Growatt setup with export limit on and set to zero, Priority set to Batt first. What could cause my inverter to feed to grid? Is it the delay on the RS485 over ethernet?
All help would be helpful.
 
The inverter has its own lag, on some you can calibrate the set point to be say 40 or 50w incoming as the target point to reduce feed in. I use a mk2pvdiverter that monitors at 50 times a sec and I am still feeding a 0.2kWh per month
 
The inverter has its own lag, on some you can calibrate the set point to be say 40 or 50w incoming as the target point to reduce feed in. I use a mk2pvdiverter that monitors at 50 times a sec and I am still feeding a 0.2kWh per month
I'm not sure if this will help? I'm not bleeding a few kwh's per month I've given away 5.5kwh's in the last 3 days :fp2
 
I'm not sure if this will help? I'm not bleeding a few kwh's per month I've given away 5.5kwh's in the last 3 days :fp2
So rather than a calibration issue then a setting is wrong or the firmware is mince. You best post screenshots of all your settings in the hope someone with a similar inverter can see the issue. Being written in Chinglise the setting headings are prone to meaning the opposite of what a native english speaker would assume.
 
Back
Top