modbus energy meter read-only/sniffer ?

[satmanuk2]

Evening all;

I need some guidance if possible.

I have an solar hybrid inverter that uses modbus to talk to an eastron SDM120 energy meter.

The solar inverter is already in communication with the meter very rapidly & the meter is responding with regular energy stats.

I would like to use the SDMx20 driver/variables in tasmota to read that power data, but without requesting it from the meter first (i.e. to sniff the data already on the bus)

I have already tried bi-directional SDMx20 driver coms with the energy meter from 2 devices at the same time (e.g. from the inverter & from tasmota) but while it appears to work at first, if the inverter doesnt get a timely response from the meter when it is dealing with a large load or load shedding, it starts to misbehave (e.g. not ramping power up & down properly).

I have tried the bi-directional route with multiple 8266 boards & an ESP32 (so see if it was a board speed problem), but what I really need is not to put extra load on the meter at all, as it is already giving out what i need, but tasmota won’t acknowledge it RX only.

I assume there has to be somewhere in the code that insists on send to receive, is anyone able to point me in the right direction ?

p.s. i have spent several hours searching the web before posting, but havent found anyone trying to get tasmota to co-exist with an existing meter setup read-only.

Thanks in advance

Paul C

[barbudor]

Hi
If you are sure that the meter is sendind all message that tasmota need as requested by the inverter, you can try to use the RS485 converter without connecting the TX pin from Tasmota
Then tasmota will only receive data, including the request from the Inverter but hopefully that won't be a problem

If there isn't all expected mesasge than you are in trouble as Tasmota driver is coded to be the only master and not sharing the bus (and may be the inverter too as per what you say)

[satmanuk2]

hi barbudor,

thanks for the info. if I disconnect the TX the SDM data decoding stops and the console throws SDM error 7 & SDM error 9 every time it received a modbus packet.

from looking at the modbus data on the inverter, it appears to carry all the same parameters, but there must obviously be something different, unless it is the hardcoded master/slave as you suspect (yes the inverter is likely set to be the master)

is there a way to change the master/slave in the SDM driver?

[barbudor]

Looking at the code it looks like it cannot work. Sorry for the false hope.

The driver send a read command to read a register at a given address
Then the meter respond with the value of the register without confirming the address
So as of today

1. the driver gets confused by the message from the inverter that he can't identifiy as a erad request
2. it's state machine cannot match the value sent by the SDM120 with the measurment it is expecting

To support that it would be a specific driverr which listed to the command sent by the eastron to know to which measurement the the response would match

So that requires a specific driver to be written either in C for any ESP or in Berry for ESP32

[satmanuk2]

ah, gotcha, so any viable 'read-only' solution would need to first monitor the master devices request (the inverter) & then pair it with the response from the meter.

sounds like a lot of hard work, so I can see why people just put a dedicated meter in, which is what i think i will do.

thanks for investigating/clarifying.

[Noschvie]

We need a "Man-in-the-Middle" Modbus feature :-)

[gemu2015]

This will not work with the sml driver in MODBUS mode. If you could manage to feed both RX and TX of the sniffed device to an RX of an esp then it would be easy to decode via raw mode. This would need to connect with a RS485 adapter and not use TRX at all
The SML pattern matcher could then compare the requested register number and decode the response

[barbudor]

@gemu2015

1. I was reaching the same conclusion that TasmotaModbus would not be usable and planning to use directly TasmotaSerial

2. there is no need for an special hardware or an OR gate. The MODBUS is based on RS485 and all communications goes over the same wires in differential mode
   Any RX receive all the communications whoever is TXing

[gemu2015]

@barbudor
you are right, since we are in the middle and may only listen to the bus via RS 485 adapter, don't activate TRX, at best don't connect it to the ESP but to VCC

@satmanuk2
monitor the bus by sensor53 d1

set the descriptor to raw mode
+1,5,r,0,9600,modbus

when you get a dump of the traffic you may construct the decoder lines.

[barbudor]

You just need your standard max485 adapter board and only connect to rx

The point is that I don't think that script SML will work neither because the response message doesn't tell which register it is carrying
So you need to first listen to read request to know which register will then be provided in the réponse message

My plan is to derive from Theo's generic MODBUS driver

[Noschvie]

The generic Modbus driver supports just the client part, isn‘t it? A Modbus server like libmodbus is needed, isn‘it?

[satmanuk2]

just an update, ive been working with the raw mode data & i think ive figured out what the Inverter is asking the meter. (it only ever appears to send 2 modbus requests to the meter)

request no 1 - this is the request SMI was able to decode previously (but was unstable). this is just a float for a single register pair & gives a total KWH figure.

request - 01 04 01 56 00 02 90 27
response - 01 04 04 44 a2 b2 b8 3a 44

request no 2 is much more complex & this appears to be asking for 76 registers consecutively from register 1 (basically a complete dump from the meter in a single response)

request 01 04 00 00 00 4c f1 ff
response (152 bytes of data)
01 04 98 43 7a 33 33 00 00 00 00 00 00 00 00 40 a0 a3 d7 00 00 00 00 00 00 00 00 44 9c d6 66 00 00 00 00 00 00 00 00 44 94 b3 ba 00 00 00 00 00 00 00 00 43 03 e6 66 00 00 00 00 00 00 00 00 3f 7f a3 1b 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 42 48 00 00 43 c8 cd 91 44 61 31 db 2c 56

i had to recompile tasmota with much larger serial buffers to stop it truncating/corrupting. which is also why i suspect tasmota was crashing with the SMI decoder when trying to decode this data.

now that i have this data, i dont really know what do do with it, can SMI decode this many registers in a single response?.
The response code & count doesn't seem to change (the trigger/match looks consistent), and the register locations within the message also appear consistent.

ideas are welcome, unless like you say we are looking at external decoding & not using SMI

preview of decoded response.

[gemu2015]

ok, but why is special option 7 not recognized.
i will have to debug it further, will be back at least tomorrow

[satmanuk2]

NP

p.s. i do appreciate the time you've spent on it. hopefully it will help others that want to use 2 serial devices/drivers and it isnt just me that has this crazy idea :-).

[gemu2015]

yes there have also been others that requested that.

[satmanuk2]

btw, is there a way to get the sum/total of an array's member values ?

i've gone over the doku & found [-2] gives an average (which is handy) but i couldnt find a sum

[gemu2015]

ok was misplaced #endif
now it works i testet it with your script

currently no function for array sum, but may add this in next release

xsns_53_sml.ino.zip

[satmanuk2]

phew, just before my gitpod timed out..

working, well done

reading inverter<->meter modbus data & querying 8x pylontech batteries

[satmanuk2]

[satmanuk2]

@gemu2015

all of the battery data ive worked with so far are numbers, however i'd like to extract a text value for each battery & store in a string array, however any time i declare string array e.g. is1(8 "") & save the script, the board crashes & disables the script.

is there some trick to getting string arrays enabled?

[gemu2015]

string array must not be defined in >D but in >B

>B
is1(8 "")
is1[1]="hallo"
print %is1[1]%

[satmanuk2]

thanks. im sure i tried it in B with the same result, but working now (so must have been me)

btw, will these changes (special option 7 etc) get comitted into the main build at some point ? or do i still need to compile myself in the future for this particular device (apologies, im not too familiar with how the forks & merges work)

[gemu2015]

pr done, waiting for merge

[satmanuk2]

ha, i thought i'd figured it out, but no

[gemu2015]

there were several changed in tasmota dev. may be there is an impact. i did not update my fork since a week or so.

[satmanuk2]

FYI ive compiled from current dev & replaced those 2 files from your older fork & it appears to be stable again (no crashes so far).

i am confused to say the least.

will monitor tomorrow & report back.

[gemu2015]

i again looked at your script. (you should learn how to post code for better readability, use add code icon above, see below a better readable version)

1. did you write a # sign at the end of meter descriptor ? it is not visible in your post, this is essential
2. you did not expand the script string buffers >D 128 is not enough you need
   #define SCRIPT_MAXSSIZE 128 too

>D 128
res=0
rstr=""
label=""
lraw=""
val=0
val2=""
cur=0
cnt=0
sta=""
pow=0
batno=0
dbatno=0
i:time1=0
m:volt=0 8
m:curr=0 8
m:temp=0 8
m:coul=0 8
heat1="WebSend [192.168.1.54]"
heat2="WebSend [192.168.1.77]"
trip=0
snd=""
upt=""

>B
->sensor53 r
;open serial port for RS232
res=so(16 17 115200 8N1 1024)
;declare string variable
is1(8 "")
is2(8 "")

>F
; check for serial data
res=sa()
if res>0 {
	rstr=sr(10)
	;print %rstr%
	rstr=sb(rstr 1 sl(rstr))
	label=sb(rstr 0 13)
	;make sure we have a data & not an empty line
	if sl(label)>1 {
		;get value
		val=st(rstr : 2)
		if val==0 {
			val2=st(rstr : 2)
		}

		;print decoded data.... %label% %val% %val2%
		;if we have the power header
		if sb(label 0 7)=="Power " {
			dbatno=sb(label 7 1)
			print RX battery %dbatno%
		}
		switch label
			case "Voltage "
				volt[dbatno]=val/1000
			case "Current "
				curr[dbatno]=val/1000
			case "Temperature "
				temp[dbatno]=val/1000
			case "Coulomb "
				coul[dbatno]=val
			case "Basic Status "
				is2[dbatno]=val2
			case "Soh. Status "
				is1[dbatno]=val2
		ends
		;end label length check
	}
;end serial receive
}

>S
upt=s2hms(upsecs)
; run every 4 seconds
if upsecs%4==0 {
	;call a battery on each pass
	batno+=1
	snd="pwr "+s(0batno)+"\n"
	print TX battery %batno%
	res=sw(snd)
	if batno==8 {
		batno=0
	}
	;sum up current reading as we want total not average
	cur=curr[1]+curr[2]+curr[3]+curr[4]+curr[5]+curr[6]+curr[7]+curr[8]
	pow=volt[-2]*cur
}

;state engine to calc export/import duration
if sml[3]<0 {
	if sta!="Importing" {
		sta="Importing"
		time1=1
	}
}
if sml[3]>1 {
	if sta!="Exporting" {
		sta="Exporting"
		time1=1
	}
}

if sml[3]==0 {
	if sta!="Balanced" {
		sta="Balanced"
		time1=1
	}
}
;-------------------------------------------------
;check if minimum export time threshold has been reached
if upsecs%30==0 {
	switch sta
		case "Exporting"
			if time1>100 {
				print turning on immersun
				;turn on immersun
				->power2 1
			}
			if time1>130
			and sml[3]>1000 {
				print turning on kitchen water heater
				->%heat2% power1 1
			}
			if time1>160
			and sml[3]>1000 {
				print turning on oil radiator
				->%heat1% power1 1
			}
	ends
}

>R
;close serial port
sc()
;Grid Watts {m}%sml[3]%

>W
State {m}%sta%
Delta {m}%0time1% Secs
Uptime.. {m}%upt%
Stack {m}%2volt[-2]%V | %2cur%A | %1pow%W | %temp[-2]%C | %0coul[-2]%%%

Bat1{m}%2volt[1]%V | %2curr[1]%A | %1temp[1]%C | %0coul[1]%%% | %is1[1]% | %is2[1]%
Bat2 {m}%2volt[2]%V | %2curr[2]%A | %1temp[2]%C | %0coul[2]%%% | %is1[2]% | %is2[2]%
Bat3 {m}%2volt[3]%V | %2curr[3]%A | %1temp[3]%C | %0coul[3]%%% | %is1[3]% | %is2[3]%
Bat4 {m}%2volt[4]%V | %2curr[4]%A | %1temp[4]%C | %0coul[4]%%% | %is1[4]% | %is2[4]%
Bat5 {m}%2volt[5]%V | %2curr[5]%A | %1temp[5]%C | %0coul[5]%%% | %is1[5]% | %is2[5]%
Bat6 {m}%2volt[6]%V | %2curr[6]%A | %1temp[6]%C | %0coul[6]%%% | %is1[6]% | %is2[6]%
Bat7 {m}%2volt[7]%V | %2curr[7]%A | %1temp[7]%C | %0coul[7]%%% | %is1[7]% | %is2[7]%
Bat8 {m}%2volt[8]%V | %2curr[8]%A | %1temp[8]%C | %0coul[8]%%% | %is1[8]% | %is2[8]%

;smart meter interface (READ ONLY, RX pin connected only)

M 1
+1,3,r,0,9600,Grid
1,=so3,256
1,=so7,1
;NB negative value=importing, positive value=exporting
1,01040000004cx5ffffffff@1,V,V,Volts,1
1,01040000004cx5x12ffffffff@1,A,A,Amps,3
1,01040000004cx5x24ffffffff@1,W,W,Watts,3
#

[satmanuk2]

point taken, will use code option in future.

yes there is a # at the end of the meter descriptor, dont know why it didnt carry on the copy/pasta, prob user error.

will re-compile with the SCRIPT_MAXSSIZE 128 now.

i still dont understand why it is stable replacing those 2 files, but unstable as 100% dev

[satmanuk2]

for completeness (if anyone else finds this thread), here is an overview of my project

Many thanks to gemu for all the help & adding a new SML feature to allow the 2 UARTs to co-exist on 2 different serial drivers. & barbudor for earlier guidance (I wouldn’t have got this far without them, given that ive used tasmota scripting for all of a week!!)

comments welcome ;-)

Objectives

• to read (sniff) the modbus data between my solar inverter & it's Eastron SDM120CT meter, without needing to add a second dedicated meter & clamp into the solution. i later found out that the solis inverter asks the meter for all registers in one dump, but i was only interested in the first 3 registers (volts, current, power)

• to gather data from my stack of 8 pylontech US3000C batteries (volt, current, temp, SOC, SOH, error codes etc) – this data to be used for further logic control & general health awareness of the stack at individual battery level (something the solis inverter does not provide). N.B. the serial str read code was taken from an earlier post about reading a US2000 i beleive.

• to enable my immersun DHW immersion controller heater outputs, only once certain conditions are met e.g. minimum battery SOC, exporting power for xx period, between sunrise/sunset etc. without this logic I found that the inverter & immersun would sometimes fight each other for load control, especially after a large load was turned off (e.g. Oven) and the inverter takes several seconds to ramp the power down.

• to send commands to other tasmota plugs/contactors to enable additional loads such as heaters to come on ahead of their scheduled time, to maximise any surplus power generation.

• to monitor the grid state and enable the EPS contactor if the grid fails (a contactor changes over several plug circuits in my house onto the EPS) I could not keep the house plugs on the EPS circuit all the time, as the solis inverter annoyingly shuts down the EPS circuit momentarily when the battery reaches its DOD%. thus, causing a rapid flicker & some electronic devices to stall & need resetting. A mechanism to delay the EPS changeover was required, plus a small, dedicated UPS on critical things like WIFI/router etc.

• to be an independent device (non MQQT) that is permanently powered by the EPS circuit of my inverter, again in the even of a grid outage this device needed to be able to act on its own without a broker/host involved.

Some constraints/comments
-I Cannot actively communicate with the meter on the modbus network, as this caused unstable inverter behaviour, so can only listen to modbus responses between the meter & the inverter, therefore only RX pin is connected to ESP32 & meter descriptor has no TX config.

-I wanted to avoid monitoring the pylontech Canbus for similar reasons (It already makes me nervous having all that lithium battery in house). The pylontech US3000C console port seemed to be the logical choice & would avoid any undesired issues between the inverter & master battery over canbus.

-Each pylon battery is queried sequentially (all via the master battery cable) by specifying its power index (1-8 in my case). the script then detects which battery has responded & fills an array with the data acordingly.
There is another command that returns data for all batteries in one go, however I didn’t opt for this route, as it is a much larger buffer to process & there are no per metric ascii delimiters (I might work on this though, as querying them all one by one does take 20-30 seconds to build up the stack averages/totals)

-Note that my batteries are stacked vertically (short edge), 4 above 4 in a custom-made rack. Batteries 1-4 are above batteries 5-8, hence you will note an average 3 degree higher temperature in batteries 1-4.

-I have only pasted the code to sniff the SDM120, read pylon battery data & a basic websend trigger, as the rest of the logic/outputs will be unique to my implementation.

Hardware
ESP32 WROOM devkit (ESP8266 lacked the 2nd UART, IO & grunt for this task)
TTL-RS485 board (from amazon/ebay, pictured is the actual board I used)
TTL-RS232 adapter (an old sparkfun DB9 level shifter board i had spare)
Serial to RJ45 with pylon gnd, TX/RX pins only.
5V relay module powered from Vin on the ESP32 (generic az delivery board) used to engage the larger 32 & 60A contactors
A 240V neon indicator lamp in a small light proof case with an LDR-TTL adapter for grid detection (this seems low tech, but compared to pricier modules to opto isolate grid detection & filtering out grid fluctuations in restarts or capacitor discharge if using a PSU, this solution works surprisingly well)
A 5V 2A USB power supply (on the EPS circuit). It has enough capacitance to survive the previously mentioned flicker caused by the EPS change in the inverter @ DOD.

schematic (excuse ms paint)

compile includes
#ifndef USE_SCRIPT
#define USE_SCRIPT
#endif
#ifndef USE_SML_M
#define USE_SML_M
#endif
#ifndef USE_SML_SCRIPT_CMD
#define USE_SML_SCRIPT_CMD
#endif
#ifndef USE_SCRIPT_SUB_COMMAND
#define USE_SCRIPT_SUB_COMMAND
#endif
#ifdef USE_RULES
#undef USE_RULES
#endif
#ifndef USE_SCRIPT_WEB_DISPLAY
#define USE_SCRIPT_WEB_DISPLAY
#endif
#ifndef USE_SCRIPT_SERIAL
#define USE_SCRIPT_SERIAL
#endif
#ifndef SCRIPT_LARGE_VNBUFF
#define SCRIPT_LARGE_VNBUFF
#endif
#ifndef MAXSVARS
#define MAXSVARS 30
#endif

script

D 128
res=0
rstr=""
label=""
lraw=""
val=0
val2=""
cur=0
cnt=0
sta=""
pow=0
batno=0
dbatno=0
i:time1=0
m:volt=0 8
m:curr=0 8
m:temp=0 8
m:coul=0 8
heat1="WebSend [192.168.1.54]"
heat2="WebSend [192.168.1.77]"
trip=0
snd=""
upt=""

B
->sensor53 r
;open serial port for RS232
res=so(16 17 115200 8N1 1024)
;declare string variable
is1(8 "")
is2(8 "")

F
; check for serial data
res=sa()
if res>0 {
rstr=sr(10)
;print %rstr%
rstr=sb(rstr 1 sl(rstr))

label=sb(rstr 0 13)
;make sure we have a data & not an empty line
if sl(label)>1 {

;get value
val=st(rstr : 2)
if val==0 {
val2=st(rstr : 2)
}

;print decoded data.... %label% %val% %val2%

;if we have the power header
if sb(label 0 7)=="Power " {
dbatno=sb(label 7 1)
print RX battery %dbatno%
}

switch label

case "Voltage "
volt[dbatno]=val/1000

case "Current "
curr[dbatno]=val/1000

case "Temperature "
temp[dbatno]=val/1000

case "Coulomb "
coul[dbatno]=val

case "Basic Status "
is2[dbatno]=val2

case "Soh. Status "
is1[dbatno]=val2

ends
;end label length check
}
;end serial receive
}

S
upt=s2hms(upsecs)

; run every 4 seconds
if upsecs%4==0 {

;call a battery on each pass
batno+=1
snd="pwr "+s(0batno)+"\n"
print TX battery %batno%
res=sw(snd)
if batno==8 {
batno=0
}

;sum up current reading as we want total not average
cur=curr[1]+curr[2]+curr[3]+curr[4]+curr[5]+curr[6]+curr[7]+curr[8]
pow=volt[-2]*cur

}

;state engine to calc export/import duration
if sml[3]<0 {
if sta!="Importing" {
sta="Importing"
time1=1
}
}
if sml[3]>1 {
if sta!="Exporting" {
sta="Exporting"
time1=1
}
}
if sml[3]==0 {
if sta!="Balanced" {
sta="Balanced"
time1=1
}
}
;-------------------------------------------------
;check if minimum export time threshold has been reached
if upsecs%30==0 {
switch sta
case "Exporting"
if time1>100 {
print turning on immersun
;turn on immersun
->power2 1
}

if time1>130
and sml[3]>1000 {
print turning on kitchen water heater
->%heat2% power1 1
}

if time1>160
and sml[3]>1000 {
print turning on oil radiator
->%heat1% power1 1
}

ends
}

R
;close serial port
sc()

;Grid Watts {m}%sml[3]%

W

State {m}%sta%
Delta {m}%0time1% Secs
Uptime.. {m}%upt%

---

Stack {m}%2volt[-2]%V | %2cur%A | %1pow%W | %temp[-2]%C | %0coul[-2]%%%

Bat1{m}%2volt[1]%V | %2curr[1]%A | %1temp[1]%C | %0coul[1]%%% | %is1[1]% | %is2[1]%
Bat2 {m}%2volt[2]%V | %2curr[2]%A | %1temp[2]%C | %0coul[2]%%% | %is1[2]% | %is2[2]%
Bat3 {m}%2volt[3]%V | %2curr[3]%A | %1temp[3]%C | %0coul[3]%%% | %is1[3]% | %is2[3]%
Bat4 {m}%2volt[4]%V | %2curr[4]%A | %1temp[4]%C | %0coul[4]%%% | %is1[4]% | %is2[4]%
Bat5 {m}%2volt[5]%V | %2curr[5]%A | %1temp[5]%C | %0coul[5]%%% | %is1[5]% | %is2[5]%
Bat6 {m}%2volt[6]%V | %2curr[6]%A | %1temp[6]%C | %0coul[6]%%% | %is1[6]% | %is2[6]%
Bat7 {m}%2volt[7]%V | %2curr[7]%A | %1temp[7]%C | %0coul[7]%%% | %is1[7]% | %is2[7]%
Bat8 {m}%2volt[8]%V | %2curr[8]%A | %1temp[8]%C | %0coul[8]%%% | %is1[8]% | %is2[8]%

;smart meter interface (READ ONLY, RX pin connected only)

M 1
+1,3,r,0,9600,Grid
1,=so3,256
1,=so7,1
;NB negative value=importing, positive value=exporting
1,01040000004cx5ffffffff@1,V,V,Volts,1
1,01040000004cx5x12ffffffff@1,A,A,Amps,3
1,01040000004cx5x24ffffffff@1,W,W,Watts,3

[relyd]

@satmanuk2 Related #18618 Did not have time yet to compile/test this branch.

[MichaelEFlip]

Thank you for your prework here. I extended the script for the usage of Eastron sdm630 meter (three-phase) which is polled by a Solis hybrid inverter.

>D
>B
=>sensor53 r
>M 1
+1,3,r,0,9600,Grid
1,=so3,31
1,01040034000a31c3010414ffffffff@1,Total Power P,W,Tot_P,2
1,01040034000a31c3010414x4ffffffff@1,Total Current I,A,Tot_I,2
1,01040034000a31c3010414x8ffffffff@1,Total Apparent Power,VA,Tot_VA,2
1,01040034000a31c3010414x8x4ffffffff@1,Total Power Factor,,Tot_PF,2
1,01040034000a31c3010414x16ffffffff@1,Total Angle,°,Tot_Ang,1
1,0104000c0006b00b01040cffffffff@1,Power P L1,W,P_L1,2
1,0104000c0006b00b01040cx4ffffffff@1,Power P L2,W,P_L2,2
1,0104000c0006b00b01040cx8ffffffff@1,Power P L3,W,P_L3,2
1,010400000006700801040cffffffff@1,Voltage U L1,V,U_L1,2
1,010400000006700801040cx4ffffffff@1,Voltage U L2,V,U_L2,2
1,010400000006700801040cx8ffffffff@1,Voltage U L3,V,U_L3,2
1,010400060006900901040cffffffff@1,Current I L1,A,I_L1,2
1,010400060006900901040cx4ffffffff@1,Current I L2,A,I_L2,2
1,010400060006900901040cx8ffffffff@1,Current I L3,A,I_L3,2
1,01040046000691dd01040cffffffff@1,Frequency,Hz,Freq,2
1,01040046000691dd01040cx4ffffffff@1,Import Energy,kWh,Imp_E,3
1,01040046000691dd01040cx8ffffffff@1,Export Energy,kWh,Exp_E,3
1,010400120006d00d01040cffffffff@1,Apparent Power S L1,VA,S_L1,2
1,010400120006d00d01040cx4ffffffff@1,Apparent Power S L2,VA,S_L2,2
1,010400120006d00d01040cx8ffffffff@1,Apparent Power S L3,VA,S_L3,2
1,010400180006f00f01040cffffffff@1,Reactive Power Q L1,VAr,Q_L1,2
1,010400180006f00f01040cx4ffffffff@1,Reactive Power Q L2,VAr,Q_L2,2
1,010400180006f00f01040cx8ffffffff@1,Reactive Power Q L3,VAr,Q_L3,2
1,0104001e0006100e01040cffffffff@1,Power Factor L1,,PF_L1,2
1,0104001e0006100e01040cx4ffffffff@1,Power Factor L2,,PF_L2,2
1,0104001e0006100e01040cx8ffffffff@1,Power Factor L3,,PF_L3,2
#

I struggeled a lot to debug and to get this to work, so here some hints:

• I used Hterm to send the hex values to the esp32. You have to fill the buffer (defined by '1,=so3,31' = 31 bytes) completely to have the values read or appear in the web interface. Best is to repeat sending the message.
• If the buffer is too small (here defined by '1,=so3,31' = 31 bytes) the message cannot be read
• If the buffer is too big (here defined by '1,=so3,31' = 31 bytes) the read out is delayed
• There has to be an hash "#" at the end of the script (this is missing in the last post from satmanuk2 due to formatting)
• If you show all messages on the console using the command 'sensor53 d1' the web interface will not update the values. You have to stop displaying all messages on the console using the command 'sensor53 d0' first
• Be aware: The CRC check sum in the modbus messages is not compared. Bit errors may lead to irregular wrong figures.
• Check if your TTL/RS485 has an 120 ohm termination resistor on the board. If you are using the board to sniff inbetween you have to remove this resistor. Termination should be done at each end (so normally at Eatron and at Solis).