Modbus Communication Error

This sounds like ground loop issue caused by different ground potentials between the Quasar side and the other side manifested through the RS-485 comm hardware wiring.

2 wire RS-485 is a misnomer, it's really supposed to be 3 wire, 2 communications lines and a signal reference line, but most implementations use the case ground for the reference line. The local grounds can be at different potentials.

If it were me, I'd buy an RS-485 isolator module to isolate the problem Quasar unit from the others.

Dear carl

sir much thanks to answer to problem

Actually we are using converter (Rs232to 485) having connection terminal showing signal ground(SG) and aso tx+,tx-,Rx+,RX-.but energy meter having only TX+and TX-

so how can i connect signal ground .

For using Rs 485 isolator: how many energy meter of different make can connect,is it costlier.

please as any other way to avoid ground looping in transmission line.

please help us. your information are very valid to us.

waiting for ur replay

regards

rahul

It sounds to me like there is an address conflict. Ensure each device has a unique modbus address.

dear sir,

thanks for your Replay

actually what you mean by address conflict. modbus address of both meters starts from 30000,is this is address conflict. how can we avoid this.please replay

waiting for urs answer.

regards

rahul

I assume you have the manual. If not it's here:

http://www.larsentoubro.com/lntcorporate/uploads/product/Quasar_Manual.pdf

See Appendix 2 and 3 about setting the modbus slave address.

dear sir,

we are giving all settings as per Appendix 2 and 3

regards

rahul

Rahul,

You write that, when connected together with other meters, the reading from your L&T Quasar Energy meter is not the actual value. By actual value, I assume you refer to a value you obtain from the energy source using a multimeter or the like.

However, what do mean by a reading from the Quaser meter:

1. Is this the value your eyes read from the screen of the meter?
2. Or, the value your computer reads from the meter MODBUS memory register?

If your answer is 1, then somehow your wiring connections are affecting the meter's performance. Perhaps there's a voltage drop in a common power supply?

If your answer is 2, then the computer reports a different value from that which you read off the meter's screen. The MODBUS RTU protocol includes a CRC-16 error check that enables your computer software to verify the data frame coming from the meter. What software do you use to read these meters? Does it notify you of data errors?If so, you may have ground loops or other interferences with data integrity. Are your meters connected to circuits fed from different power transformers? Can you "float" them and use a system common third wire instead of local grounds? Otherwise, you may require one or more repeaters or isolators. These are classic RS-485 communication issues. You might try using Wireshark to help you solve your problems with the MODBUS communications.

Dear sir

thanks for your replay

i have clear my problem based on your question

what do mean by a reading from the Quaser meter:

The reading shown in display is correct value. but value from modbus register is incorrect but at same time meter shows response in software(software used is modbus.pl also we checked in l&ts test software )

in out of lot of parameters from l&t meter only some of showing error value like current readings,kw,powerfactor(error means if value is 25amp it shows 60amp,30amp,rapilly varying ) at same time display shows correct vale

At same time i also vary capacitor bank step in capacitor panel but it does not correct value.

sir also i have no idea to how to correct this ground loop issue(all are feed from same transformer) because only two wire is present in modbus(tx+&tx_) how i connect third wire .

from l&t side we have no good support. even they do not know about problem

please help us its urgent.

i also include snap shot of lt software

The values you are reading are integer values for each successive Modbus register.

The problem is that the Quasar values are floating point (sometimes called 'real') values, with a format like xxx.x or xx.xxx.This evident from the Memory map:

Note that the register addresses increment by two, because floating point values need 32 bits, or two Modbus registers. The sequence is 30001, 30003, 30005, not 30001, 30002, 30003, 30004.A typical display of floating point values would show every other register, like 30001, 30003, 30005 (or 30000, 30002, 30004) because a floating point value needs two registers (32 bits) for its value.

You need to program or configure your Modbus master HMI/SCADA/PC/whatever so that it reads and interprets the data as floating point values, not integer values.

There is two other issues that are common:

1) Floating point has two major different formats, sometimes called little or big endian, or byte/word swap.If the floating point values your HMI shows are really bizarre, like 3.42E-24, then you need to figure out how to get the Modbus master to interpret the data using the other floating point format.

2) The memory map starts at 30001. But some masters address that as 30000. This is known as the 'one-offset'. It can also cause bizarre values.

To all viewers of this threads

dear sir,

thanks for all helping us we are nearly reaching solution

for l&t meters they are send data as LSB & MSB (real value)in seperate Registers so there is a 2s complement calculation is requried for combining values and also a multiplication factor for all register values combinded. then only actual reading.

thanks and regards to all for supporting us special thanks to Carl_E:

keep in touch

regards

rahul

Instrumentation

merchem limited

Dear sir,

As the part of getting actual value (l&t meter) i used plc (selec mm30302) for reading modbus the required calculations are

1. splitting real value(in Dw format) in to msw and lsw.

2.convert lsw to hex

3.prefix lsw to msw.

4.convert (3) to decimal

5.multiply with mf

this is procedure obtained from l&t .

since i used plc i can directly convert double word to hex(no need of spliting and prefix lsw and msw ) but plc does not any instruction for conversion of decimal or real to hex so i cannot perform this calculation by converting it in to hex

unfortunately no option for go to other plcs

so can u please tell me any method for conversion to hex or any other logical operation perform this operation in directly

pls help us its very useful

regads

rahul

I do not understand what the L&T format is for its values in Modbus registers.

I do not understand what number format your PLC needs to work with.

1) Which format does the L&T meter use for its Modbus registers? IEEE 32 bit floating point? something else?

If something else,

- what do they call the format

- what is an example of a value and how would it look in the registers?

be sure to label what number system you're using, decimal, hex, or binary

2) If the L&T uses floating point, for the example of a decimal IEEE floating point value, 55.32, which is hex 425D47AE in IEEE floating point,

Where the IEEE format for the example is defined as 4,3,2,1

or 4th byte = 42, 3rd byte = 5D, 2nd byte = 47, 1st byte = AE

Is the format used by the L&T meter
reg 30001: 425D
reg 30002: 47AE ?

3) Which data word formats can your PLC handle?

dear sir,

device ID conflict

from my knowledge the node id is same as slave id in modbus and i give it different for each of energy meter (like 1,2,,3,up to 16) to each meter. the slave id of qusar is 1 and nova starts from2,3.. so i think no device ID conflict occures. and in situation of all meteres are connected the qusar meter is respond only if its baud rate and parity is different from other(say 4800 and enen parity) when other meter have 9800 and none parity or vice versa.

pls replay

rahul

>please as any other way to avoid ground looping in transmission line.

Zvi gave you an answer that mirrors mine:

Can you "float" them and use a system common third wire instead of local grounds? Otherwise, you may require one or more repeaters or isolators. These are classic RS-485 communication issues.

I would add that some repeaters are not isolators, I would only use an isolator/repeater.

sorry sir i cant understand what you mean . can you explain more

pls replay

rahul

One approach to get rid of ground loops is to 'float' one (meter) or both (Meter and HMI) devices, so that they have no ground reference, but that can cause safety issues.

The other approach is put an RS-485 isolator module between them.

Not sure what you have but in some cases the application requires impedance correction resistors to be added when multiple meters are used.

Consult the setup manual and verify if the resistor(s) are needed.

If this is not the cause, it may be the configuration of your RTU input channel.

dear sir,

thanks for your replay

i have addeded impedance resistance of 120 ohm at the input terminal value still have some error

please can you tell me it is qusar meter manufacturing defect in modbus communication.is any need to replace meter also how can i correct ground loop issue in energy meter

pls replay wating for ur answer

rahul