Modbus RTU problems CT Commander SK Drives

Kindly advise me if there is any solution to the problems faced by us at site.

The Modbus RTU Master is Johnson’s Control Gateway “ Metasys Integrator”.

This gateway has RS232 Output port

We have connected a RS232-485 converter and have connected 8 AC Drives SK series of Control Techinque to it.

The distance between the Master and the last drive is about 350 Mts.

We are able to communicate with the drives Using “Metasys Integrator”, However the communication is not very stable.

Using Modscan PC based program we tried to analyze the polling results. (http://www.win-tech.com/html/modscan.htm)

The results are as Follows:

Out of 100 polls 20 poll requests results in error, The CRC error is the maximum 80% of the failure and 20% of error we get is Time out.

We have reduced the baud rate to 4800 with no great change in the failure rate.

We do not doubt the cable and the screening.

Question:
----
In master “Metasys Integrator” (Johnson’s Master Gateway) can we program to define failure of value if consecutive 3or 4 poll failures occur then declare the value on the BMS screen as “***** “ .

1. How do we get rid of the CRC errors
2. Could there an issue with the firmware of the Drives ?
3. The Drive Manufacturrer specifically states that the RS232-485 converter to be used should be “Amplicon Magic 485F25 or Magic 485F9”
   1. Any other used converter must have the hardware and software support to tri-state the transmit buffer following the message transmission, Else the drive will not respond.
   2. We are using “ICP con” 485 converters, this works with other Modbus Integrations well, not sure if this is supporting the so called TRI STATE transmit buffer specifically mentioned by the Drive manufacturer, can any one provide more details?

I presume most of the 485 converters block message transmission from slaves when master is ready to send a message.

Johnson’s Metasys Intergrator People always poll a Block of 16 holding register, we have asked them to reduce the Block/ group registers to 5 as only 5 values are required by them per drive. Is it not possible to define in this gateway to make request only for Just 5 Registers instead of always 16 registers per drive loading the buss with useless values.

There are another set of Values that are required from the drive which are not in the consecutive addresses but since we do not know great details of the VCT file that defines the Modbus implementation in Metasys Integrator , we are unable to tell them how to fetch these values. We do not have gret knowledge of Metasys Integrator and its VCT file , can some one help us so that we can indicate the same to the Integrator about his own GATEWAY.

Is there any parameter where we can define the number of retries the master (Gateway) should make before it ensures ***** values appear on the BMS ( SCADA system)?

Waiting for PLC GUY : bernie_carlton

sorry for pushing this post up.

I had seen some post by PLC GUY berni carlton who had earlier posted some issue regarding Control Technique SK serese MODBUS RTU implementation. I hope he could be in a position to throw some light on this issue regarding special RS485 converter to be used with these drives.
Regards

My experience is with the AutomationDirect PLCs which have RS485 and MODBUS protocol built in. The connection, as detailed at my web site, was easy. I would suspect that you may have other problems (stop bits?).

Dear bernie,

Well I have tried with 1 Stop Bit and 2 Stop Bits at masters level. Unfortunately the Master level we are not in a position to write a protocol as it is a standard gateway , However we are first testing with the modscan tool first.

I could not locate the setting parameter in the drive for this stop bits, I guess these drives have Just the Baud rate and Address setting parameter , If there is any setting in the Drive FOR stop bit please do let me know and correct me on this.

I am now trying to get a 485 converter with 485 Turnaround Time set as 10 ms inside the 485 converter.

If i succeed in this i will send you additional notes that you could add in your site, so wish me luck.

These SK Drives seem to be very sensitive to cable and noise , I hope there is no problem with the ICs on the SK drives, which would be my second last thought, and last would be to check with Control Technique HO if there is some issue with the MODBUS RTU implementation for NETWORKING these drives.

We are connecting 8 drives on a network with length of 400Mts. Even tried the 120 Ohms Terminator resistors to avoid any wave reflection problem , but situation is same.

In your case how many drives were connected to DL PLC and what was the distance between end to end.

Thanks for your reply,

Regards

As you found, parameter 43 sets the baud rate. Otherwise the other communication parameters are fixed at 8 data bits, no parity bits, 2 stop bits.

Also, our applications have been for near (< 10 feet) communication. The application manual (Commander SK Adnvanced Users Guide) shows suggestions for shield grounding if necessary.

Dear bernie,

Thanks for the stop bit issue, we did try with 2 stop bits.

Yes we have seen the MANUAL stating 0V of the DIVES to be connected to Shield and the ground at the 485 converter end.


However they have a foot note later which says do not connect the shield to 0V at drives if the bus cable is Long. In a longer bus cable if you connect the 0V of dives to the shield at Dives end, you have good chances of blowing up the 485 port at the SK drives, they are too sensitive and no isolation or protection exists at SK drives 485 PORT end. Even if the earthing is excellent and you do not have more than 2V AC between neutral and Earth point, the chances of your SK drives PORT blowing out are good(CASE SK drives only).

Have i learned this the hard way, yes I HAVE ...and wont ever do this again

Thanks for the feedback, Shall keep looking for more..

I have no experience with either of your devices specifically, only serial Modbus in general.

1 or 2 stop bits is not an issue. If any of the messages at all go through, then whatever you've used for the word size/stop bits/parity is correct. When these settings are incorrect, no message gets through. Your problem is a percentage problem (probably) from noise, not a 'no comm at all' problem.

If I interpret your situation correctly, you are getting some message and responses through, others not, as Modpoll indicates a percentage of CRC errors and time outs. Correct?

I don't think this is a firmware issue. I think it's a noise problem.

A percentage of CRC errors is not the fault of the master or the slave generating or interpreting the CRC value, it is an indication of a bit or bits missing. Bits go missing because the noise makes the line look like a 1 when it should be a zero.

I suspect that your drives are creating noise that is picked up in the comm cable.
I know you don't doubt your cable and screening, but drives create noise hash by chopping the AC wave form and low voltage communications is susceptible to noise drowning the signal. Although RS-485 is differential mode (one side subtracts from the other) which is inherently more noise resistant than single ended RS-232, noise can still be a problem.

How long is the RS-232 cable run?

Is it shielded cable, too?

Is the 232-485 converter box shielded from drive noise? Is there a drive or hash noise generator (some arcing/sparking device) in the same panel the converter is mounted in?

Are the serial comm cable wires twisted inside the cable beneath the shield?

Is the cable shielded (I assume so because you have faith in the screening)?

RS-485 is daisy chained: master to slave 1 to slave 2 to slave 3 etc.
Have you connected both the incoming and outgoing cables at the slaves directly to the drives, under the same terminal screws, without creating a spur length of

cable running from the comm run to a drive (like the leg of a Tee)? I hope not, spurs create reflections.

Is there a cover (noise shield) over the drive wiring that hasn't been replaced ?

Do the comm cables run in a wire tray/conduit with other wires? What kind of wiring? power wiring? Or low voltage wiring?

Have you followed to a Tee the recommendations for all cabling the drives, both the power and comm?

Is it possible to run drive #1 by itself, disconnecting the comm cable to the others, to see how its comm performance is under operating conditions?

There might be a point along the comm link at which noise is introduced, which, by breaking the comm line, at various points you could isolate one or more drives which run without seeing the noise, so that you can concentrate on solving the noise problem further down the line.

Is there any means of interpreting the Modpoll data to see if close drives run OK, but far drives have CRC errors in their messages or responses?

I suspect that the converter contains the biasing needed for the RS-485 drivers, from your comment "Any other used converter must have the hardware and software support to tri-state" I've had RS-485 systems not have biasing built into the RS-485 port which resulted in symptoms similar to yours.

The attached tech note explains biasing the need to maintain a 200mV 'gap' between the A and B lines.
RS-485_tech_note_biasing_termination.pdf

One time, a factory came up with a 4.7K ohm bias resistor between B line and 5V power ground.

The need for terminating resistors dramatically increases with baud rate. The pulse that a reflection creates on the line is much more likely to be 'seen' at 1.5Mb (Profibus rate) than at slow speeds like 4.8Kb or 9.6Kb.


(Courtesy Profibus Foundation)

I concur with B&B Electronics' opinion:
"Contrary to popular belief (and to propaganda from some other vendors), termination is rarely appropriate for RS-422/485 systems. While our engineers are happy to go into transmission line theory, the bottom line is that termination generally causes many more problems than it solves. (Jan 2007)"

Is your reference to 0V a point 0V reference for the RS-485 on the converter, which is in addition to the A & B lines?
And there's a 0V RS-485 terminal on the drives too?
Can you use 3 wires, 0V, A, & B for half duplex 2 wire RS-485?

I would tend to float the shield at the drives, like the recommendation.

Dan

WOW DANW,

That was indeed a great reply and i need some time to address all your questions point wise.

Let me study all that you have explained in details and come back.

Thanks for this detailed writeup. Makes me wonder should i really be using Modbus RTU for the drives we are presently using.

There are parallel modbus networks at the same site with identical installation but not with Drives on that network (SO no choppers) working very well.

Thanks for all the feed back and let me look into all issues and come back to you soon.

Regards

Dear Dan w,

If I interpret your situation correctly, you are getting some message and responses through, others not, as Modpoll indicates a percentage of CRC errors and time outs. Correct?

Click to expand...

CORRECT, More of Time out at present. CRC we have tried to eliminate by testing a direct connection to one Drive for testing / analyzing purpose.

How long is the RS-232 cable run?

Click to expand...

The RS232 Length = 1.5Mts.

Is the 232-485 converter box shielded from drive noise? Is there a drive or hash noise generator (some arcing/sparking device) in the same panel the converter is mounted in?

Click to expand...

The 232-485 converter is near the GATEWAY(MASTER) which is far away from the AC drives In a different room.No arcing / sparking device is near by. Shield is grounded at the 232-485 Converter end.

Are the serial comm cable wires twisted inside the cable beneath the shield?

Click to expand...

Not clear which cable we are talking about, the 485 or the 232 cable ? and we are using Twisted pair for the 485 communication bus link. However the other 485 Bus (working at the same site is not twisted one . Ours is and still we have a problem

Have you connected both the incoming and outgoing cables at the slaves directly to the drives, under the same terminal screws, without creating a spur length of

Click to expand...

No, there is a 30 cms of cable from Drive to the bus ( At the AC Drive end we have a RJ 45 terminal ( No terminal blocks), so there is a shorth length of cable there) and at the Bus end like "T" short for incomming and outgoing bus, will such a short cable create wave reflections even at such lower baud rate (4800).

Is there a cover (noise shield) over the drive wiring that hasn't been replaced ?

Click to expand...

Not very clear what is this exact point refering to, sorry.

Do the comm cables run in a wire tray/conduit with other wires? What kind of wiring? power wiring? Or low voltage wiring?

Click to expand...

yes, and it is Low voltage wiring

Have you followed to a Tee the recommendations for all cabling the drives, both the power and comm?

Click to expand...

"T" for power , sorry did not get the point here. Well the "T" connection is a manual one for the communication cable with the shield also connected manually and taped. Sorry not a very clean way although. Would this create a problem too ???

Is it possible to run drive #1 by itself, disconnecting the comm cable to the others, to see how its comm performance is under operating conditions?

Click to expand...

We did that and did not get CRC errors (very rare) but are now getting TIME our errors only, however reduced to now only 5% of polls, earlier the failure rate was 20%. This cable was laid totally outside the cable tray so there was no other power or any other cable near by, in an IT office set up.

There might be a point along the comm link at which noise is introduced, which, by breaking the comm line, at various points you could isolate one or more drives which run without seeing the noise, so that you can concentrate on solving the noise problem further down the line.

Click to expand...

This is what we are planing to do today , feeling tired , because even in under a good condition ( One drive one master, new cable laid as a test setup outside the cable trays)we were still getting Time out errors (Now only 5% we were hoping 0%)

I suspect that the converter contains the biasing needed for the RS-485 drivers, from your comment "Any other used converter must have the hardware and software support to tri-state" I've had RS-485 systems not have biasing built into the RS-485 port which resulted in symptoms similar to yours.

Click to expand...

The 232-485 converter man says he has that , But inside the drive port i have no idea of the circuit details, should the 485 port at drive end have something also?? If the AC Drives is not having this what should be done? The tri state is controlled by master usually ( 485 converter) or also the slaves should have this control ?Sorry for this strange question.

This is not a profibus installation and is Just a MODBUS RTU installation and the speed we have tried as low as 4800 baud.

I will study the details you have given for the biasing and come back after all the eperiments we conduct.

However we suspect the turnaround time too at the 485 converter end and trying to reduce the turaround time to 10 ms now and checking with that.

Is your reference to 0V a point 0V reference for the RS-485 on the converter, which is in addition to the A & B lines?
And there's a 0V RS-485 terminal on the drives too?
Can you use 3 wires, 0V, A, & B for half duplex 2 wire RS-485?

I would tend to float the shield at the drives, like the recommendation.

Click to expand...

True DAN , We have left the shield open at Drives end and not connected to the 0V trminal of the AC drive.

Will come back later today with what new things we try.

Thanks for all the info Dan.

I wish communication was just a plug and play , and we could concentrate more on the application part

Best Regards

RS-232 link appears OK, shielded cable grounded on one end, no arcing/sparking stuff to put noise on it.

RS-485 cable with shield & twisted pair should be ideal.

I can't provide a hard answer for whether or to what magnitude the 30cm spurs will generate reflections.
My gut feeling is that noise from exposed wiring, where the shielding is pulled away, at the junction of the 30cm spur to the 485 bus line could be a noise injection point, because this is in the neighborhood of where the noise generator, the drive, is.

I suspect that your "single drive" test setup has test cable that runs direct from the 232/485 converter to the drive, with no spur, correct? If adding drives with spurs increases the CRC faults, then look at the spur to bus junction as a potential noise injection point.

Timeouts:
Timeouts can occur if the master's query is not recognized by the drive, because the slave drive's address byte is corrupted in the query message. Then the slave won't generate an error response because it doesn't think the message was directed at it. But my experience is that timeouts are more often due to marginal response times. The ICP converter has 'self-tuning'. Hopefully it can flip without excessive turnaround.

Noise shield on the drive:
I've seen a drive or two that have a recessed area for wiring, even for RJ-45's, which is covered by a sheet metal cover that apparently functions not just a dust cover but as a noise shield, too. Sometimes, the covers are left off, lost, disappear, whatever, with the result that sensitive control signal wiring is exposed to electrical noise.

"To a Tee"
Sorry, this is an American idiom, meaning "follow instructions exactly". Your English is so good, I neglected to not use regional idiom.

I'm not a drive guy, but I notice from the chatter about drives on this forum and others, that there is now shielded power cabling available to reduce noise generation. So my question was whether the installation was done according the drive manufacturer's cabling requirements. Do they call for shielded power cables?

Biasing:
Having the converter provide biasing is sufficient.

Dan

Dear Dan,

I am happy and sad at the same time

You are on dot regarding many issues.

Now I need to solve it.

I just need few minutes to prepare what all we did today and finally fount the culprit , but how to nab him and put him behind i may need some valuable inputs from you.

You have almost identified the bad guy and we have also reached that conclusion.

Please give me few minutes make my write up.

I am happy, as "Now" i know i have an excellent 485 communication problem solver here on PLCTALK

With best regards

When do you sleep?

According to Widipedia, Banglore is GMT +5.5 hours, I'm at GMT - 6 hours, so we're almost half way the world from each other.

So it's night there now, and you must have had one long work day.

regards, Dan

Dear Dan ,

Yes i know its late out here.

But i enjoy your company more as you are my Identified problem solver.

Sorry ,

Have been busy on calls . Let me sit down now and get back to you.

I admire you Dan.

Regards

Hi Dan Master ,

Who are you ?

Sorry for that question , because you did hit every nail that is required to ensure these "VVF low cost Drives" don not create issues and let the demon play.

How many follow instructions strictly as mentioned in the Bible hmmm Tough to tell.. and as usual interpretations are many to same phrases.

Ok back to what we tried and to answer some of your earlier questions and waiting eagerly to save our face at site.

Bias
-----

The RS485 converter has a bias as shown below:-

oops, how should i insert the image ... let me study....sorry too tired to study today, help me how to insert PIC here. Any way you are too smart to catch all the problems without any Pictures...

Anyway the R485 has biasing Resistors. AC Drives 485 port , i have no Idea , I have called them and asked them to give me a feed back.

Tests Conducted Today.
------------------------

1. Removed the 485 BUS, and connected the PC masterRS232 ---> to 485 converter--->Twisted pair----> AC Drives RJ 45 port 30 cms cable.( No cable trays , Just outside installation , like a lab set up)
   1. Distance RS232 CABLE ( 2 Mts), RS 485 Twisted pair (60 Mts, in coil form) and then a small Link cable FROM AC drives to This 485 cable.
   2. RESULT : Problem same.. Time out errors...
2. Lowered the turnaround time of the 485 converter to less than 10ms
   1. Result : No great difference. Thus the problem of Turnaround timing being the culprit has been eliminated.
3. Removed the biasing resistors at the converter end.
   1. RESUT : WORSE condition.
4. Moved the RS485 Converter near the drive so that the 485 Bus cable length is not more than 1mts and Made the RS232 cable longer , Bad i do know , The RS232 was 20 Mts.
   1. RESULT:
      1. Perfect communication. Although it is a stupid installation.
      2. Culprit [font=&quot]declared [/font]hiding near the AC DRIVES.
5. The drive 485 port biasing we have not done, as we did not have the 5 V available from the drives 485 PORT.
6. The drive 485 Port does have 0v pin available( 3 Wires can be used ) , however we have not used that as distances involved are more than 350 to 400 Mts.

Dear DAN ....

How do i shield the drive from throwing this JUNK on 485 BUS.

I can't provide a hard answer for whether or to what magnitude the 30cm spurs will generate reflections.
My gut feeling is that noise from exposed wiring, where the shielding is pulled away, at the junction of the 30cm spur to the 485 bus line could be a noise injection point, because this is in the neighborhood of where the noise generator, the drive, is.

Click to expand...

The noise is being created at DIVES END ( Not cable tray or any other place)!!! You are correct, how do i resolve it !!! But when i place the 485 converter very near to the drives it behaves well...and does not irritate me , but then i am being a fool and not providing a solution. This is ok for testing.

Please dont ask me details of the [font=&quot]Electromagnetic[/font]/ Any other protection the Drive manufacturer should have taken care of while providing a 485 port near the chopper circuit.

Timeouts:
Timeouts can occur if the master's query is not recognized by the drive, because the slave drive's address byte is corrupted in the query message. Then the slave won't generate an error response because it doesn't think the message was directed at it. But my experience is that timeouts are more often due to marginal response times. The ICP converter has 'self-tuning'. Hopefully it can flip without excessive turnaround.

Click to expand...

Very TRUE, you are a very smart man...you sit in a remote place and identify the problem so well .. I hope i could be as smart as u are , I shall try to be like you after this ISSUE IS SOLVED .

The drive choppers are killing the question asked by the master and thus does not respond to the so called garbage sent by master . So the master makes us enjoy the time out messages. And all the time we were thinking about turnaround time for 485 switching, sorry i was totally wrong.

Now i have not provided any Biasing at drives end as i do not have 5 V at the drive end PORT.

1. Should i do biasing at DIVES END TOO ??
2. If so how ? Sorry to ask you as you are not [font=&quot]responsible [/font] person from the drives manufacturer's end.
3. How do i filter out the garbage the AC drives generate on the 485 cable ??

I'm not a drive guy, but I notice from the chatter about drives on this forum and others, that there is now shielded power cabling available to reduce noise generation. So my question was whether the installation was done according the drive manufacturer's cabling requirements. Do they call for shielded power cables?

Click to expand...

Yes true.. but they do it to protect the drive and the heating of bearing at the motor's end not allowing to damage these two equipments , but never have they addressed it from the communication end. This is also an issue which you have raised well. Let me discuss more with UK on this if i can.

My understanding is that the power shielding has not been done as they usually recommend that when the distance between the motor and the AC dives is more than 100 Mt , sorry if i am wrong on this issue. But they have been more bothered about AC drives and Motors and have ignored us communication guys even if the distances are small , did i make a very bold statement lol.. forgive me if i did .

Help me in biasing the drives port i don't know where to pick up the +5 V inside the drive ..

With best regards

And GOD BLESS you MY FRIEND !!!

Waiting for some solution .. I know u can solve this ...

One solution is to give a 485 Optical isolator at each Drives end , hmmm expensive at this moment as many drives are there in many sites..

Help me sir...

With best / warm regards

Dear Dan Guru,

The basic block Diagram for the system is as shown below:




The picture of the drives with the way the communication cable is connected is shown below


Should i run a longer spur inside a metal Tube and bring it outside the drive and then connect to my 485 bus , would that help?

Noise shield on the drive:
I've seen a drive or two that have a recessed area for wiring, even for RJ-45's, which is covered by a sheet metal cover that apparently functions not just a dust cover but as a noise shield, too. Sometimes, the covers are left off, lost, disappear, whatever, with the result that sensitive control signal wiring is exposed to electrical noise.

Click to expand...

You can see above , we never had any metal cover

Best Regards