DNET help -Replacing a Drive

Hi All,

I am looking for some advice regarding the correct steps that should be taken if I was to replace an existing (but now obsolete) Third party drive, with a Rockwell Powerflex 70.

The network scanner is a 1747-SDN residing in a chassis with an SLC500 5/05.

The adapter will be the usual 22-Comm-D.

There are several obsolete drives on the network.

I've seen several different means of achieving the above, with some subtle differences.

I'm fairly familiar with RS500/RSNetWorx for DNET etc, but looking for some expert help in the correct /most effective sequence of steps to do this.

Also, assuming that the network has been setup and operating for several years, I would guess that I will not have to make any changes between SDN card and controller?
This is currently done in the controller program by means of a copy instruction that moves I/O data between M0 & M1 buffers to/from the PLC/Scanner into Integers.
I have also noticed that there are also discrete input /output addresses used in the program (e.g. the run signal to the scanner) and have never fully grasped the difference between the "M0/M1/Integer data" AND typical discrete addresses such as "I:2 / O:2" etc also used for each device on the network when it comes to the scanner mapping and PLC code

My thoughts on the correct steps would be:-

1.Go online to the current network using RSNetworx and backup the working configuration as a fall back.

2. Go Offline and remove the device from the Scanlist then download the revised scanlist to the 1747-SDN
(Do I need to stop the scanner from running before downloading?)

3. Remove the old device from the network.

4.Configure the Comm-22-D adapters baud rate to that of the established network.

5. Physically connect the new device to the network with its default node 63
(or can the new adapter have the original node address as the old one is no longer in the scanlist?)

6. If required, use the EDS wizard to register the file for the new drive.

7.Use the drive commissioning tool to change the node from 63 to the same node as the decommissioned drive


6. Add the new drive to the Scanlist

7. Assuming that the I/O for the new drive will differ from the old one, then remap to the M0 & M1 buffers via the scanner properties tab
(Not sure if this is wrong and that maybe I should be mapping to the discrete I/O areas directly)

7. Download the revised Scan list to the Scanner.
(Do I need to take the scanner out of run at this point?)

8.Assuming all is healthy, then add the relevant PLC logic for the drive.

Anything incorrect or missed or folks?

Welcome to the PLCTalk forum community ! This sounds like a fun little challenge.

First, can you describe the obsolete drive and its DeviceNet interface ?

While DeviceNet has AC Drive "generic profiles" that theoretically would allow an almost seamless migration, it's more likely that the size of the I/O connection and the meaning and scaling of the bits and words in it are different between the old drive and the PowerFlex 70.

Also: the 20-COMM-D is for the PowerFlex 70/700. The 22-COMM-D is for the PowerFlex 4/40/40, so be sure you get the correct interface.

the difference between the "M0/M1/Integer data" AND typical discrete addresses such as "I:2 / O:2" etc also used for each device on the network when it comes to the scanner mapping and PLC code

Click to expand...

You probably have an increased degree of difficulty because your configuration uses "segmented mapping" to break up the I/O connection into both discrete memory and the "M-file" memory of the 1747-SDN. That was done to conserve memory in the small 32-word I/O image for data that needed to be exchanged quickly, while allowing reference/feedback/monitoring words to be relegated to the slower M0 and M1 files that are periodically moved by COP instructions to SLC data table files.

I recommend uploading the Scanlist and I/O map from the 1747-SDN before proceeding, as well as determining for certain which I/O Assemblies the existing drive is using, and how to configure the PowerFlex 70 and the 20-COMM-D to be the same or similar.

You will definitely have to take the Scanner out of RUN mode any time you want to download a change to the scanlist.

Ken Roach said:

Welcome to the PLCTalk forum community ! This sounds like a fun little challenge.

First, can you describe the obsolete drive and its DeviceNet interface ?

While DeviceNet has AC Drive "generic profiles" that theoretically would allow an almost seamless migration, it's more likely that the size of the I/O connection and the meaning and scaling of the bits and words in it are different between the old drive and the PowerFlex 70.

Also: the 20-COMM-D is for the PowerFlex 70/700. The 22-COMM-D is for the PowerFlex 4/40/40, so be sure you get the correct interface.



You probably have an increased degree of difficulty because your configuration uses "segmented mapping" to break up the I/O connection into both discrete memory and the "M-file" memory of the 1747-SDN. That was done to conserve memory in the small 32-word I/O image for data that needed to be exchanged quickly, while allowing reference/feedback/monitoring words to be relegated to the slower M0 and M1 files that are periodically moved by COP instructions to SLC data table files.

I recommend uploading the Scanlist and I/O map from the 1747-SDN before proceeding, as well as determining for certain which I/O Assemblies the existing drive is using, and how to configure the PowerFlex 70 and the 20-COMM-D to be the same or similar.

You will definitely have to take the Scanner out of RUN mode any time you want to download a change to the scanlist.

Click to expand...

Well firstly Ken, I would like to say thanks so much for the quick (and detailed) reply.
You deserve a global recognition award for the years long service in the field of Controls and automation my friend

Regards the old drives, I haven't presently got the detailed information, but will have this early next week.

But already you have provided me with some priceless points regarding the 20-COM-D adapter and Memory / Discrete mapping, which despite searching in the usual places (Rockwells Knowledge Base etc) I could not find, so thanks a bunch.

I will certainly be using the backup of the existing network as a mapping reference as I recall that once a device is removed from the scanlist, its mapping is also deleted from the scanners memory.

Do you think that the steps I plan to take are correct? Or anything missed?

Also, after doing more reading on here, noted that there can be a conflict between the config of the Powerflex 70 parameters and the DNET Comms to the drive via its adaptor.
Possibly on its safety settings??

Memory / Discrete mapping

Click to expand...

You can break up a single I/O connection into up to four "segments" and place any amount of that data in four different places in the Discrete memory and the M0/M1 memory of the 1747-SDN. It was very useful for moving bulky AC drive "DataLinks" or slower FLEX Analog module I/O into the slower but larger M0/M1 data transfer mechanism.

Be sure to limit your M0/M1 access, though: literally every time you address a bit or word from one of those files, the whole M0 or M1 gets transferred and you take a scantime hit. I frequently used COP to move M1 to the N11:x file at the start of my program, and to move N10:x to the M0 file at the end of my program. Sometimes I would add a little scan counter to do it every 10 or 20 scans.

once a device is removed from the scanlist, its mapping is also deleted from the scanners memory.

Click to expand...

Yes. Deleting a node from the graphical network view doesn't remove it from the 1747-SDN scanlist, but deleting it from the scanlist does delete the mapping.

there can be a conflict between the config of the Powerflex 70 parameters and the DNET Comms to the drive via its adaptor.

Click to expand...

I think the best way to describe that is "you need to both Enable and Configure DataLinks". There are parameters in the 20-COMM-D and in the PowerFlex 70 that have to be consistent with one another.

Maybe the most common example is the "DPI I/O Config" versus the "M-S Input" and "M-S Output" parameters. To enable DataLinks that are part of the I/O connection with the SLC-500, you have to enable them in DPI I/O Config as well as in M-S Input and M-S Output.

This is because there is a very-seldom-used peer-to-peer DataLink feature, so you have to explicitly say that the DataLinks will be part of the I/O Assemblies.

And then, of course, you have to go to the DataLinks parameter group and actually configure which parameters are assigned to A1/A2/B1/B2/C1/C2/D1/D2 for both Input and Output DataLinks.

Possibly on its safety settings ?

Click to expand...

Probably not in this case. To my knowledge, PowerFlex 70 always had hard-wired Safe Torque Off, with a tiny little (flimsy !) daughtercard. It never tried to use CIP Safety over DeviceNet.

I used to be one of the experts in RA's regional sales on using drives with DeviceNet, but it's been long enough that I don't still have all the details memorized.

But I'm happy to help on this, because of you take prudent steps to be sure you know all the details, it doesn't have to be confusing or difficult.

Ken Roach said:

You can break up a single I/O connection into up to four "segments" and place any amount of that data in four different places in the Discrete memory and the M0/M1 memory of the 1747-SDN. It was very useful for moving bulky AC drive "DataLinks" or slower FLEX Analog module I/O into the slower but larger M0/M1 data transfer mechanism.

Be sure to limit your M0/M1 access, though: literally every time you address a bit or word from one of those files, the whole M0 or M1 gets transferred and you take a scantime hit. I frequently used COP to move M1 to the N11:x file at the start of my program, and to move N10:x to the M0 file at the end of my program. Sometimes I would add a little scan counter to do it every 10 or 20 scans.




Yes. Deleting a node from the graphical network view doesn't remove it from the 1747-SDN scanlist, but deleting it from the scanlist does delete the mapping.



I think the best way to describe that is "you need to both Enable and Configure DataLinks". There are parameters in the 20-COMM-D and in the PowerFlex 70 that have to be consistent with one another.

Maybe the most common example is the "DPI I/O Config" versus the "M-S Input" and "M-S Output" parameters. To enable DataLinks that are part of the I/O connection with the SLC-500, you have to enable them in DPI I/O Config as well as in M-S Input and M-S Output.

This is because there is a very-seldom-used peer-to-peer DataLink feature, so you have to explicitly say that the DataLinks will be part of the I/O Assemblies.

And then, of course, you have to go to the DataLinks parameter group and actually configure which parameters are assigned to A1/A2/B1/B2/C1/C2/D1/D2 for both Input and Output DataLinks.



Probably not in this case. To my knowledge, PowerFlex 70 always had hard-wired Safe Torque Off, with a tiny little (flimsy !) daughtercard. It never tried to use CIP Safety over DeviceNet.

I used to be one of the experts in RA's regional sales on using drives with DeviceNet, but it's been long enough that I don't still have all the details memorized.

But I'm happy to help on this, because of you take prudent steps to be sure you know all the details, it doesn't have to be confusing or difficult.

Click to expand...

Thanks very much Ken.

I've done a bit of digging in the SLC program and found that the status and command words used for the current, obsolete drives, follow the same format as the new Powerflex 70, which would I assume only require the "standard" 4 input and 4 out bytes mapping from the drive.

However, the M1 file data (the M0 data isn't used), which is copied into 4 integer words, seems to be the fly-in-the ointment, as 6 bits within the first word have been used for each of the 6 current adapters/ drives for a "Communication fault flag".
Assume this is from each of the 6 drive adapters.
I'm guessing that to replicate this with the PF 70's, I'm going to have to use the Datalinks, although I haven't actually seen this particular Node comms fault functionality in the adaptor or drive yet.

I have little personal experience with Datalinks, but reading the Adaptor manual, it seems to show that enabling them all requires a step up from the standard 4 bytes input 4 bytes output, with the full datalink mapping requiring 20 bytes of each instead.

I also read that the datalinks need to be specified in the drive parameters, but not sure how they are handled when mapped within the 1747-SDN scanner. Is this done using the advanced mapping feature and splitting the discrete I/O then the extra 16 bytes into the M1 / M0 area?
Pretty sure tat this is what you alluded to in your last post?

Looking at the scanner advanced config (in the SLC I/O Config), there are 32 input / 32 output words (discrete) and 394 each M1 / M0 words.

Other question is regards the hierarchy.

From the 20-COMM-D Manual

If a controller is used to control the drive, set adapter Parameters 25 -
[M-S Input] and 26 - [M-S Output] for Master-Slave Hierarchy.

So pretty sure that this config is indeed Master-Slave?