Ethernet Switch Question NTRON

I have a operating 7026 managed switch and will be adding 16 drives which will require an additional switch. Can I purchased a unmanaged switch and use a patch cable from the 7026 to the unmanaged 16 port switch and communicate that way?

Or will I have to purchase another Ethernet Card for my ControlLogix rack and by a managed switch and configure a new IP?

I'd definitely recommend against the unmanaged switch. You need a new managed switch, probably one with IGMP snooping. If you are doing CIP MOTION, then you need a switch that supports IEEE 1588.

Whether you need the 2nd Ethernet card or not I'll leave up to EIP experts.

Agree with mk42.

Can you? Technically. Should you? Absolutely not.

What drives are you using? Not that it should matter I guess... But to save actual switch ports, I use a lot of AB 1783-ETAP 3 port modules for drive comms. 1 port in from switch, 1 port out to drive, 1 port to next ETAP.

I would check out the N-ring, sounds like NTRONS version of a DLR. You may not need a bunch of extra switch ports if you can do a ring... that is if the drives support it... but you didn't tell us which drives you're using.

The drives I will be using are PF 755's. Basically we are going away from older Toshiba drives hard wired for HMI controls. So I will be adding these 16 drives which I do not have enough free ports on my 7026 NTRON switch. I will not have any motion, all of these run exhaust fans.

As I understand it.

an unmanaged switch polls each port constantly for data.
a managed switch waits for the data request which speeds up the throughput.

If i'm wrong, someone please correct me.

james

From the Phoenix Contact website:

Hubs transmit all data out all connected ports. Switches automatically learn the network layout and then only transmit data where it needs to go, making the network more efficient. Managed switches add an additional level of configuration and monitoring capability to a network.

Ethernet networks are created with many point-to point connections, and infrastructure devices such as hubs and switches are used to connect the devices.

There is often some confusion between Ethernet switches and hubs. Hubs essentially act like an “extension cord” on the network. All incoming messages are retransmitted out the ports to every device. Each device sees messages for all other devices, regardless of the intended destination. The unnecessary traffic can slow the network and cause widely varying response times (nondeterministic) due to message collisions.

Ethernet switches, both unmanaged and managed, avoid collisions by routing messages to only the intended devices. When a switch receives an Ethernet message (packet), it reads the address of the device for which the message is intended, and then transmits the message out of only the port to which that device is connected (even if it is connected through several other switches).

So why couldn't you use an unmanaged switch?

I have seen 32 AB525 VFD's (exhaust fans also) on multiple PhxCon Switches with no problems.

Its not a matter of managed or unmanaged really but it needs to support IGMP. Managed/unmanaged refers to the ability to log in the switch and check port status, create vlans and many other neat stuff.

What IGMP does it route the multicast traffic to the needed ports only, where a normal switch would only do that for uni-cast traffic.

If you can't fit a 2000$ switch in your budget, I've successfully used NETGEAR GS108E's which do support IGMPv3, gigabit Ethernet, and do have a limited UI to manage some functions on it. There is also a 5 port and 16 port version.

They are not DIN rail mountable(but have screw holes) and have the cheesy wall adapter but you can always wire it directly to a 12vdc industrial power supply if you do not want to add an outlet to your panel.

Its not ideal but for roughly 100$ it does the job. Make sure it's the E variant which is gray and not the standard blue one that I'm sure everyone has seen.

rdrast said:

What drives are you using? Not that it should matter I guess... But to save actual switch ports, I use a lot of AB 1783-ETAP 3 port modules for drive comms. 1 port in from switch, 1 port out to drive, 1 port to next ETAP.

Click to expand...

Do you use an ETAP per drive then? Creating a DLR network with your ETAPs and each drive is just a drop? Similar to ControlNet taps w/VFDs? Haven't considered this approach before.

OP -
To add to rdrasts's suggestion on the 1783-ETAP, you could add the 20-750-ENETR dual port Ethernet/IP module to each drive and create a local DLR network to the drives, eliminate the need for a switch all together. Might add too much to the cost though guessing each is in the $300-$400 range. rdrast's suggestion would cost about the same, but allow any combination of drives to be taken off the network w/o interrupting the network. Where-as going w/the the ENETR card at each drive you run a possibility of taking multiple drives offline, if you break the DLR ring it goes to linear, but if you take another drive offline then the drives between go offline. However you can circumvent this by putting a temporary patch cable in place of the offline drive to restore the ring/linear network. rdrast's suggestion takes more panel space due to the footprint of multiple ETAPs.

If you choose to add another switch I would go with the same model you already have to be consistent. But if you aren't using any of the management features, using an unmamaged switch will most likely work without any obvious consequences, just be sure you set everything up as unicast in Logix. You might be cornered into this position pending access to a bucket-o-money.

James Mcquade said:

As I understand it.

an unmanaged switch polls each port constantly for data.
a managed switch waits for the data request which speeds up the throughput.

If i'm wrong, someone please correct me.

james

Click to expand...

Both managed and unmanaged switches work the 2nd way, what you listed for managed switches.

The main difference is that unmanaged switches have no IP address, and no configuration or diagnostics possible. They are "dumb" devices, and just move data where they are told. An unmanaged switch can still be a quality device that moves packets quickly, but it doesn't do anything else.

A managed switch has an IP address, which means you can configure it, you can pull diagnostics from it, and much more. For EIP, one of the most important settings is IGMP snooping, which allows the (traditionally) multicast data to only go where it is needed, instead of flooding everywhere.

@Paully

I use the ETAP's, as actually, they are slightly cheaper than using the dual port Ethernet in the drives themselves, but also because with the ETAP's, I can still unhook one or more drives, without disturbing the others.

I very rarely use the DLR ring configuration, but even if I did, disconnecting two drives would leave the ones in-between orphaned. The ETAP prevents that.

I think you answered my question mk42. I would rather not configure a second switch with a additional IP if I can just go with a 16 port unmanaged (junction) for lack of a better term. We are not doing anything special with these fans so I really just want to get these drives online with control logix.

rdrast said:

@Paully

I use the ETAP's, as actually, they are slightly cheaper than using the dual port Ethernet in the drives themselves, but also because with the ETAP's, I can still unhook one or more drives, without disturbing the others.

Click to expand...

Completely understand this logic, I actually didn't think to use the ETAP in this manner (DLR marketing at it's finest) and instead when breaking the ring just put a patch cable it to jump offline drives if needed. Only good if the drives are all in the same panel.

I suppose it makes you wonder when DLR is advantageous. Given the footprint and costs of a bunch of ETAPs of your setup, a switch and star architecture seems just as appealing. Is there something else driving your architecture choice?