Setting up a DH+ network

Hi guys,

I am looking to connect four existing, currently standalone, SLC 5/04's (1747-L541) on a DH+ network. The controllers are compatable for DH+, so I am guessing that all I need to purchase is the connecting cable (1770-CD) and the data highway terminating set (1770-XF)? Is this the correct conncetion for the SLC?

Thanks.

You will have to assign each SLC a unique node address, the default is station 01, can't have four station 01's on the same network. You change the node address by clicking on the channel configuration in RSlogix and then channel one system tab, remember the addresses are in octal not decimal. Since you only have four nodes, I would number then one through four.

Then purchase the amount of cabling you need, I don't remember the exact Belden number, but you can call any AB rep and just order "blue hose", that will get you the right cabling. It's best to use a daisy chain type network with a 1/2 watt 150 ohm resistor across the terminals at each end of the network. No need to purchase any special AB termination kits.

Thanks a million Ken

It's best to use a daisy chain type network with a 1/2 watt 150 ohm resistor ...

Click to expand...

Actually, electrically, it's best to build a trunkline-dropline network. Much more robust than daisy-chaining. But since most don't understand the difference and daisy-chaining is easier, that's what usually gets installed. Good luck with the daisy-chain if that's what you choose to implement.

And watch out for the value of the terminating resistor. It is baud-rate dependent ... 150 ohms for 57.6K but 82 ohms for 230K.

(btw, the cable is Belden 9463.)

SLC's only support 57.6 for DH+, so 150 ohms is correct.

Actually Ken - I've got a 5/04 running a 230.4 KBaud right now. Maybe the limit you recall is for a different series processor?

So - pablo2000, choose your speed appropriately and let that determine your end of line resistor . . .

Good luck,

Marc

Actually, electrically, it's best to build a trunkline-dropline network. Much more robust than daisy-chaining. But since most don't understand the difference and daisy-chaining is easier, that's what usually gets installed

Click to expand...

Would you care to explain why the trunk line would be much more robust?
From my experience the robustness of DH+ is in the following:
1. It is transformer coupled (no it is not RS-485) differential
transmission.
2. There is an SDLC (IBM standard) and Manchester encoding.
3. The PCCC (aplication layer) and the Data link layer (token
passing) are robust and proven technology.


Drop lines require capacitors and the length of the drop line is an issue, that makes it less robust. The reason for using the trunkline and T drops is the economy of scale. You need more cable to do the
point to point.

noise immunity

I was under the impression that Rockwell didn't recommend a trunk/drop line configuration for DH+ only a daisy chain.

many things suggest otherwise, for example pub 1770-6.2.1 and an understanding of what's going on electrically on a real DH+ in the field

more information ...

there’s a nice little picture (with notes) on page C-4 [Adobe Reader page 220 of 291] of the

SLC-500 Modular Hardware User Manual

and you can find some handy “how-to” text in that same section ...

also (with no offense to Ken intended) this next one isn’t a big thing but ...

I'd recommend that you use the node addresses #2 through #5 rather than #1 through #4 ... the reason: when a new processor comes from the factory it's already set at #1 ... a processor also defaults to node #1 if you have to "reset" it with the little "VBB-to-GND" trick ... anyway ... plugging another processor with node #1 into a DH+ network that already HAS a node #1 assigned will automatically generate a network conflict ...

yes, I'll agree ... a knowledgeable technician will be aware of this and be able to work around the problem ... but Doofus Dave working at 3:00 o'clock in the morning could have some serious difficulties getting things back up and running after a "let's-see-if-this-fixes-it" processor replacement ... especially if he’s going to try to download the appropriate program over the DH+ network connection ... worse case scenario: he downloads to the WRONG node #1 ... oops! ...

like I said, it isn't a big thing ... but why set yourself up for trouble down the road? ...

good points about the node number, ron

And look on page 221 where they're adding the capacitor at the middle node for noise immunity. Easy to draw; not-so-easy to implement in the daisy-chain configuration ... six wires and a cap on that connector? C'mon. Again, that's a reason where the trunk/drop is superior. LIke Ron said, why set yourself up for trouble?

My bad on the 57.6 only, sorry for any confusion.

Ron you'r right, node addresses other than 01 would be best.

Ken

Originally posted by mmw:
...six wires and a cap on that connector? C'mon.

Click to expand...

Be careful with this. In this implementation NONE of the shield wires go to the shield terminals on the connectors. All the shield connections are made away from the connectors. Terminating the shields to the connectors AND adding the caps would defeat the purpose. Not that it makes it any easier to implement. You still need a place to stick the shields and the caps.

Keith

with or without the shields, it's still a mess. Just wad it up in a ball of electrical tape ... that'll fix it up!