1794-AENT Module Configuration

[rammin48 1]

I have 8 drops of 1794-AENT Modules with a 32 input, 16 output, OE4, and an IR8 card. I have experimented with the RPI's and I am baffled. With the AENT set for Rack optimization and an RPI of 300ms I am seeing multicasts at 20ms? So I looked at the RPI of the other modules and one was set to 20ms. I have now set the other modules to greater thatn 300ms and that cleared alot of bandwidth up. My question is how do you set the RPI's for the most efficient opperation? Rack optimization to me, means that the AENT controls the packet transaction, not the other way around. Unless the RPI for the AENT is for the AENT module only. I have alot of traffic and the opperation is inconsistent, with traffic peaking and subsiding to the point of affecting the network update times. In Control net I know that the RPI's have to be set specifically in relation to the NUT times. So how is the Ethernet supposed to be set? Thanks for any input.

[Ken Roach 194]

You're getting the hang of it. The RPI for the 1794-AENT's "Rack Optimized" connection will transfer discrete data for 16 points per module, for up to 8 modules, in a single connection. That's the one you had running at 300 milliseconds. Your assembly of FLEX modules includes several that have their own individual connections, which run at their individually configured rates. The 32-point discrete modules, and any analog modules will have individual connecctions at their own RPI rates. The inconsistency you report is confusing, though. Usually the bottleneck for EtherNet/IP networks is the controller interface and how many packets per second of I/O traffic it is handling. Unless you have some software applications also on the network that perform group reads and writes of large numbers of tags, the amount of data on Ethernet should be pretty stable when averaged over a period of at least the largest RPI. I have never closely examined EtherNet/IP traffic to see if the "scanner" modules attempt to spread out their traffic for multiple similar RPI's, the way that ControlNet schedules do. If you set up four connections for 100 millisecond RPI's, for example, ControlNet would schedule 1 of them every 25 milliseconds so that all four have arrived before 100 ms has elapsed. I don't know if EtherNet/IP does the same thing, or if it just waits until 100 milliseconds have elapsed and sends all four packets.

[rammin48 1]

I finally got through to tech support and this is what I found. The AENT determines the update rate if the module is able to have Rack optimization enabled. And due to random luck, only one of the four modules can do that. So I have essential 4 devices on each drop, times 8 trying to update at the individually set RPI. Add a couple of Touchscreens and then what. So the solution is to find a set of RPI's that are not multiples of each other( prime numbers) and distribute them through the devices to prevent peak demands occuring. you also have to allow enough space between RPI's for the specific transaction to occur. Another issue/limitation is I can not set the Flex I/O with an RPI of greater than 750ms. So I have to coordinate RPI's in 3/4 second or less. Impossible given the traffic. Because of this issue I am now screwed as I have exceeded the 32 device limitation of the Compact logix. Guess I have to suck it up and tell the boss we have go to the Control logix which allows 128 devices and upgrade three in the field, and three in the shop, plus miss ship date. I won't even go into the flex I/O not being rack optimized.

[Ken Roach 194]

Could you replace the 32-point Input module with two 16-point Input modules ? That would eliminate one I/O connection per FLEX adapter.

[lotuseater 0]

After I found this out about my 32-point modules a while back, I was able to exchange them for 1794-IB16XOB16Ps. This is a 32-point module, 16-in and 16-out, that you CAN rack optimize.

[Ken Roach 194]

If you need a more powerful CompactLogix, the 1768-L43 controller can handle 64 CIP connections through the 1768-ENBT module. If you already have 1769-series I/O, this might be your best solution.

[rpraveenkum 5]

don't have a battery Edited 7 Mar 2007 by rpraveenkum

[rammin48 1]

Well there has been several issues with this. The main one was my vendor was told that there is no 24vdc power supply. We have now found that there is one available. The next issue is the L43 is an 8 day lead time. I am supposed to ship this panel on Monday. so I am looking at Control logix for this lineup, then retrofitting the rest with L43 and ENBT cards. It is amazing that Allen Bradley does not stock parts.