Welcome to the Solutions Expert Community EGX300

[JamesJeanette]

Welcome to the Solutions Expert Community for the EGX300!  This site is currently under development, but expect to find both technical and commercial information here in the future, along with (hopefully) an active forum for everyone involved in the Gateways and Energy Servers business within Schneider Electric.

[sesa357348]

Hi ,

there was a debate between me and one of my colleagues regarding the following:

I know that at RS-485 network communication is considered as a master / slave network, where the EGX is the master and IEDs connected to it are slaves, the question is:

Is the EGX -  by practicing  - only can communicate with only one TCPIP device even though they are classified as a client server network?

- I need to assure if we can make a number of sepams which are connected to an EGX400 can communicate simultaneously with both scada system PSE or Monitoring system PME and with  PLC at the same time.

-  Is it available to use a card module )ACE969-2) interface ports S-lan and E-lan for separate networks without negatively affecting losing any data?

- How I can make a sepam 20 , 40 and 80 to speak to two devices such as a PLC pulling data from it to do a sharing or shedding application, and concurrently speaks to a PSE or PME system?

Could you send me a document supported with tested architecture for such a system?

Could you guide me to a document that illustrate how many device can be simultaneously communicated by an IED, sepam, tesys, micrologic, Modicon PLC m340, premium, quantum, m580, twido, STB, OTB and EGX 100, 300, and 400?

Thanks a lot.

[JamesAZZO]

Are you asking about multiple TCP Clients or multiple Masters in the 485 network. Because the EGXs allow for plenty of concurrent Modbus TCP connections by multiple devices. RS-485 is not the same. The Masters will conflict. 

That said, Conrad  van Rooyen  is working on a piece of hardware that can help with this issue.

[Hal_Etheridge]

The EGX300 (and many other similar modbus gateways) will accept multiple incoming ModTCP connections from master software.  What the gateway will do is interleave the various master requests with each other.  Other than a delay in the request/response time, the individual masters will typically not have any other obvious problems with this configuration.  Note that the delay will increase as more masters are added and/or more requests are generated per unit time.  You may need to increase various timeout parameters in the master software to handle this delay - typically the master software will default to values that are reasonable for a *single* master (ie itself).  You would be responsible at system commissioning time to calculate/measure how much the parameters need to be changed (they probably will not increase linearly so some experience with this configuration is recommended before doing this for a customer).

An additional consideration is that certain device types do not always work with multiple masters.  For instance, the Sepam devices have a specific set of Modbus registers for log download so if multiple masters both try to download logs at the same time then the masters will conflict and cause download failures.  The Sepam event logs are actually *deleted* as part of the log download process so the 2nd master will never see the event records (or each of the masters will get some of the event records).  Luckily the Sepam devices have multiple register blocks available so you just need to configure the each of the master softwares to use a non-conflicting register block (or only download the records using a single master software).

[sesa357348]

James DiLiberto

EGX is a protocol converter divided into two parts:

the first is that part which speaks to the downstream IED devices in RS-485 (modbus serial master slave protocol so only one IED device can speak to the EGX which considered as the master at this time .
I'm asking about the other part of the EGX that speaks with the upper stream in Modbus TCP/IP server client protocol,

what means that several clients can speaks/command simultaneously the EGX ( which act as a server at this time),

the question is:

is that is just a theoretical issue but in practice there will be a conflict among several clients speaks to the EGX at the same time?

[sesa357348]

Thanks Hal.

what I get from you is as follows:

You've spoken about two issues or about the two parts of the EGX - as a Modbus gateway example;

First you've spoken about the upper stream network of the EGX which speaks in Modbus TCPIP server client protocol/language.

So several clients or masters are able to speaks to the EGX at the same time, where the EGX will handle their requests by responding to them alternatively- for example he'll respond to the fist master request then convert to the second master request then back to the first master second request and so on ...- and that's would only affect on the response time of the EGX so it's advised to adjust the response time out of the clients/master. Am I get your point right?

and I'm just afraid of the EGX shuffling the interleaving among masters so he would respond to client/master no.1 for example for twice or more times, what may lead to ignoring client/master no.2 not only for a specific time but it may go on like this, and I don't have the ability to test that now. is it a possible reason for some who say/claim that EGX only speaks to a one master/client software by practical trial?

second, there is a consideration that is not related to the EGX itself but to its related downstream serial speaking devices.

1- for instance, when an EGX request a specific measurement from sepam, and that speam responds, so this measurement is actually deleted from it.

may be that meets somehow what've been mentioned in that link: Two Modbus network for one Sepam with ACE969

I understand that in DNP3 and IEC103, but Modbus would be an exception, as per mention their ( some of the data zone can be available from a side but not available for the other).

moreover it's mentioned that (simultaneous access to data in indirect mode is forbidden), even though I didn't understand what've been meant by indirect mode, but the result is one at the end where the simultaneous access  is forbidden.

2- a nice work around could be done depending on configuring multiple registers in the sepam and the requester software or device. but that's depending on the availability of sepam registers with regards to the amount of required data by the master of the EGX or the master software

   from 1&2, can I say that achieving network redundancy at the IEDs level is almost can't be done !

Is this right?

[Hal_Etheridge]

TCP side of EGX:  the interleaving should occur with the master requests so you are correct.  Increasing the timeouts on the master software may be required depending on how many masters are trying to communicate at the same time.  Exactly how the interleaving is done has not been documented as far as I know so it could be the EGX processes requests in the time order they arrive or it just alternates between the masters - this is unknown.

Sepam Event handling:  the referenced document about the data zones is exactly what I am referring to.  If two masters both read from the same data zone then they will a) possibly cause comm errors for the other master when downloading log records and b) delete events from being downloaded from the other master.  Basically all the masters must use unique data zones on a given comm port to avoid issues.  If you want to have redundancy then you will need multiple comm ports and/or use different data zones.  How this needs to be configured depends on the number of masters and the number of comm ports available on the device.

[sesa357348]

EGX100 & EGX300 simultaneous connection.pdf

please take a look for the attached link and document.

thanks for all of you.