Hello dear colleagues,

I will describe my application and problem as best as i can:

In our machine we have a profinet network in which we have IWLAN comunication with several ET200S, we choose to use IWLAN because ET200S are moving arround in the machine, each ET200S is mountend in a sort of cart, we have 20 carts in the line in a kind of carrousel arrange (please check Fig #1 in attached PDF).

All of the carts moves and turn arround the machine as showed by the arrows in Fig #1, the speed at which they move is about 12.3 Meters per minute, the length of the machine from side to side is about 45 meters, and the width is about 10 meters.

Inside of the machine carrousel we have 3 access points Scalance (6GK5788-1GD00-0AA0), this access points are located in the positions AP1, AP2 and AP3; inside each cart cabinet we have 1 client Scalance (6GK5747-1AA30-6AA0) which we use as gateway to link the corresponding ET200S with the acces points, and finally the acces points are linked to the Profinet IO network; the network topology is shown in Fig #2 of the attached PDF.

The problem that we have is that under this conditions, we loss comunication with 1 or more clients in a ramdom way, the comunication loss is normally instantanius (it does last no more than 4 seconds), we think that time, is the time that the client needs to switch from a acces point to another.

As mentioned before the comunication loss happen in random way, there is no repetition pattern, all of the clients may loss comunication in different zones of the machine, and under different circumstances; even when the machine is not moving.

We have reviewed the hardware conditions, by checking that all connectors (in the acces points, clients and ET200S) are clean, tight and in good shape, we also have repalced some of the ethernet cables used to link the ET200S with the clients, but so far this did not make any difference.

Based on this description would you please suggest were to start so we may avoid to losse comunication with our ET200S, this fault may appear very offten, maybe 2 or 3 events per hour.

Thanks all of you for your kid help.

Regards

Dear Siemens experts,

I can fully support what TEBANCH described based on our observasion at one of our projects. The WLAN system consists of:

- 2 APs W774-1 RJ45

- 4 Clients W734-1 RJ45, 5GHz

Each client is connected with an S7-315-2PN/DP on a moving platform with max. speed of 1.5m/s. The main PLC is an S7-317-2DP with a CP-343 which takes care of hand-shaking mechanism with the S7-315s.

The hand-shaking signals are based on TCP communication or S7 communication(we tried both). The algorithm was proven to be effective through test with wired Ethernet in office.

The disasters occur on site after the WLAN is established and the software runs. The hand-shaking with 2s period never goes through, no matter if clients are in still-stand or moving. What we don't understand is the connection of the client to any of AP is always under excellent condition with following figures:

- signal sthrength 93%

- TX rate 150Mbps

- RX rate 90-150Mbps

- Ping tests showed max. 2ms response time

With 4s hand-shaking period we have one failure within several hours. However, this period is not acceptable due to harzadous system condition.

Somehow the AP discriminates high-level data packages on TCP. It is just my guess.

We expect urgent solution from your guys. Thanks in advance.

I want to add more infos to the thread:

We made many tests based on suggestions of Siemens service hotline, such as

- move a client close to the AP

- switch off one AP

- Click "Out Door" option

- Shut off all possible interference sources such as frequency inverters

None of the measurements made any visible improvement.

The data records of one of the clients are attached here.

Hello,

very interesting prj. Sry only reading your post and simply think about some feedback from clients, probably interesting will be know if disconnect was made due communication lost (any way) or simply client close communication in normal way - disconnect. In this situation it comes to me question, what kind of communication instructions do you use, TSEND_C/RECV or TCON+TDISCON+TSEND+TRCV. Only think what can be better to find some relevant status of lost comunication. Can you provide some relevant answer? With some kind of this prj I was working, but that was inteligent guided wehicles giving parts to workmans in car plant.

I am updating this thread with result of our newest experiment.

We replaced the 5GHz antenna 6GK5 793-6MN00-0AA6 with 2.4GHz antenna 6GK5 792-6MN00-0AA6, then "down-grade" the APs and the clients to operate on 802.11g.

It works!

The result indicates deep-rooted problem of 802.11n implementations of 5GHz devices, although the firmware of them can be dated to Oct. 2015.

So avoid the 5GHz and 802.11n products of Scalance W.

Sicolinx

I am updating this thread with result of our newest experiment.

We replaced the 5GHz antenna 6GK5 793-6MN00-0AA6 with 2.4GHz antenna 6GK5 792-6MN00-0AA6, then "down-grade" the APs and the clients to operate on 802.11g.

It works!

The result indicates deep-rooted problem of 802.11n implementations of 5GHz devices, although the firmware of them can be dated to Oct. 2015.

So avoid the 5GHz and 802.11n products of Scalance W.

 Hello Sicolinx, thanks a lot for sharing your knowledge and experience, at my machinne we are working at 5GHz, and we are working on 802.11g, so for shure i will change my system to work at 2.4GHz, and test it.

Also i have upgraded the firmware of the acces points as well as the clients to the latest avaiable version, after this i let the system work for a couple of days, i monitored it, but the problem is still there.

Then i have downloaded the event log files from all acces points as well as for some clients, by monitoring the online status of some of the clients i could not found any error or event report generated in the event log when the comunication is lost, but if someone would like to take a look to them, i'm attaching Several TXT files with them.

Local Siemens product support (Mexico) suggested to change the network speed from 54MBbps to an slower rate (maybe 11MBbps) which sounds reasonable, and also change the update rate of the ET200S from 16 to 32 ms (in HW configuration) which from my point of view may not help, what do you think about this suggestions?

Thanks all of you for your hel, have a grat day .

Hi Tebanch,

I really doubt that the update rate of IO modules has something to do with WLAN communication. They run on  different processors and have only certain data momory to share. I still believe that the 5GHz products are inmature and only tested in labs under ideal conditions. So once on site with complicated environment such as strong reflections from steel structures, the APs get confused somehow and buffers are jammed.

I noted there are many errors of received packages and FCS. The retrial number also increased to 100 within a minute at 5GHz operation. But at 2.4GHz operation there is hardly any error to be watched.

We have other two projects coming soon into commisioning which are equipped with 5Hz, too. Hopefully Siemens could provide a final solution asap.

Hello,

yes reflections are not needed. What can be maybe better is add other routers (placed in good visibility place between each other) as amplifiers, maybe problem can be solved. This is my first idea.

I wish to make a correction on my previous messages about 2.4GHz WLAN. The situation seemed to be improved alot, but stilled not solved completely. With 2.4GHz devices we still encounter mysterious communication lost. At one site we are able to reduce the watch-dog period to 2 seconds, but at another site it looks worse than 5GHz.

When we disable "Background scanning", the situation is improved, but still not satisfactory.

Any idea?