Frequency inverters on profibus using S7-300 PLC

In some projects we use frequency inverters on profibus connected to an S7-300 plc. We send information to/from the inverters using the SFC14/SFC15 blocks from Siemens.

In one application I have a motor that has two speeds. A proxy detects a box and then the conveyor slows down. For some reason it looks like the conveyors sometimes starts to slow down to late. We checked the proxy, inverter and software and could not find anything. No I'm thinking it has something to do with the profibus/system functions (because simulair application without profibus do not show this problem).

(Project information. Profibus speed 3Mbit/sec, 20 inverters 3PD)

Is it possible to prioritice an inverter in some situations?
Anybody seen these problems before?
Is it possible to skip inverters once in a while? (So don't call SFC14/15)
Any other suggestions?

Have you thought about accessing the process image table directly using PIW, PQW statements...

How much time is this "to late" ?

Look up some of the posts on ISOCHRONE mode -

You may also find slowing the bus down to 1.5Mbit may help .

Are you calling SFC14/15 for every motor every scan ? What is your scan time ? Which CPU are you using ?

SimonGoldsworthy said:

Are you calling SFC14/15 for every motor every scan ? What is your scan time ? Which CPU are you using ?

Click to expand...

Yes all motors every scan cycle of the plc
Scan cycle time is approx 10ms
PLC is S7-317-2DP

What are the values in STEP7 Hardware Config for your Profibus configuration? Every device has the details in its GSD file covering the volume of data exchange, the speed of response, etc. From this, and knowing the master's data and network speed, the network configuration can calculate the typical token rotation time around the network. I would normally expect this to be a fraction of the CPU's scan time i.e. the DP network controller chip should be sending/receiving this data all the time.

Of course, for large volumes of consistent data (well, greater than 4 bytes anyway) you have to use SFC14/15 to send data from the program memory to the DP controller - you are not actually sending it to the slave device at that point.

I'd be surprised if something like transmission speed affected response, unless you were running at very low Baud rates indeed. I'd still be looking at the inverter settings - could it be a slightly different ramp rate has been set for the rogue one compared to others?

Regards

Ken

Ken , I have found with a dodgy network that slowing it down sometimes alleviates a problem - this is of course not a solution , just fault finding . Does the network comply with the maximum segment length for 3 Meg , worth checking , and also worth checking that you are using the correct GSD file , not as silly as it sounds . If you have access to a BT200 , also worthwhile doing a check .

The network was faultless a couple off weeks ago. We have a profibus analyser, this did not show any problems at that time. (We had lot's off problems in the past).

This profibus analyser shows that the complete network is send and received in approx. 4ms. This is faster then the plc cycle time. But I had some other experience with Siemens blocks costing up to 3 plc cycles to complete.

I forgot to mention that the proxy involved is placed on an ASI network. This ASI network is connected to the profibus network via an DP/ASI link.

I will investigate this problem in more detail the next time i'll be in Denmark.

I've experienced unexpected results with ASi I/O in close proximity to inverter fed motor cables. Greater separation and routing the motor/control cables at ninety degrees eliminated the problem.

SimonGoldsworthy said:

I've experienced unexpected results with ASi I/O in close proximity to inverter fed motor cables. Greater separation and routing the motor/control cables at ninety degrees eliminated the problem.

Click to expand...

We also had some problems with the ASI network together with the inverters. We use clixons (bimetal contact in the motor) to detect if the motor gets to hot. Via this contact inverter noise came into the ASI input block. This block then requlary gave errors. Siemens just replied: "Use a relais to seperate the motor and the ASI". Although I'm sure this will solve the problem, I still think it's a stupid solution for the ASI system.

I'll check the ASI bus the next time I get there. Thanks for the info.

You told it was faultless earlier, so check the EMC-cable-entryes for the motor cables and also remove paintings off from there.