ML 1400 -> PowerFlex 4M - Strange comm problem

Hi!

Yesterday a customer called me and said that he couldn't start a motor.
This motor is controlled via:

HMI->PLC-> Modbus->Powerflex4M->Motor.

1. When he pressed the start button in the HMI the motor didn't run. Since I'm on distance I started out with testing if the PF wasn't faulty. We set it to control via buttons on the PF and we could run the motor. So the PM is OK.

2. However when the modbus cable was unplugged from the PF there was no fault. I should get F0081 (Comms loss) fault on the PF. Strange.

3. We did a factory reset on the PF. Still not working.

4. We switched the Micrologix 1400 to stop and then run. Still not working.

5. We have a memory module with the program so we made a transfer from the MM to the ML1400. Still not working.

6. Turn the power to the cabinet off and the on. Still not working.

As it became late I made one final suggestion for the day. They have six identical machines, so I told them to switch the PF with a working PF on another machine.

7. They switched the PF and tried it. Now BOTH of the machines work.

Here comes the big problem. WHAT was the problem? Does anyone have any suggestions? I haven't made this program myself but as I can see the modbus writing is copied from one of Rockwells example projects.

Since I didn't get a fault when I unplugged the modbus cable between the PF and ML1400 I'm thinking that there is a problem with the modbus? But we have several of these machines been running for years without any problem so it seems unlikely. But why did it start to work again when we switched? When they switched the PF there was a longer power break on the cabinets ofcourse, could this be the reason for it to work again?

I just hate it when things is solved by themself...

Is there a way that the modbus comm can be on a temporary hold or paus or locked out? I read this by Ken Roach (I think) in a post about modbus:

"And once you get more MSGs than buffer spaces (it depends on the controller and network port), a few message failures can cause a cascading backup of the communication buffers and "lockup" of communications on the controller."

I've been in contact with my rockwell support, but unfortunally he is in hospital and will be back next week.

Here's a really old post I did a few years ago with example 1400/Powerflex 4 modbus MSG code examples. In my opinion, programming the MSG instructions is a bit difficult to get correct - to avoid the 'wedge' problem you describe. You want to ensure the MSG code has properly dealt with the DN and ER bits of the MSG instructions so that any hiccups in the communication do not 'wedge' the whole communication loop.

http://www.plctalk.net/qanda/showthread.php?t=44818

Thanks alot I will look into that as soon as I can!

I will show you what is used right now in this project, this is from a reference project from Rockwell that I think he have used. It looks exactly like it.

I have never myself programmed modbus but it seems a little bit strange that the MSG that initiate the whole cycle (MSG11.0) is reactivated by itself the whole time?

Could you explain this "wedge" problem, how and when it accure? I like to understand what my problems are so I can wrap my mind around it more easy.

And thanks alot for your time, it is really helpful

Here is sample of what we use. A few more bits but I'm very confident of it.

That 'what is used' looks suspicious to me: I can't see where it waits for the final MSG to be completed before it starts again at the first MSG (or else I'm missing something). It looks to me like the MSG instructions will start running in parallel (instead of serially) which will cause the modbus communications to eventually just lock up (wedge).

If you look at my code, or Bernie's, you'll see we both cycle through each MSG, get to the last one, then have a mechanism to start again at the first one. In my example, I use the first scan bit to kickstart it all.

Like Bernie, I'm very confident in the code I provided. It was used on many,many machines and they were rock solid for years.

Yes it looks wiers to me to and thats exactly what I thought. Strange that this is an example from Rockwell :/

I will for sure change it for my upcoming cabinets.

Once again the advice "If you wan't something to be good, you should do it yourself" is proven correct