S7 300 delays on AC signals

[becs 0]

I have a small machine that stitches books together. Currently this is a fairly simply relay machine but I'm adding a Siemens 312 CPU to count missfeeds independently of the 14 feeders on the machine. There is already a missfeed sensor on each of the feeders that outputs a 120VAC signal. I am also adding a sensor to detect the chain conveyor spacing, 1 pulse for each load pocket. My worry is the speed at which this machine runs. I figure it loads one section of paper every 75mS or so. Each feeders missfeed sensor would be on ~200mS when it's triggers, but I need a repeatable pulse from the conveyor spacing signal which might only be high for 50 to 100mS each cycle. The AC input card have a delay of 25mS, is there potential I could drop a signal that's only high for 50mS? In the past I have almost exclusivly worked with DC machines, so I don't have very much experiance with AC inputs to PLCs. Thanks, Rebecca

[ianbuckley 0]

From a "seeing the input change" standpoint, it only matters that the signal stays in each state long enough for the I/O scanner to see the change and record it. That depends on the update rate of the I/O table, which I believe is tied to the main scan overhead for this PLC. So if the I/O is updated every 5ms, any state change less than 5ms long will be missed. As for the application, it will depend somewhat on the response time of the sensor itself. There are ~8.3 ms peak to peak of a 60Hz signal, so that is typically the worst case delay for an AC signal to be recognized. If the card has a 25ms filter delay, you have ~33.3ms + [sensor repsonse time] before data that a target in front of the sensor is ready to be registered by the I/O scanner. You don't mention whether the input card has a 25ms filter for the on to off transition as well. If so, then from a "seeing the input change" standpoint, it can be ignored (everything is just shifted 25ms). You are back to the response time of the sensor and the ~8.3ms as long as the I/O table is scanned often enough. From a timing and coordination standpoint, it depends on whether ~8.3ms (or ~16.6ms if the output to the next device is AC) is a small variable or a big variable to the application (how fast things are moving). If it is a big variable, use DC devices instead. HTH, Ian

[becs 0]

Timing I'm not too worried about. I am limited to where I can put the conveyor sensor so I already assume that the conveyor and the missfeed signals will not happen at the same time. I set a bit on the rising edge of the conveyor sensor and then reset either when I see the falling edge of the missfeed sensor or if I have two or more pulses from the conveyor sensor without a missfeed. There are only 4 AC input cards in the catalog for the S7 300 systems. They only mention delay, no filtering. I was looking at the only 16 input card they have, 6ES7 321-1FH00-0AA0. The 8 input cards they have have similar stats. The 32 input card has a delay of 15mS instead of 25mS, but that's way more inputs than I need.

[ianbuckley 0]

The input card you listed show a maximum delay of 25ms to switch states, so essentially 0-25ms (though 0 is very unlikely). So for worst case, the signal takes 25ms to turn on and 0ms to turn off. Next you have to determine how long the sensor will see or not see its target before its output changes state (on delay - off delay), and then the scan rate of the I/O table (main scan overhead). If the input can turn on and then off again between scans, you run the risk of missing a signal. It might help to draw a timing chart of the target in front of the sensor, the output of the sensor, the update of the I/O table by the input card and the main program scan.