5/12/2020 6:05 PM | |
Posts: 4 Rating: (0) |
Dear community, I have a question about the conversion time of the analog input module AI 4XU/I/RTD/TC ST. The manual (page 44) states that the minimum basic conversion time should be 9 ms. However when I measure a input with the trace function at a cycle time of 0.001 s (1 ms) the measurement shows a time step of around 18ms between input changes, which is twice the minimum conversion time as stated in the specs. I’m trying to measure the switching time of a pneumatic valve, from measurements taken with a high speed camera I know this valve opens in 12ms. There for the 9ms conversion time is essential for my application. In the device configuration I have only activated Channel 0, diagnostics are all uncheck, interference frequency suppression is set to 400 Hz, Smoothing is set to None and there are no hardware interrupts configured. Is there a way to reduce the now achieved conversion time of 18ms to the conversion time of 9ms stated in the specs? |
5/12/2020 10:53 PM | |
Joined: 7/7/2010 Last visit: 7/15/2024 Posts: 14952 Rating: (2393)
|
Have you right-clicked on the PLC from project view and selected "download hardware configuration" to the PLC to ensure your offline and online hardware configurations are identical? If you want high speed analog input acquisition, you need the HS or HF module, not a basic or standard AI module. If you really need fast and deterministic I/O, you need to configure the IO for ISO update. Now on to the speed of the valve and measuring it reliably. Because this is an ST module, the 4 analog inputs are not electrically isolated. That means they are multiplexed. When you hear the word multiplex and analog input, think immediately - configure all modules for voltage (the default), and jumper the + / - inputs of each unused input. Is this module in a remote IO system, or local IO? Local IO cannot update faster than 1ms (as of V16 and PLC firmware 2.8). Remote IO can go as fast as the profinet configuration and modules allow. I would approach this differently, but that would also depend on current PLC scan time. If it is < 10ms and relatively stable with no OB's queuing up according to the PLC diag buffer, I would create a cyclic interrupt OB that operates as fast as the IO module can report new data. You can find this out using a signal generator and the trace feature. Set the cyclic interrupt to 0.5ms. Have it copy the analog input from the periphery to some DB tag with the raw input value. That means IW128:P if the input is word 128. The :P is essential because it grabs the latest available value from the module. Configure the analog input to accept data from the signal generator. Send a continuous ramping function between 0-10-0... VDC with the ramp value changing fast enough that the PLC trace see's flat lines between adjacent values - instead of one pixel per vertical value. It will take some fine tuning, but what you will find is the absolute fastest your hardware can respond. If it is not fast enough, install an HS or HF analog input, reconfigure, and try again with the same experiment. FYI, the highest speed analog input can update faster than 250µs. But to capture anything faster than 1ms, you must use remote IO. I went through this exact exercise a few months ago.
|
science guy |
|
Follow us on