Analog Sample Rates Different Between Sensors

Good Morning :)

I'm working on a pressure data logging system, and I'm finding different values when different sized pressure sensors are used...

I'm using an S1215F PLC with the integrated analog inputs to measure pressure and temperate during a test. I have a 2ms flip-flop acting as the sample trigger, and the data logging function to log the data and save on the webserver.

The pressures of the tests are typically 1 bar, and these tests work absolutely fine with a pressure sensor that goes up to 6 bar. However, I've been asked to install a 30 bar sensor instead, due to other tests being completed on that rig, and the data log results are much more 'stepped'.

I'm monitoring the data live as I'm doing the test, and it looks like the PLC is almost 'sticking' on some values, even through the pressure is increasing. Or that the resolution of the 30 bar sensor is much bigger, causing bigger steps between pressure changes.

However, the 30 bar sensor shouldn't be a problem at 0-10V at 0-30 bar / 27648 = 0.0157 PSI increments. But as an example, the data log is jumping from 2.14 PSI for about 7 samples, then to 2.68 PSI for about 7 samples, then to 3.21 PSI for around 7 samples. This makes the data log graph look very stepped.

I could just reduce the sampling speed, but the point is to get the very fine 0.015 PSI increments during the test.

Any ideas?

(Also, the sensors are from the same supplier IFM)

Regards,

AD <3

Have you tried to change in properties integration time or smoothing?

Maybe this will help you

Mr.V

Have you tried to change in properties integration time or smoothing?

Maybe this will help you

 Hello! I have tried this, and it is currently set to None for smoothing and I've also tried changing the noise frequency.

It just seems a bit strange that it works fine with the 6 bar sensor, but the input values seem to stutter/ stick when using the 30 bar sensor.

Also, I have the pressure trending on a HMI as the tests are being done. Again, the HMI trend looks very smooth when the test is being done, but the data log CSV data is stepped and the curve looks awful.

- AD

So i think the problem will be in sensor it have big range 0-30Bar so if scale 4-20mA, then 0.1mA will be 0.1875Bar

Before was 0-6Bar so 0.1mA will be 0.0375Bar

Or maybe try another manufacturer sensor maybe it will work better then IFM i don't know

Automation Dad

However, the 30 bar sensor shouldn't be a problem at 0-10V at 0-30 bar / 27648 = 0.0157 PSI increments. 

What made you think that a resolution of analog inputs on S7-1215 is 16 bit ? 

S7-1215 data sheet:

jacek d

 

Automation Dad

However, the 30 bar sensor shouldn't be a problem at 0-10V at 0-30 bar / 27648 = 0.0157 PSI increments. 

What made you think that a resolution of analog inputs on S7-1215 is 16 bit ? 

S7-1215 data sheet:

Analog value generation for the inputs

Integration and conversion time/resolution per channel

● Resolution with overrange (bit including sign), max. 10 bit

This is just my poor assumption that because the NORX function is ranging from 0 to 27648, and the sensor range is from 0 to 30 BAR, I just simply divided the sensor range by the NORMX range, and assumed that the PLC would read a different result within that range.

Apologies for my lack of knowledge, but when you say 16 bit and 10 bit, to you mean that the 27648 is 16-bit, but if the PLC can only read 10-bit then the space between the differently ready pressures would be much higher? i.e. not 0.0157 PSI, but something like 0.25 PSI? That would make sense based on the readings I am getting.

To counter this, would the S1500 range improve the sample rate? If so, then I might just get one of these ordered to test it. I could get away with keeping the 6 bar sensor on the rig, but would be better if I had 30 bar.

I've added an attachment which shows the two differences with the different sensors. Hope this helps.

Thanks to everyone so far! Ver helpful people as always :)

- AD

Automation Dad

Apologies for my lack of knowledge, but when you say 16 bit and 10 bit, to you mean that the 27648 is 16-bit, but if the PLC can only read 10-bit then the space between the differently ready pressures would be much higher? i.e. not 0.0157 PSI, but something like 0.25 PSI? That would make sense based on the readings I am getting.

Yes. In case of S7 analog modules, the max value is always 27648 "counts" no matter what a bit resolution of a module is. But what changes is step the reading changes between two different levels. In case of 16 bit module it is 1 "count". Less bits means more "counts" pre step.

How about extending your set up by adding e.g.6ES7231-5ND32-0XB0? It should be less expensive than S7-1500 and still meeting your requirements.

jacek d

Automation Dad

Apologies for my lack of knowledge, but when you say 16 bit and 10 bit, to you mean that the 27648 is 16-bit, but if the PLC can only read 10-bit then the space between the differently ready pressures would be much higher? i.e. not 0.0157 PSI, but something like 0.25 PSI? That would make sense based on the readings I am getting.

Yes. In case of S7 analog modules, the max value is always 27648 "counts" no matter what a bit resolution of a module is. But what changes is step the reading changes between two different levels. In case of 16 bit module it is 1 "count". Less bits means more "counts" pre step.

How about extending your set up by adding e.g.6ES7231-5ND32-0XB0? It should be less expensive than S7-1500 and still meeting your requirements.

 Thank you for this - This may be what I need to resolve the issue. As you say, it's 16-bit so the resolution should be much higher than the standard 1215 PLC CPU.

One question though... Am I able to use 0-10V on this? It states -10V/+10V, and I assume this just means the full range of the analogue input card so would still allow for 0-10V by setting the NORMX function from 0 to 27648, rather than -27648 to +27648?

- AD

Automation Dad

Am I able to use 0-10V on this? 

 Yes.

Automation Dad

It states -10V/+10V, and I assume this just means the full range of the analogue input card so would still allow for 0-10V by setting the NORMX function from 0 to 27648, rather than -27648 to +27648?

Set NORMX to -27648 to +27648 and it will be fine. In case you will get some small negative readings, you can limit the reading to clip negative values to zero.