Possible causes for loss of pressure in pressure measurement

[Celnurien]

I am not sure.  We have had the area tested and the is definitely a lot of harmonics going on.  The frequency should be 50Hz.  I have tried using a bandpass filter with little success, which made me question whether it is harmonics causing the noise.  I switched to a moving average filter which cleaned up the signal more.  You may have put me onto something with the sampling rate, I might have to reconsider the placement of the filter.

I don't know anything about the versadec system either, it is a total black box so I can't access what kind of signal conditioning they are doing.  We did have the same pressure transducers (they output a 4-24 mV signal) but we decided to switch because I was getting no signal into the cRIO (signal was being transformed through a 250 ohm resistor).  The new transducer is the same as the old one, just with the 0-10A output.

The low pressure signal is something we unfortunately have to deal with in our work.  Unfortunately these transducers are the ones that best fit the range of pressures we will experience.  Some samples we have will generate pressure close to 200 kPa.  At best I can use the lower pressure measurements to see that something is being measured that is supposed to be there (in this case, the production of steam).

These graphs are drawn from the run, with one probe running to the cRIO and one to the versadec.

[Celnurien]

Thanks for your input!

Yes, the smaller measurements are not going to be incredibly accurate, but we are more interested in the pressure peak, we can sometimes approach 200 kPa.  I take your point on my comment about the lower pressures seeming to be "accurate", I didn't consider this.

[wiebe@CARYA]

---

@Celnurien wrote:

I am not sure.  We have had the area tested and the is definitely a lot of harmonics going on.  The frequency should be 50Hz.  I have tried using a bandpass filter with little success, which made me question whether it is harmonics causing the noise.  I switched to a moving average filter which cleaned up the signal more. 

---

Over here in Europe, 50 Hz would be a terrible choice. Our main power is 50 Hz, and lots of hardware (NI HW included) even have optional 50 Hz filters (to filter it out).

IIRC, USA is 60 Hz, still pretty close to 50 Hz.

If we get interference from power supplies, 90% of the time it is 50 Hz.

If you have a choice, I'd stay away from 50-60 Hz, and maybe even it's harmonics.

[AaronTeitlebaum]

---

@Celnurien wrote:

Sorry, definitely a typo.  Should be Amps.

---

It says 0-10 amps, presumably it should be volts? And your previous sensor probably didn't output 4-24 mV, the standard is 4-20 mA.

[Celnurien]

Sorry everyone, I need to get more sleep.

That should be 4-24mA and 0-10V.  

[wiebe@CARYA]

---

@Celnurien wrote:

That should be 4-24mA and 0-10V.  

---

That sounds much better.

Usually it's 4-20 mA though. The full range is usually a bit more, but the signal (0-100%) is usually 4-20 mA.

I don't think it's relevant for the problem, but you might want to check. If your scaling is 20% off, scale between 4-20, not 4-24.