pressure transducer

[jarvism]

A little background info I work at a underground coal mine. We have a Dust tank capable of holding 135 tons of rock dust on the surface. We use this tank to drop dust underground. I have a issue we use Omega 0-50psi 4-20ma pressure transducers to calculate the remaining amount of dust in the tank. It feeds into a 1734 RIO with a 1734 IE4C analog card with a L32e processor. The issue I am having is my reading is fluctuating a lot. was trying to find ways of smoothing it out a little

[Nova5]

you can try and average it over a minute. that should help reduce fluctuations. The time you average it over should be such that it doesn't impact a good reading.

what sort of fluctuations is it, only when certain things are on? ac line noise from nearby power can induce noise into the 4-20ma signal and cause erratic readings.

[GaryS]

The best way would be to install load cells on the tank. they can be retrofitted in most cases. Lasers, ultrasonic, radar would not give you good results

[OkiePC]

First order filter can do a lot to help out and it's easy math. Load cells are ideal, but can be costly to implement on existing equipment. Emery Winslow makes great load cell system for all sorts of machinery, and their hydrostatic systems can weigh nearly anything.

Don't judge them by the website...either under construction or I found the wrong link.

http://ew.mbcsitedev.com/scale-type/tank-hopper/

[drbitboy]

Originally Posted by jarvism

A little background info I work at a underground coal mine. We have a Dust tank capable of holding 135 tons of rock dust on the surface. We use this tank to drop dust underground. I have a issue we use Omega 0-50psi 4-20ma pressure transducers to calculate the remaining amount of dust in the tank. It feeds into a 1734 RIO with a 1734 IE4C analog card with a L32e processor. The issue I am having is my reading is fluctuating a lot. was trying to find ways of smoothing it out a little

I agree with Nova5 if load cells are a non-starter.



How is pressure used to calculate the remaining dust in the tank? Is it a fluidized bed?

[James Mcquade]

you said the magic word, dust.
so i have to ask you several things.
1. is the device going through an intrensic barrier and if so, what is it and the model number?
2. are you using the electrical installation requirements for class 2?
3. is the 4-20 ma signal shield connected at the plc side only?
4. is this signal near any high voltage motor wiring and if it is, is the motor cable shielded? is this a vfd motor?
i'm asking questions to help promote safety, and also answer your 4-20 ma issue.
james

[Steve Bailey]

Have you been using this method for a long time or is it a new installation? To me, this sounds like a misapplication of a pressure transducer. It depends the dust behaving like a liquid. But I presume the absolute accuracy of the measurement is less important than establishing a benchmark for when to dump the contents.
It might be enough to simply add a timer to delay dumping the tank until the pressure transducer signal has remained above the trigger level for the duration of the timer preset.

[James Mcquade]

after re reading your post, i agree with Steve.
you need load cells. dust will accumulate on the transducer orifice and cake up giving false readings.
james

[Mickey]

If you decide on a filter see txt below...

[drbitboy]

I am still not sure how level affects pressure, but Logix5k has the MAVE instruction, which may do what you want, depending on the level of fluctuation.

[Peter Nachtwey]

The first answer should not be a filter.
The first question I have is are you filtering process noise, transducer noise or wiring noise.What Steven mentions is process noise.

When the system isn't changing, neither should the transducer output. The process noise should be zero. I would first measure the output at the transducer with a scope and look for noise. If it is noisy there could be a problem with the power supply being noisy. I would then compare that to noise at the input to the analog card and look for differences to see if the noise was induced on the wires. The problem here is that if the plant is shut down the may be no noise induced on the wiring because none of the equipment that would induce noise is running. Also, the problem Steve mentioned wouldn't occur when the system isn't running. Given that all that checks out then I would consider filtering with low pass filter.








BTW guys, there are better filters than the low pass filter. Yes it is simple but for a little more effort one can implement a two pole Butterworth filter.
https://deltamotion.com/peter/Mathca...worth%20NG.pdf


Another question, if you are going to use a low pass filter have you ever considered putting a capacitor across the input?