looking for a solution to get read of noise caused by 3phase motor/generator working as part of my rig

[labexpert]

i wrote a code to measure temp, pressure and RPM of my experiment. the RPM signal input to labview is a puls (open collector) 

as soon as i start my motor (3phase) with a motor drive all my reading start reading rabbish!

i have tried to change the power supply of every thing, covering some cables with aluminum foils, faradey cage, ... but steel no useful measurement.

i a certain speed motor will change over to a generatore and i dont know what will happen to my readings them!

any help will be appriciated!

also if any of you have a proper code reading open collector puls as RPM please post it for me!

[johnsold]

As you have found, it is not very effective to try to shield cables with aluminum foil.

First, consider that the signals you want to measure are in the milliwatt or microwatt range while the interferring device (the motor/generator) is in the kilowatt range.  So the interference is 10^6 to 10^9 or more times larger than the desired signal!  This is not a trivial problem.

Second. Very few motor drives are designed with the idea that someone will be measuring small signals adjacent to the motor or drive.

One of the first things to check is to see if you have ground loops or ground-coupled interference.  The motor circuits will have safety grounds (or earth, depending on the convention used in your part of the world).  These grounds are NOT intended to provide a low noise reference level for measurement of small signals. It is common to have hundreds of mV between various points which are all nominally grounded and a few volts may occur occasionally.  These voltages may appear as interference to your measurements.  It is possible for substantial currents to flow through these grounding conductors. If some of those conductors are in your computer or sensors, destruction can occur.

The use of isolated transducers and signal conditioning circuitry may help separate the small signals from the interference.

Lynn

[ben64]

To add to Lynn comments you may also consider using isolation transformer and line filter for you data acquisition hardware.

Ben64

[labexpert]

thanks a lot Lynn,

i have connect my experiment table (metal frame) to the earth (main power supply)

any suggestion for a separate powerful ground?!

also all of my thermocoples and pressure transducers signals connected to my DAQ USB 6343 with protected cables and as you mentiones aluminum foli does not work in such big noise

is there any kind of filteration as a code to subtract unwanted frequencies? i have used frequency filter in my code but does not work well

i have tried to just connect a cable to the DAQ counter (it has 4 of them) with out connecting to the speed reader, with a distance of about 2 meters, as soon as i start motor in any frequency the noise jump on my RPM reading, such a strange condition!!

[johnsold]

USB data acquisition devices may be among the worst for noise issues.  The USB ground is connected to the computer ground and the data acquisition ground is connected to the USB ground, all through rather small conductors.  If any motor ground currents flow through the USB ground (whether coupled conductively, capacitively, or inductively), the measurements may be corrupted. If large currents flow, the USB device, the cable, the computer, or all of them may be destroyed.

Are the thermocouples grounded (to the motor ground)? What kind of signal conditioning are you using with the thermocouples? Are the pressure transducers grounded? Do they have internal signal conditioning or are you using external signal conditioning or trying to read the millivotl outputs directly?

Counters can measure MHz signals so a transient on the encoder line lasting less than a microsecond will easily be counted although your encoder probably does not generate pulses faster than a few kilohertz or less.  A circuit which requires a signal to remain in the same logic state for a time clsoe to the period of your highest encoder frequency can eliminate many of the transients.  Search the web for digital debounce circuits.  A low pass filter followed by a comparator with hysteresis may also work.

As Ben64 pointed out an isolation transformer may help.  Make sure you are careful about safety grounds if you go that route.

Lynn

[labexpert]

Dear Lynn

i have double check the condition

the thermocouples are not grounded, they directly conected to the USB DAQ

the pressure sensors receive a 24 v DC power and send the 0-10 v signal directly to DAQ (Omega pressure sensors)

no signal conditioning in between

they are perfectly work when the motor drive is off

as soon as i turn on the motor drive to run the motor every thing receive noise

[johnsold]

This has reached the point where it becomes much less likely that someone will be able to guide you to a successful solution via a few lines of text.

I recommend that you find an electrical/electronics engineer experienced in instrumentation and measurements to look at this.  An expert may spot in a short time something critical to the process that we would not come across in weeks of questions and answers.  The fix may not be difficult but identifying what needs to be done could be a substantial challenge.

Lynn

[labexpert]

Thanks Lynn

as you mentioned it is a challenge and it is why i have asked this question in this forum!

[djs_at_eaton]

When you find the electrical / instrumentation engineer, one of the first things he / she should do is trace all the currents in the system.  Each current source should have its own wire path back to its sink.  That path has to use wire appropriately sized to the current.  Once the paths are identified, separated, and wires sized correctly, you can determine which signals might benefit from isolation.  This is also the time to figure out where to connect shielding.  Foil and Faraday cages will not help you if your problem is poorly configured current paths.

Look at how (physically) close the signal and power wires are to each other.  Shields are more effective if the wires aren't close to each other.  Remember that most shielded wire protects from capacitively coupled noise but does nothing for magnetically coupled noise.  Motors / Generators are magnetic devices and you should expect that kind of noise.

Here are some resources from NI's website.  This is only a start.  Noise isolation and signal integrity is not an exact science.

http://www.ni.com/white-paper/3546/en

http://www.ni.com/white-paper/3035/en

http://digital.ni.com/public.nsf/allkb/4A441BC4E49541F4862573A000789203

http://www.ni.com/white-paper/5362/en