Multifunction DAQ

cancel
Showing results for 
Search instead for 
Did you mean: 

Pressure Sensor Ground Loop Problem

Solved!
Go to solution

Hello,

 

I have an interesting problem I hope someone can help me out with.

 

I am measuring pressure with a pressure sensor.  We are seeing a lot of noise with the measurement, and I believe we have a ground loop.  I verified the noise is 60Hz.  The interesting thing is the case of the pressure sensor is connected to ground.  When we thread the metal pressure sensor into the stand we see the noise.  When we disconnect the pressure sensor from the stand noise goes away.  The stand is connected to ground.  I am powering the sensor with a 24V power supply that is also grounded.  This I think is the ground loop.

 

My problem is the 24V power supply should be grounded for safety, and the I can't isolate the pressure sensor from ground.

 

I have attached how I have the system wired.  How do I hook the pressure sensor up to my NI 9215 Analog Input module to break this ground loop?

 

Thanks

 

Dan

Dan Shangraw, P.E.


   

0 Kudos
Message 1 of 16
(5,708 Views)

Dan,

 

Just some thoughts based on past trials and tribulations.  24V supply has an earth ground, but the actual +24V and com terminals should still be floating/isolated from earth ground.  Is 24V supply dedicated to pressure sensor or does it supply power to other things in the enclosure?  If its not dedicated, you might try temporarily using a dedicated supply for the pressure sensor and see what the effect is.

 

Make sure that all earth ground connections in your enclosure are all connected together at one place in the enclosure.  The same should really be the same of your test stand.  You might try connecting earth ground of enlclosure and teststand to same physical grounding point.

0 Kudos
Message 2 of 16
(5,697 Views)

The 24V power supply supplies other things in the enclosure.

 

I tried to tie the ground of the sensor the the enclosure ground.  Didn't work.

 

I will try to float the power supply and see what happens.

 

Dan

Dan Shangraw, P.E.


   

0 Kudos
Message 3 of 16
(5,687 Views)

This is quite an old thread, but I'll see if someone is listening.

 

I am currently having the same problem. Did floating the power supply help? Is there anything else to try?

0 Kudos
Message 4 of 16
(5,301 Views)
Solution
Accepted by topic author ASTDan

Ahhhh Yes,

 

I did solve this problem.  It was a ground loop.  I had the ground tied in 2 places.

 

We removed one of the references to ground and poof everything worked well.

 

One of the problems I have found is some external power supplies reference ground and some don't.  If you provide an additional reference to ground... Ground Loop!   

 

Ground loops can be nasty awful things.  In my experience they are the number one causes of noise and electrical goofiness.

 

I have compiled some resources on my website to help folks

 

http://www.autosofttech.net/resources

 

The best one is Field Wiring and Noise considerations for Analog Signals 

 

If you want the definative word on the subject check out this book

 

Noise Reduction Techniques in Electronic Systems

 

Hope that helps!

Dan Shangraw, P.E.


   

0 Kudos
Message 5 of 16
(5,280 Views)

Rule of thumb (ROT)

 

All grounds must be referenced to the same point.

 

If not you will have a ground loop.  See the above links for an exhaustive explanation.

Dan Shangraw, P.E.


   

0 Kudos
Message 6 of 16
(5,278 Views)

Thank you for replying.

 

We will try to see where we might have two ground connections. The tech tried lots of things. The problem only went away when he disconnected the cable to the data aq card or the hydraulic hoses. All of the grounds in the outlets (motor, computer, etc) had continuity to each other. Should the neg on the power supply output for the instruments not be connected to the same ground as the others? Where could another ground pop up?

 

Thanks!

0 Kudos
Message 7 of 16
(5,273 Views)

Hello Chris,

 

To answer your question depends...  Diagnosing ground loops can be very tricky and time consuming.  Sadly the method you are using, connecting and disconnecting stuff, is a good way to try and diagnose what is going on.  Hang in there....

 

Is the problem you are having noise?

 

A motor throws out a lot of magnetic noise.  If you are not using twisted pair cable you could be seeing noise from the motor.

 

Here is my checklist for noise.

 

1. Check for ground loops.  Grounds must be referenced in ONE place.  If you motor is connected to earth ground and the DAQ card is connected to earth ground they need to be tied in at the same point.  If not you have a ground loop. 

2. Check Grounds.  Sometimes something isn't grounded that needs to be.

3. Use Shielded twisted pair cables.

4. Connect Shield at one end

5. ALWAYS use differential measurements

 

Good Luck!

Dan Shangraw, P.E.


   

0 Kudos
Message 8 of 16
(5,270 Views)

It is a noise problem, and it sounds the same as yours. When I hold the transducer in the air, it reads zero. When I touch the shell to the table, the reading offsets about 0.2 bar and gets much noisier. The problem appears to come and go, and some times are much worse than others.

 

We found continuity between the grounds for the computer and the table which is connected to motor ground via the hoses. Is that enough to say that they are tied in at the same point or should they be plugged in as close as possible? There shouldn't be much difference because the outlets are on either side of a wall, but I have no idea how long the connection is between the motor ground and computer ground is.

 

We do have twisted wires, but no shielding. We could try this.

 

Thanks for all the help.

0 Kudos
Message 9 of 16
(5,265 Views)

One way to nip ground loops in the bud is to use an isolation module, here are some examples. Just be aware of the bandwidth depending on your signal frequency content needs:

 

http://www.dataforth.com/functional.view.aspx?ptypeid=2

 

-AK2DM

 

 

~~~~~~~~~~~~~~~~~~~~~~~~~~
"It’s the questions that drive us.”
~~~~~~~~~~~~~~~~~~~~~~~~~~
0 Kudos
Message 10 of 16
(5,262 Views)