NOISE REDUCTION FROM BNC Connector 2120

[Stame]

How can I reduce the noise that I am getting from NI BNC-2120? The signal that I am getting from the connector is very noisy.

[johnsold]

Stame,

The BNC-2120 is just a connector block. The noise it contributes is very likely much lower than the noise from your signal source or from the DAQ device to which it is connected.

Tell us what DAQ device you are using. What is the amplitude range? What is the sampling rate? What cables are connected? How is everything grounded? What is the signal source? What is the amplitude and frequency range of the signal you want to measure?

How much noise is "very noisy?" What does the noise look like? Is it random or does it have a periodic component?

Lynn

[Stame]

Hey Lynn,
I have used labview to generate a signal. The block diagram generated the signal and is outputed to the BNC 2120- through DAQ Assistant from labview. The amplitude that I have used is 2 volt dc ramp and 0.05 V sinusoidal. I have used 1000 samples.I was trying to see the signal in an oscilloscope( I guess it is already grounded) and the signal is generated from labview installed in a laptop.I want to measure the signal of amplitude 2.005 V(2.00 V ramp + 0.005 Volt sinusoidal). The ramp looks clean but the sinusoidal is noisy( very noisy, can see the sinusoidal but bearly can read it). THe noise is aobut 500 mV.

Please see the sample VI that I am working on right now.

[billko]

---

@Stame wrote:

Hey Lynn,
I have used labview to generate a signal. The block diagram generated the signal and is outputed to the BNC 2120- through DAQ Assistant from labview. The amplitude that I have used is 2 volt dc ramp and 0.05 V sinusoidal. I have used 1000 samples.I was trying to see the signal in an oscilloscope( I guess it is already grounded) and the signal is generated from labview installed in a laptop.I want to measure the signal of amplitude 2.005 V(2.00 V ramp + 0.005 Volt sinusoidal). The ramp looks clean but the sinusoidal is noisy( very noisy, can see the sinusoidal but bearly can read it). THe noise is aobut 500 mV.

Please see the sample VI that I am working on right now.

---

Is the probe really grounded?  Is it good?  Is it the right type?

[mikeporter]

As BLM (Bill the Lowly Minion) said, you need to check the probe. An oscilloscope is a type of instrument that requires each input to be grounded correctly. This is because you can do a lot of neat things with an oscilloscope like measure differential signals (neither input is referenced to ground, but the ground on the two channels are tied together),

Mike...

[johnsold]

You still have not told us what DAQ device you are using.  The DAQ Assistant is software.  What kind DAQ hardware is connected to the BNC-2120? In particular what is the resolution of the D/A converter used for the analog output?

Is your sinusoidal signal 0.05 V or 0.005 V?  You gave both numbers.

If you have a 5 mV sinusoidal signal, you will have a rather jagged or step-like approximation to a sine wave. If your D/A converter is 12 bits, then the resolution on a 20 V range is 4.88 mV, or approximately equal to your peak value.  Even with a 16-bit D/A you will only have about 16 steps over 5 mV.  Either way the signal might look noisy on a scope.

Your program has some strange things. The inner for loop calculates the same simulated signal 10 times and only uses the last one. The outer loop does nothing except to repeat everything inisde.  If you want to repeat the analog output, just put the DAQ Assistant inside the loop with all the sign You are using floating point values for indexes. You have duplicated code. You have hidden controls with identical default values. Why not use constants?

Lynn

[billko]

---

@johnsold wrote:

You still have not told us what DAQ device you are using.  The DAQ Assistant is software.  What kind DAQ hardware is connected to the BNC-2120? In particular what is the resolution of the D/A converter used for the analog output?

Is your sinusoidal signal 0.05 V or 0.005 V?  You gave both numbers.

If you have a 5 mV sinusoidal signal, you will have a rather jagged or step-like approximation to a sine wave. If your D/A converter is 12 bits, then the resolution on a 20 V range is 4.88 mV, or approximately equal to your peak value.  Even with a 16-bit D/A you will only have about 16 steps over 5 mV.  Either way the signal might look noisy on a scope.

Your program has some strange things. The inner for loop calculates the same simulated signal 10 times and only uses the last one. The outer loop does nothing except to repeat everything inisde.  If you want to repeat the analog output, just put the DAQ Assistant inside the loop with all the sign You are using floating point values for indexes. You have duplicated code. You have hidden controls with identical default values. Why not use constants?

Lynn

---

I was kind of way confused, also.  If the noise was 500 mV, how can you even see the 0.005 or 0.05 signal in the first place???  Normally when I have such a noisy signal, I look for issues like a) is something wrong with my probe, or b) am I even hooked up to the right place?

[johnsold]

I have designed and built systems (mostly hardware) which worked fine with signals 30 dB below the noise. But I knew exactly what the signal was and what kinds of noise and interference we had to fight.  Until we get more specifics from the OP, I will not speculate further.

Lynn

[Stame]

Hey,

Sorry my bad. I am using 68 pin, DAQ card 1 Inputs/2 Outputs, 200ks/s and 16 bit multifunction I/Q. It was an typo, I am trying to make a signal with a sinusoidal voltage 0.05 V that will be converted to constant current(1mA/Volt parameter= 0.05mA) by analog Stimulus Isolator. The noise voltage received from the BNC 2120 when plugged in with the DAQ Card (without running the VI) is about 10 mV. I understant the loops in the VI is kind of weired but I can resolve it. The main problem is the noise voltage level. I have to minimize the noise level to hit my  desired sinusoidal voltage target. What will be the solution or suggestion?

I have another question too. Can we use both analog output of BNC 2120 to generate the signal at a same time. Say ramp up and hold signal in one output while sinusoidal on the next and combine them together in breadboard.

[johnsold]

Can you describe exactly how you are measuring the 10 mV of noise? Do you have any load connected? How is the scope probe grounded?  What is the noise floor of the scope?

Certainly you can use two analog outputs to generate the separate components of your signal.  If you generate the sinusoidal signal at 1-10 volts and attenuate it to 50 mV in the summing circuit, you can take advantage of the capabilities of the D/A. You might be able to use a circuit as simple as three resistors if the Stimulus Isolator input impedance is high enough.

Lynn