Offset voltage on PCI-6123

[jchrisj]

I'm running a PCI-6123 8 channel card with the BNC-2110 attached. I have an audio signal generator set up to generate a 1 KHz signal connected through an equalizer and power amp to a speaker. I've Tee'd the output of the generator to the AI0 input and I see a nice 1 KHz sine wave centered at 0V (the peaks are +/- about 0.4 V)

I have a LinearX M31 electret microphone connected directly to AI1. The microphone has an XLR (3-pin) connector with +9 V power on pin 3, the signal on pin 2 and pin 1 common.. I transition from XLR to BNC for connection to the BNC-2110 and am using a 9V battery connected between pins 1 and 3 on the microphone to provide power. Pins 1 and 2 connect to the braid and center conductors on the BNC.

With the battery disconnected I see a signal near 0 when I sample AI1 (16000 HZ rate). With the battery connected I see a nice sine wave between +5 and +7.5 V. I put a little BNC box (BNC male on one end and female on the other) with a 0.01 ufd (I think) capacitor connecting the center conductors. Without the battery the no signal voltage goes up to ~0.18V and  with the battery the signal is back to the +5 the +7.5 V range - again offset from 0 V by about 6 V.

Any ideas about what is going on? the only thing I can think of at the moment is that the battery is lifting the "ground"  and maybe I beed to replace the BNC box and connect two pieces of BNC cable with capacitors on both the braid and the center conductor.

Also, I've noticed that with nothing connected to the BNC-2110 the sampled input voltages on any channel are never zero and typically change as I move the cable around..

Any ideas would be greatly appreciated. I can probably CorrecText my input data but I'm using up some of my available range since the input signal plus the offset has to stay within my +/- 10 V range. If the signal was centered on zero I could run higher amplitudes.

Thanks,

Chris

[jchrisj]

Two additional pieces of information:

I made a capacitor-isolated BNC jumper, with small capacitors connecting the center and braid conductors of 2 pieces of coax cable and it made no difference that I can see.

Also, I think my last message ran afoul of the spelling checker, I meant to say that I could DC-correct my data as it comes in but that I would lose dynamic range in the process.

thanks,

Chris

[hoffiz]

Hi Chris,

Good coincidence I got you post again. If I understand orrectly you’re stripping out the XLR connection from the microphone, that has
3 cables and connecting it from to a BNC connector. I’m imagining the BNC strip out in a very homemade kind of way, is that right?

Is the nice sine wave from +5 to +7.5 the expected voltage? If not, what is the expected voltage? The 0.18 volts might be the capacitor been
charge, although I need to check from where since the battery is not connected.

How are you connecting the signals to the 2110? I would uggest looking at the BNC-2110 nstallation guide and making sure you connection is set for differential,
page 5.

Let me know if it helps.

[jchrisj]

The microphone has a male XLR connector. I've used a female XLR connector to connect to it, with the center conductor of the coax cable connecting to Pin 2 (The microphone's "signal" pin) and the braid connected to pin 1 ("common"). I'm using a 9v battery on a battery connector, with the wires for the connector snaked through the strain relief into the female XLR connector. The 9v battery has it's "+" terminal connected to pin 3 ("PWR") and the "-" to pin 1.

I tried connecting the male BNC on the other end of the cable directly to the "AI1" connector on the BNC-2110  with the FS/GS switch in both positions with no change. I tried to couple the center signal line through a small capacitor to block any DC, but that didn't work. I coupled both the center and the braid through capacitors (that **does** look pretty home-made) and saw no change.

I was wondering if the 1 KHz signal was "riding on" a much lower frequency signal but I can run the test panel numerous times and I get the same offset.

With no input connected, I did notice that I could change the voltage on the channels in the test panel just by moving the BNC2110 around on its cable while it iwas acquiring data.

The channel is set for differential on the test panel and in my program (I'm running NIDAQmx under Linux)

thanks,

Chris

[hoffiz]

Hi Chris,
Thanks for you notes. The fact that you are running in Linux should not make any difference. Quick note on your state: “riding on a much lower frequency signal”, running test panel would behave the same way as running a task in LabVIEW.

What is your expected peak to peak voltage? I’m guessing it has to be centered at zero, right? You are using this kind of an approach instead of using a XLR to BNC adapter since you need to power the MIC, right? What other connector blocks do you have beside the BNC-2110?

We have checked everything but?

1. Connect the BNC cable to one of the AI <0..7> BNC connectors on the front panel.1
2. Configure your software to measure this channel differentially.
3. Move the FS/GS switch to the applicable position, depending on whether you are measuring a floating source (FS) or ground-referenced source (GS) analog input signal. Refer to the Measuring Floating Signals and Measuring Ground-Referenced Signals sections for more information about these signal sources.

Here is a picture of the BNC internal hardware and the difference between GS and FS, and for your case you should be in FS.

 

Message Edited by Jaime F on 09-29-2008 11:14 AM

[jchrisj]

Ok, I've got the microphone cable hooked directly to AI1 on the BNC-2110. The acquisition code is set for differential with a range of +/- 10V and the BNC-2110 is set to "FS"

I'm seeing a reasonable sinewave - the approx. 4V pp is fine, it is the DC offset that is the problem.. I'm seeing about +4 - +8V instead of +/- 2V.

 I've attached the VI that I'm using to take the data. I would think that capacitively coupling the microphone's signal line to the A1 terminal would eliminate the DC offset, but it does not seem to have any effect at all.

Chris

[ChrisR.]

Is it possible that the microphone includes an AC coupling capacitor internally?  If so, maybe you just need to tether the output to ground with a resistor.  I couldn't find any specifications for the device on the LinearX website, but I'd think 100kΩ between the signal pin and ground (pins 1 and 2) would do the trick without affecting frequency response too much.

(Another) Chris 

Message Edited by Chris R. on 09-30-2008 12:06 PM

[jchrisj]

Thanks for the idea. I can certainly try that and see what happens.  I'll report back what I find out.

Chris

[jchrisj]

Connecting a 100K resistor between the BNC signal and ground conductors recenters the sine wave around zero volts. It has to be connected on the BNC-2110 side of the DC blocking capacitor, which does make some sense, and as you say I may not need the capacitor. I'll find that out next.

Chris

[sherwinlalap]

Hi Chris,

Im using the same type of microphone (LinearX M31 series), and I encounter the same problem before. If you read the manual of M31 microphone I think its design for RSE connection. Im using differebtial before but I can't get a clear signal that why I shift to RSE connection. Just today I tested the microphone with 1kHz audio gen and I got about 468mVp-p. (from 2.74 to 3.21V). 

Any idea on how to convert this value into dBspl Chris. Thanks and hope it will help.