Customized Attenuation Probe

[Marcelo Fernandes]

Hello:

I'm using a NI-5112 to measure the signal of a magnetic/hall sensor - 100Vpp.

I want to use 10:1 attenuation but I can't use a common probe. I want to connect the CH0 to sensor with a special/customized cable&connector.

I've tried to make the same circuit with 9MOhms and capacitor like explained on site, but I reached bad results. The industrial probes has special characteristics for capacitance and noise rejection.

Someone knows how can I solve it ?

Thanks.

[DFGray]

You really do not want to use a 10:1 attenuation with this experiement unless you data acquisition demands it.  The lowest range on the 5112 is 50mV.  If you use a 10:1 attenuation, your peak-to-peak signal will be 10mV.  You will reduce your available resolution from 256 values (8bits) to about 50.  Add the 1.5 bits of noise of the 5112 and you are down to a resolution of about 20.  I would strongly recommend you use a 1:1 input.  Your circuit will be much simpler and you will get better resolution.  Use the 5112 input amplifiers as your amplifier.  The 5112 range can be set in 10% increments, so you can almost always optimize it.

[Marcelo Fernandes]

Hello Gray:

OK, but I'm worried about higher voltages like 100Vpp.

If I am correct, the max input range for NI-5112 is +-42V, ok ?

So, how can I protect the input for high voltages without distortion of the signal ?

Thank you very much.

[DFGray]

My apologies, I read 100mV, you wrote 100V.  Too early in the morning.  You are correct, you need some attenuation, since the maximum calibrated range of the 5112 is 50V.  However, there are a couple of options that may make this easier.

1. What is the output impedance of your sensor? The 5112 has a switchable 1MΩ and 50Ω input. If your sensor has a high impedance output, choosing the 50Ω scope input will usually halve the input signal - giving you exactly what you want. Make sure you are not overdriving the sensor before you do this.
2. Is your signal a full 100Vpp or is it a lower valued AC signal on a high DC background? If the latter, you can use either AC coupling or the offset capabilities of the 5112 to reject the high DC level and only look at the AC portion.

If neither or these options help, let us know and we can go from there.

[Marcelo Fernandes]

Hello DFGray:

        Unfortunately, neither than two options is my solution ...

        The output impedance of my sensor is lower. My signal is full AC with no DC component.

        My signal range of interest is from 0 to 50Vpp. I don't want to see levels higher than 50V AC.

        I'm thinking to put two zener diodes to reject voltages higher than Scope limits but I don't know if them can cause some distortion in my measurements ...

Thanks.

[DFGray]

Two zeniers ought to work (with current limiting resistors), but how well depends on your choice of zenier/resistor and the frequency range you are interested in.  I am not an electrical engineer, so I have contacted a coworker for help on this one.  We are also considering the attenuation option. It would help if we had the following information:

• Output impedance of your sensor
• Current sourcing abilities of your sensor
• Voltage range of your sensor (100Vpp, I assume)
• Frequencies you are interested in
• Frequencies that can be attenuated

Thanks.

[Marcelo Fernandes]

Hello DFGray:

My sensor has a coil with ~ 500Ohms

Voltage Range: 0.01 to 100Vpp

Max Frequency: 100KHz

Thank you.

[DFGray]

All you need is a simple compensated voltage divider (although at 100kHz, you may not need the capacitors). Check out the circuit diagram in the attached graphic. Use non-inductive resistors (carbon film are great). Make sure they can handle the power, although it won't be much at 1MΩ. If you want a bit better frequency compensation, add a variable capacitor to the top resistor and take a little back from the bottom one. The values are nominal based on the input values of the 5112. This divider should give you about 3X when used with a 5112. Use the 5112 amplifiers to maximize your signal in the input to avoid digitizing noise.

Good luck! Let us know how things work out.

[jemmy]

Hello,

This thread was of great help to me as I too was stuck with similar problem.

In my case it is piezo device (having a capacitive load of 150pF)

 the signal details are as follows.

Freq : 20Khz Max

Volt age 150Vp-p

On time : 20uSec

Gray, I am planning to try your circuit in my application, would like to know the significance of the capicators parellel to each resistor. Is it going to act as compesator, if so how do they help us in rise time or fall time..

Do  I need to consider anything more to use this circuit. Kinldy advice.

Again thanks a lot..

Jemmy

[DFGray]

The capacitors are used to ensure that that gain at all frequencies is the same.  This is usually only a problem starting at frequencies in the megahertz.  If you have sharp edges, you will see it.  A variable capacitor is used to match the impedance through the entire circuit.  If you modify the capacitance (with a variable capacitor) while watching your waveform, you will see sharp edges go from undershoot, to normal, to overshoot.  Tweak until good.  At the frequencies you mention (20kHz base, 50kHz edge transients), you will probably not even notice.  For your application, you can probably ignore the capacitors and use a simple resistor divider.  Remember to factor in the input impedance of the scope.