How to build an auto- range measurement?

[hanh]

I use a DAQ card PCI MIO 16E4 . I want to build an autorange measurement. My input signal is from nA to mA. I am a LabVIEW user, pelase  help me.

[Sean_C.]

Hello Hanh,

Please explain what you mean by "autorange measurement." 

Also keep in mind that you must use a shunt resistor to take a current measurement.  You send the current through the resistor then measure the voltage.  Using Ohm's law you can calculate the current.  I = V/R.  Your range of measurable currents is dependent on the value of the resistor that you send the current through.  The higher the resistance, the smaller the current that you can measure.

Please give me some more details and I will be glad to help you out.

Regards,
Sean C.

[AnalogKid2DigitalMan]

hanh and Sean:

I have been thinking about this and it does not appear trivial. A nA to mA span is 6 orders of magnitude. WIth a single 1K resistor (value arbitrarily picked) this current range will generate 1uV to 1V. Not an ideal situation at the lower end for good resolution. Using a 10K is still 10uV to 10V, again 10uV is just a few bits of full scale max input. I think a better approach could be to switch various ranges of resistors (i.e. 1Mohm, 100Kohm, 10Kohm, 1Kohm) in circuit (via digital outputs driving a circuit to drive DIP reed relays) to bump the voltage up to get good resolution for whatever the current is. 'Autorange' could be applied by setting DAQ input level programmitically (do not know if this can be done with this card) at highest sensitivity, looking if overrange, and if so then switch to next lower value resistor. Repeat as needed until in a good usable voltage range.

Of course you need precision resistors whose actual value is known for optimal accuracy. And can the circuit supplying the current provide the voltage burden (voltage drop across resistor)? Will sense resistor load down circuit, etc....

If it were me, and the application constraints and accuracy demanded it, I would rather buy a higher end Fluke, Agilent, or Keithly meter with a GPIB interface.

Or, would a PCI-4070 suffice?

I am not trying to shoot down the use of the MIO 16E4, but without further details like accuracy, is it AC or DC, measurement thoughput, etc, I am pointing to some potential issues you may come across.

[hanh]

Thank  Sean. C and friends

My problem is :

1. I want to measure the light intensity from a sample by heating it to high temperature (may be 300 oC) and  I couldn't predict it intensity !

   Using a photomultiplier the light is converted to small current, ranging from nA to mA! (six decade order in magnitude)

2. Note that we can only do the experiment ONE time ! ( after the heating, the information stored it our sample will be destroyed!)

3. Some instrument as Keithley can do this easily (auto range mode)

 I hope you understand me !

If we dont have a good method, we will lose our data when we receive the message : overrange !

Thanks .

[hanh]

Hello. Thank for your kind reply

I hope you will read my reply to Sean. C ( I am sorry for not knowing how to reply to both of you one time!)

The problem is  the way to measure a signal (volt/ current) which ranges six order of magnitude in ONE TIME ( I explained this in my reply to Sean). I will try your hints.

In fact, I have a Keithley to do this but I want to find the way to do this by a DAQ card.It will be more useful for my measurement system 

[AnalogKid2DigitalMan]

I would then suggest a logarithmic transimpedance amplifier. The transimpedance portion is commonly used to convert current to voltage in photomultiplier tubes and photdiodes. The logarithmic portion 'compresses' the 6 order of magnitude down to a more usable span of voltage.

[WM_John_Weeks]

Another possible method would be to use an array of resistors of different values chosen to give a variety of different ranges. Sample each resistor on a different input channel. Without thinking about it real hard, it seems like a linear array and differential inputs would do it.