signal generation

[jackie_brown]

Good evening,

I 'm trying to generate an output voltage and I'm facing some problems. I'm using a NI 9263 analog voltage input module, I have the two wires (AO0 and COM) connected to the positive and negative terminations of a BNC connector; I want to generate an alternating voltage, 60Hz and amplitude from 1 to 5 V. Unfortunately what I measure with a multimeter is something always scaled, for example: if I generate a signal of amplitude 1 I measure 0.7 volt, if the signal is 2 in amplitude I measure 1.4 and so on. I would like to know if some calibration is needed or there are possible errors in what I 'm doing.

Thank you.

[Dennis_Knutson]

A BNC connector has an impedance of 50 ohms and that will cause the error you are seeing.

[Buechsenschuetz · Accepted solution]

Sorry, but do not at all agree with this answer. 

A BNC connector does not have any impedance of its own. The output impedance of a signal generator is determined by the circuit, not by the type of connector. 

Anyhow, even if the output impedance of the circuit is 50 Ohm, it is not the reason for your readings. The input impedance of mulitmeters is high enough not to create any voltage drop together with a rather low impedance output. The reason probably is the difference between a peak-to-peak voltage of alternate current (AC) and the RMS value. In your case, probably the software sets a peak amplitude and the multimeter measures a RMS value. The RMS value is always the peak value divided by (root of 2). 

See also 

http://86.43.94.97/moodlecp9a/mod/glossary/showentry.php?courseid=1&concept=RMS+Voltage

[Dennis_Knutson]

https://en.m.wikipedia.org/wiki/BNC_connector

[Buechsenschuetz]

It is true that BNC connectors match a 50 or 75 Ohm impedance. However, the input impedance of any multimeter is much higher and therefore any source impedance which is several orders of magnitude lower will not affect the measurement. 

Just check whether the voltage set on the generator is the Vpeak value. Multimeters always indicate RMS values. (And in case the AC voltage is sinusoidal, this is "true RMS", with other waveforms the RMS value will not be correct if you do not use a - more expensive - "true RMS" multimeter.)

[johnsold]

The impedance of a connector is like the impedance of a cable: it applies to propagating waves. If you measure the shunt resistance with an ohmmeter, you will get a very large value - 1E8 ohms or larger with good cables or connectors.  That is the load seen by any signal whose wavelength is long compared to the lenght of the cable or connector.  For a BNC connector the frequency probably needs to be greater than a few hundred megahertz before any connector impedance effects will be observed.

The comments about the values measured with a multimeter are generally correct. Most inexpensive multimeters measure the average value of the waveform and are scaled to dipslay the rms value for a sine wave.  For any other waveform the ratio of the average to the rms values will differ and introduce a corresponding error. Some, typically more expensive, multimeters measure actual rms values and those will be correct for any waveform with all of its energy within the bandwidth of the instrument.

If you want to see peak values or to measure the actaul waveform, use an oscilloscope.

Lynn

[johnsold]

To elaborate on my earlier post:

Connectors and cables are specific instances of transmission lines. The transmission line is an interesting device which "lives" at the intersection of circuit theory and electromagnetic wave theory.

<sea story> Electromagnetic wave theory and circuit theory are not separate and independent. I took an Advanced Field Theory course while I was an undergraduate studying electrical engineering.  One day the professor derived Ohm's Law and Kirchoff's Current and Voltage Laws from Maxwell's Equations and the constituent relationships. The classroom had four large blackboards across the front of the room. The professor started at the left, filled the blackboards, then went back and filled them again with the derivations. It was a thriiling experience to have the two key theoretical areas of electrical engineering tied together elegantly. (OK.  Maybe that makes me a bit strange, but that is the way I felt). <end sea story>

A losssless transmission line can be modeled as a sequence of a series inductor and a shunt capacitor. The values of each are determined per unit length. The resulting structure has a low pass behavior over frequency. At zero frequency (DC) everything gets through. At infinite frequency nothing gets through. The "characteristic impedance" of the transmission line is defined as sqrt(L/C) where L and C are the per unit length values.  

Real transmission lines have losses. Those losses are repesented by the series resistance of the conductors and the shunt resistance of the dielectric. Those resistances can be measured with an ohmmeter but have little to do with the characteristic impedance of the transmission line. There can also be other losses which vary with frequency such as skin effect and dielectric losses. Practical cables are designed so that the loss resistances are negligible at the frequency and power levels intended for use with those cables. 

Search Wikipedia for "characteristic impedance" for more details.

Lynn

[Buechsenschuetz]

This is all true but negligible with frequencies around 60Hz. This is almost DC from a RF designer's point of view and the main things that matter are the outpout impedance of the signal source and the input impedance of the test instrument. If the first is several orders of magnitude lower than the second, it can be neglected, too. In any case, the effect the OP sees cannot be explained with transmission line theory. 

[jackie_brown]

Hello,

thank you to everyone. I think the problem was that I was actually reading the RMS value with my multimeter,

so doing the right proportion I was able to obtain the correct amplitude.

Thanks.