1X or 10X Passive Voltage Probe?

[pybb3ezv]

Suppose you have applied 800kHz, 30ppV sine signal (from a pulse generator) to a 115pF, 10Ohm RC circuit and you want to know the value of the applied voltage as accurate as possible what passive voltage probe would you use, 1X or 10X? Is there a way to check the accuracy of the two measurements? What I do is I use a resistor of known resistance and I divide the current passing through it into the voltage across it measured alternatively with the 1X and the 10X probe. What I see is that throughout one period of the sine wave the 1X probe sustains an R value closer to the known R value than the 10X probe. This makes me conclude, although I'm measuring the voltage across the R and not the voltage across the RC, that the 1X probe is more accurate. Am I right and are there other ways I can objectively check the accuracy of the two probes with regard to measuring the voltage across the RC?

 

[IRstuff]

What accuracy do you need?

TTFN

FAQ731-376

 

[IRstuff]

In general, this is a very complicated question, impedances and reactances abound, as well as numerous parasitic effects:

http://cp.literature.agilent.com/litweb/pdf/5988-5021EN.pdf

TTFN

FAQ731-376

 

[pybb3ezv]

@IRstuff,

Thanks for the link. What do you think of the method I'm applying to check the accuracy (I wish it to be on the order of 1 mV) and do you have a suggestion for an experiment to study that in the concrete case?

 

[IRstuff]

Well, good luck with that. You're asking for 0.0033% accuracy when you don't even know the output impedance of the signal generator to better than 5%.

Is this for school?

TTFN

FAQ731-376

 

[pybb3ezv]

No, that's just something I'm curious about. What would you say the best accuracy be and which one of the probes would be more accurate in terms of mV (as well as in showing accurately the real phase shift)? Can this somehow be verified experimentally in this concrete case?

 

[Comcokid]

Much of this also depends upon what the probe is hooked to.

A oscilloscope?
Analog or digital?
Which model?
Which manufacturer?
Are you setting the V/Div to get maximum deflection on the screen?
If digital, are you using built in measurement functions?

I have found measurement variations from all of the above regardless of the probe used. Even well know popular scope models will give you different rms voltage measurement readings for just a 115VAC 60 Hz depending upon how many cycles are shown on the screen (time base or sweep-rate setting).

Even a good Fluke handheld will give you bad measurement if you don't realize that it's 3db AC measurement spec may only be 10 kHz.

If you don't know the specification and limitations of your instrument, then the probe used is the least of your measurement issues.

 

[pybb3ezv]

Can't agree more. I'm using a Tektronix DPO 2014 scope which I believe has 1MOhm input impedance. What I'm doing is I'm dumping the data to an USB flash drive to observe it later in Excel. Like I said, I tried to recover the value of a resistor with a known resistance value (I measured its resistance prior to the scope experiment using a Keithley 2000 multimeter -- four point measurement).

 

[Skogsgurra]

Are you measuring across the 115 pF capacitor or across the 10 ohm resistor?

In the former case, you have around 1700 ohm impedance and in the latter case, the impedance is 10 ohms.

Loading 1700 ohms with a 1 Mohm probe or a 10 Mohm probe doesn't change much. The error will be less than 0.1 % in both cases (the capacitor's voltage is at 90 degrees and that makes the error very small).

Your biggest error source is probably how well the probes are compensated and the inherent accuracy of the scope. Worry about basic things first and you will find that the things you are asking about will be of very little importance.

Also, remember that your capacitance value probably has a 10 or 5 % tolerance. And that your test leads and wiring add tens of pF, which can't be ignored if you want to reduce errors.

Gunnar Englund

http://www.gke.org

--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...

 

[IRstuff]

There are probably too many error sources to even investigate. However, your failure to read the datasheet is probably number 1.

http://www2.tek.com/cmswpt/psdetails.lotr?ct=PS&cs=psu&ci=14403&lc=EN

2% tolerance on input resistance
Almost 20% tolerance on input capacitance
8-bit resolution on vertical, which means that its best accuracy with a 30V input is around 0.1 V

Anyway, back to your original question: neither, because an oscilloscope cannot support the accuracy you're asking for with any probe. Even a high performance meter would probably have trouble with an AC measurement near 1 MHz. Just in terms of digital measurement, you need at least 17 bits of single-ended resolution, which you're not going to find in most oscilloscopes.

TTFN

FAQ731-376