What is the best equipment to obtain the correct I and V of an RC Circ 2

[pybb3ezv]

This is an ongoing question I have for which I still have not found a satisfactory answer. This concerns measurements with oscilloscopes in general and to understand it I am considering one of the simplest model circuits -- the RC circuit. Suppose you have a 10Ohm resistor and a 100pF capacitor in this RC circuit and you apply 800kHz of 8ppV amplitude. I have asked a question regarding this circuit earlier in this forum and I remained with the impression that studying such model circuit even with the state of the art equipment is so unreliable that the I and V data obtained hardly reflect the real I and V that would be expected theoretically. Is that the case? Is it really true that there is no equipment at present that would provide true I and V? If it isn't then what exact equipment (oscilloscope and probes) would you recommend that would yield true current and voltage of the said sine curves measured across the capacitor and the resistor, with all parasitic effects (capacitances and inductances) at a negigible level?

 

[VEBill]

The usual garden-variety passive 'scope probe will impose capacitance on the order of XX pF (should be in the product spec.). This will obviously have an impact on circuits using 100pF capacitors; this almost goes without saying. One must *always* contemplate the impact of the measuring instrument on the circuit. Often it just takes a second or two to estimate that the impact should be negligible in most cases. But high speed digital, RF, and pF are warning signs to be careful.

In the real world, one would not often have cause to require the sort of data in your question. Prototyping an RC circuit would typically rely on modeling. Trouble-shooting would be looking for gross failures, perhaps by comparison to parallel-channel circuits. Is this for school?

 

[pybb3ezv]

No, not for school. This is a fundamental problem I'm encountering and I'd like to have it resolved first on the model circuit mentioned in order to proceed further. The voltage probes I have access to are Tektronix P2221 (the current probe is Hall effect TCP0030 which is the best on the market). It's true they have 100pF capacitance but, curiously, turned out to give better results compared to the Tektronix active probes I tried and returned because they were influencing the phase shift badly while P2221 keep it just fine. In addition, for those parameters the 10X turned out to be less appropriate because of its lower resolution. What is of great concern is the low input impedance of the probes, even the most expensive ones. Wish there could be something like a Keithley electrometer type of a probe. But there isn't, apparently.

As for modelling, I've dropped it altogether after I saw how sloppy PSpice (which is at the basis of all else having to do with modeling) is. So, at this point it seems that there is no reliable apparatus that can measure correctly even the simplest of circuits. Hard to believe.

 

[IRstuff]

Since you know there is a problem, why insist on beating on a dead horse? Use a bigger load and move on. Where's the theoretical break that prevents you from scaling the problem?

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize

 

[Skogsgurra]

Quote: "it seems that there is no reliable apparatus that can measure correctly even the simplest of circuits. Hard to believe"

That's why EE:s are there to fix your problem. There's always a way. And it is usually easy to follow. But you seem to concentrate more on how difficult things are. Go find a qualified EE. He will sort this out for you.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[pybb3ezv]

I need that particular phase shift or of similar order. Unfortunately, the problem persists for all measurements on the lower end of caps and that's unexpected for me. How come that simple of a circuit cannot be measured properly?

 

[IRstuff]

I guess a more germane question is WHY?

It's not like RCs are complicated or mysterious, so why is there a need to obtain "correct I and V of an RC Circ?"

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize

 

[pybb3ezv]

@Skoggsgurra,

Thing is, is it my problem or that's an inherent problem in measuring circuits. If it's the latter EE can't help, won't you agree?

 

[pybb3ezv]

@IRstuff,

Just curious. It appears that while theoretically a, say, RC circuit in its ideal state is well understood, experimentally it is quite a challenge to obtain the current and voltage which theory requires. This may have unforeseen ramifications, don't you agree?

 

[ReggiePerrin]

Can you attach a resistor to your test point without adding too much capacitance? Take your measurements (test point and reference) through 10k resistors.

 

[Skogsgurra]

@pybb

Is this something personal? Why the whining?

Heisenberg told us long ago that you cannot observe without influencing the observed object. And resistive and capacitive loading effects on a circuit has been known for centuries and those effects can be handled with great success.

So, the answer is that an EE can help. Believe it or not. If you want to make a philosophical statement, just do so. But do it in some other forum. This is for practicing engineers.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[VEBill]

As above. In other words, one steps away from the 100pF RC circuit and finds another, higher up, method to answer the practical question.

 

[MacGyverS2000]

py,

Look at it another way... consider a circuit with a 1 ohm resistor and a 1F cap. That scope probe isn't going to affect the measurement one whit. Theory will match up with actual spot on. As you approach your circuit with much lower capacitance, the measured will start to roll away from theoretical... but you know why it's rolling away, the effect of the measurement tool is getting larger. Theory tells us how it will effect the measurement and to what degree, so a good scientist will account for those changes in his data (or at least know where the measurement limits are). I don't see a problem with that.

Dan - Owner

http://www.Hi-TecDesigns.com

 

[pybb3ezv]

So, I take it as established that there is no apparatus at present, no matter how much state-of-the art it is, which can test the theory in the region of nF or fractions thereof and R on the order of Ohms. Supercaps are no substitute because of many other problems and most importantly because of the already mentioned insufficient phase shift.

 

[Skogsgurra]

If that makes you happy: Yes. You cannot measure anything 100 percent correctly. There's always some influence and there are tolerances. There's also thermal noise and other things.

Happy now?

Why supercaps? They are in an entirely different capacity range. Some ten billions or more higher capacitance value. I really cannot see your problem. And why you are so persistent. Are you designing power supplies for guitar amplifiers?

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[VEBill]

"...test the theory..."

We're getting close to a hundred years past the point where testing the (R-C) theory is necessary, or even interesting.


I'll give you a couple of typical situations where one could tune capacitors to pF precision without even raising a sweat.

Let say you had several channels in a circuit that all had to be matched as identical as possible (hey, I've got this T-shirt). One can sort through the components, measuring them out-of-circuit using a $10,000 4-wire instrument with lots of digits, and bin them into exact measured value values. Then build the circuit (carefully), and all three channels WILL BE exactly the same (unless you've done something wrong).

Another example. Presumably the real-world circuit has an output. Presumably these components have some real-world measureable effect on the output in some circumstances (otherwise, why not remove them). So one can arrange for the circumstances to be true (bench test), and then measure the husky output to see if it's what you want.

 

[pybb3ezv]

@VE1BLL, yes, that approach is one way to go but it is an indirect way of proving theory experimentally. Thus, you would rely on the fact that because the signal is periodical the errors will even out.The nagging question will remain, though -- where's the direct experimental proof of theory. By the way theory is always subject to testing, a hundred years or not.

 

[VEBill]

...reliable experimental evidence to support theory as to how a 100pF...

There's no science to be discovered here; only the frustration of uncontrolled strays.

You might wish to look-up 4- and 5-wire measurement techniques. That's exactly where you're headed if you are serious.

 

[pybb3ezv]

@VE1BLL, tell me more about the 4- and 5-wire measurement techniques and how they can solve the problem at hand.

 

[MikeHalloran]

He said, "...look up...".

See "Google".

Or buy a book. I strongly recommend this one:

http://www.amazon.com/Electronic-Devices-Frustration-Mountains-Engineer/dp/0316368083

Or rent an EE to tutor you.


Mike Halloran
Pembroke Pines, FL, USA