Agilent and Tektronix scope probe loading

[h1386343]

Please forgive me if this is a bit of a stupid question, but I was doing some research into a suitable first oscilloscope and came across a video on Youtube comparing Agilent scopes to Tektronix. In the video the presenter stated that the fundamental difference between the two manufacturers was the way they took measurements of the circuit. Agilent's approach is to display exactly what is measured by the probe, including any loading effects caused by the probe, whereas Tektronix display the signal that would be present, after removing any loading caused by the probe in software.

Im trying to work out why Agilent chose to display a signal that has been altered by the probe? In what circumstances would it actually be beneficial to see what effect probing the circuit has? If the Tektronix approach is better, why doesn't Agilent remove the effects the probe has on the circuit before displaying it?

[cyr]

Do you have a link?

I've never heard of this "fundamental difference". There really is no general way to compensate for the probe loading, the effect will depend on the circuit you are probing... 

Was this perhaps referring to some exotic special purpose probes?Was is also a marketing video?

[MarkL]

What video?

Agilent has an app note called "Side-by-Side Comparison of Agilent and Tektronix Probing Measurements on High-Speed Signals".  On page 9 it talks about 'Should You Measure “What was” or “What is”?'.  It sounds like this difference in philosophy is what the video refers to.

  http://www.home.agilent.com/agilent/redirector.jspx?action=ref&cname=AGILENT_EDITORIAL&ckey=406520&lc=eng&cc=US&nfr=-11143.0.00

Perhaps there's also an equivalent Tektronix app note for their view on this, but I haven't come across it.

[tautech]

Well the facts are: ALL measurements affect the DUT.
This is something we all need to be aware of.
Knowledge of the circuit and the area of interest is key as the signal levels (energy & frequency) will determine how the circuit is affected.

Absolute measurements in these cases can be misleading, however most scope work is just checking functionality.

Active scope probes have been developed precisely with these problems in mind.

[nctnico]

Basically it comes down to:
there are small lies, big lies and what is on the screen of your oscilloscope 

[tautech]

Basically it comes down to:
there are small lies, big lies and what is on the screen of your oscilloscope 

Not only is that funny, it is true. 

[cyr]

So, Tek and Agilent each design and characterize their high-end, multi-GHz differential probing systems differently and have created a little marketing war over which is the best way.

Not relevant in any way for someone looking for a first oscilloscope.

[HighVoltage]

In the German language we have a saying in the EE world:

"Wer mißt, der mißt Miss"
It really fits well in the German language, because the past tens verb of measuring, literally means "shit"

Freely  translated:
"He, who is measuring is measuring shit"

Because we always have an influence on the the part that we are testing for value.

[David Hess]

Perhaps there's also an equivalent Tektronix app note for their view on this, but I haven't come across it.

This Tektronix technical brief mentions it on page 3:

http://www.tek.com/dl/60W_18324_0_0.pdf

And here is the Agilent document for comparison:

http://cp.literature.agilent.com/litweb/pdf/5989-0553EN.pdf

Another consideration is that probes are often designed to compensate the transient response of specific oscilloscope models so two different 200 MHz oscilloscopes may give different results with different 200 MHz probes from the same manufacturer.  This becomes more important at higher bandwidths.

Also in many or most cases, oscilloscope bandwidth is specified *with the appropriate probe* at the probe tip so a 100 MHz probe plus a 100 MHz oscilloscope does not yield a 70.7 MHz system.

[h1386343]

I found the video a while ago, and now im looking for an oscilloscope that one thing has been rattling around in my head.
At first i thought that the Tektronix approach to include probe loading would be "better", but now thinking about it, having an understanding about how taking the measurements are influencing the DUT, and be able to take that into account appears to be "better" as it allows you to develop a more thorough understanding of how measurements are made, rather than hiding it from you, and blindly trusting whats on the display.

From the Tektronix brief David linked:

Suppose you have a perfect 1VDC signal that you want to measure. When you measure this signal with a probe, you expect the probe to measure 1VDC. If probe loading caused the signal level to drop to 0.95V, would you want the probe to read 1VDC or 0.95VDC? The Tektronix philosophy is that you want to see a 1V signal. Agilent’s philosophy is that you want to see the 0.95V signal.

Would it really make that much of a difference, or is all this a load of marketing rubbish and not actually relevant in the real world?
Taking their example, if you knew that it was a perfect 1VDC signal, and you callibrated an Agilent scope to read 1VDC,  then would you not have achieved exactly what Tektronix do?

[Wuerstchenhund]

In the German language we have a saying in the EE world:

"Wer mißt, der mißt Miss"
It really fits well in the German language, because the past tens verb of measuring, literally means "shit"

Freely  translated:
"He, who is measuring is measuring shit"

Because we always have an influence on the the part that we are testing for value.

To be fair, "er misst" (he measures) is present tense. And "der Mist" (single 's') means "crap", so a better translation would be "The one who measures, measures crap" 

Bit yes, that saying is absolutely accurate.

[Neganur]

any chance you have a link to that particular video?

What I noticed is that at least two Tektronix' marketing videos try too hard to make their product look better when comparing to the equivalent Agilent (i.e. they use the instrument wrong).

Also note, that most Tektronix oscilloscopes, at least those I've seen on their website up until and including the 2 GHz MDO5000 series, have gaussian response as opposed to many Agilent scopes that actually have a linear response (typically their 1 GHz and more BW scope models). This dramatically changes how system rise time has to be specified by the manufacturer for each scope&probe combination and how out of band signals are displayed (that is, pulse response).

[h1386343]

any chance you have a link to that particular video?

Im still trying to find it, it was a standard marketing video and the general vibe i got from it was that Tektronix were trying to find everything they possibly could that was wrong on the Agilent scope.

What I noticed is that at least two Tektronix' marketing videos try too hard to make their product look better when comparing to the equivalent Agilent (i.e. they use the instrument wrong).

Tektronix comparing the memory depth of two scopes when trying to capture runt pulses, at around 4:00 they 'forget' to switch on segmented memory on the Agilent, and end the video claiming the Tektronix scope is superior.

[Neganur]

Yeah, and the sample rate of the Tektronix also magically changes from 100 MSa/s 2 Mpts to 500 MSa/s 10Mpts record length when he is taking rise time measurements.
Then he goes on to complain how zooming in on the signal on the Agilent, which is at single shot capture at 200 MSa/s with 4 Mpts record length, looks utter shit. Of course it does.
Besides, he did not set the search thresholds for the runt search so naturally it will not show any events on the search.


I'm not sure those two scopes can be compared, the Agilent is 2.5 GSa/s full channel (5 GSa/s half channel) with 4M memory (2 M full channel) (even the 1.5 GHz bw seems to be some trick that squeezes another 500 MHz out of the 1 GHz version and does so at the cost of half-channel usage only for real time)

While the nice Tektronix scope is a full channel 5 GSa/s (10 GSa/s half channel) full channel 2 GHz bw with "Up to 250 megapoint record length" (which is a paid option) and half channel. Standard is 25 Mpts full channel / 50 Mpts half channel.