how to properly use a Tektronix P6046 differential probe?

[crazyhog]

Hi all,

I am doing self study of EE.  I got a Tek P6046 differential probe from ebay (in very good condition).  I discover the measurements made with the probe are lower than what I had expected.  Since I am a newbie, I am not confident that I've properly hook things up.  Before I go changing the probe's calibration, I'd like to first verify if I've made any user error.

My test setup:
lab power supply DC output: 800mv
oscilloscope channel 1: passive probe, yellow trace, measured ~800mv
oscilloscope channel 2: p6046 diff probe, blue trace, measured ~600mv using 50 ohm termination

I've connected the p6046 diff probe according to the manual (see attached manual screen capture)
probe amplifier -> bnc patch cable -> 50 ohm terminator -> oscilloscope

See "tek diff prob 50 ohm.jpg".  Notice
- the 50 ohm terminator circled in red 
- channel 1, yellow trace, passive probe, reading 800mv
- channel 2, blue trace, P6046 diff probe, reading ~600mv

I also experimented terminating with 100 ohm.  See "tek diff probe 100 ohm.jpg".  Notice I simply use a 100 ohm resistor as a terminator.  With 100 ohm termination, the passive probe and differential probe are basically in agreement! 

I've also tried no terminator and the diff probe measurement overshoots and went off screen.   

Also note that I've measured the resistance of the bnc patch cable which is ~0 ohm.   I do believe the patch cable has a "characteristic impedance" rating of 50 ohm.  However, I've not study transmission line theory yet and I am not sure it is the right cable to use.  Is my understanding correct that a bnc patch cable can be rated with "characteristic impedance" of 50 ohm but the measured resistance (with DC) is 0 ohm?

I had expect, given what is described in the manual, a 50 ohm terminator to give the correct value.  What am I doing wrong?  Or is it the probe's calibration that is off?

Thank you for your help.

[tautech]

For heavens sake man apply the KISS principle.

Sanity check:
Connect directly to the scopes probe Cal output and set the ch input attenuation to get a waveform that matches the probe Cal output.

If you're still unsure use an ordinary probe as well and compare the two traces.

Edit
And BTW we don't need massive sized pics to see what's happening, screenshots from the scope will suffice.

[crazyhog]

Tautech,


Perhaps you just quickly glanced over my message and didn't read it properly. The only difference between what you've suggested and what I've write up is I used a simple DC source and you prefer the scope cal output.  Since my focus is only on the magnitude of the measurement and not the transient issue of the probe, isn't a DC source simpler than a square wave?  I can also test voltages large and small with a twist of a knob.  Can you with a scope cal output?  You'll notice I did compare the differential probe with an ordinary passive probe.  I already know matching the attenuation will get the right reading.  But as in math, I don't just want the right answer, I want to do the problem right!

Again here is what I needed:  Did I setup the physical wiring of the Tek P6046 differential probe correctly:

"probe amplifier -> bnc patch cable -> 50 ohm terminator -> oscilloscope"

If my wiring is wrong, what's wrong with my understanding of the tek manual and how should I correct the wiring?

Tautech, I included the full photo so that folks can see my wiring. 

Tautech, man, chill with the attitude!:)  We all have to start some where.  Can we establish a common ground that the culture of a forum community matters?

[mtdoc]

A differential probe allows you to measure the "difference" in voltage between its two inputs ( and regardless of where those two inputs sit relative to ground).

Ask yourself if that is what you are doing with your setup....

[crazyhog]

Hi mtdoc,

Thanks for the reply.  It is measuring the voltage difference.  The voltage difference is between + input at 800mv and - input at ground (i.e. 0v).  Notice in the picture, I installed the grey grounding shroud on the - input (see green oval in attached pic).  It shorts the - tip to the ground.  The manual call this configuration "Single-Ended Operation".  I purposely set up the test this way to keep it simple.  Heh, another common ground I have with tautech even if he and I got off a bit on the wrong foot

But for what its worth, I also experimented with measuring difference between voltages above ground and the measured result is still lower than what I had expected using the 50 ohm terminator but matches with 100 ohm terminator. 

Regardless, can you confirm the following wiring is correct for the tek P6046 probe?
probe amplifier -> bnc patch cable -> 50 ohm terminator -> oscilloscope

If it is correct, I will tackle the probe calibration.  I wonder if the previous owner has purposely calibrated the probe for 100 ohm termination operation.  However, 100 ohm termination doesn't seem like a commonly used setup, at least not on ebay.

Thanks!

[tautech]

Ch
It's just a probe and like any other if it's connected as instructed and the amplitude is not correct/exact then there's a few tricks to adjust it........but the integrity of the signal is paramount.
So the probe Cal is where one should start, if there are only minor differences in amplitude they can be addressed with the Fine vertical gain. Personally I wouldn't piss around with ANY internal adjustments of a differential probe unless you have the right equipment and know exactly the procedure.

At higher frequencies correct transmission line termination is important to avoid reflections and ringing and with 100 ohm termination you might get into trouble.
In short don't try to attenuate with other than 50 ohm in some manner to adjust the signal amplitude, for signal integrity it need  be done in the scope of with adjustment of the differential probe gain.
Have you checked the probes PSU is supplying the correct voltage ?

Measure everything and presume nothing. 

[mtdoc]

Hi mtdoc,

Thanks for the reply.  It is measuring the voltage difference.  The voltage difference is between + input at 800mv and - input at ground (i.e. 0v).  Notice in the picture, I installed the grey grounding shroud on the - input (see green oval in attached pic).  It shorts the - tip to the ground.  The manual call this configuration "Single-Ended Operation". 

OK good.   (I missed that in the original pic).    As tautech says - measure everything and confirm that your single ended setup is as it should be.  If you haven't done so already, use a DMM to measure the voltage difference at the probe inputs (how does that compare to the scope reading?). Also, confirm that the negative input is tied to the probe ground.

Regardless, can you confirm the following wiring is correct for the tek P6046 probe?
probe amplifier -> bnc patch cable -> 50 ohm terminator -> oscilloscope

Yes, that should be fine. - But measure your terminator to confirm it is 50 ohm.

[alm]

Your connections look fine. But the scope is set to 200 mV/div. That is wrong. See the printing on the side: it is designed to work with a scope set to 10 mV/div. What you are doing is likely clipping the output amplifiers.

Connect the amplifier to the scope with the 50 Ohm termination. Set the scope to 10 mV/div on that channel, and never change it as long as you have the P6046 connected; only change the vertical sensitivity on the P6046 amplifier. Then without anything connected to the probe tip, set the attenuator balance as described in the manual (page 1-10 in my copy). On a DSO, you do not have to set the input coupling to GND. Just adjust the ATTEN BAL until the DC level is as close as possible to 0 V, and then compare to other attenuator settings (on the amplifier!).

After that, set the P6046 amplifier (not the scope!) to 200 mV/div, and check if the signal is 4 divs high, or 40 mV according to the scope. Sorry, you can not be lazy and rely on automated measurements without engaging your brain . You might be able to set the attenuation ratio to 20x (\${200 mV/div \over 10 mV/div} = 20\$), to get the measurements to match, but you would have to keep changing the attenuation ratio (not vertical sensitivity) as you change the attenuation on the amplifier.

[David Hess]

I expect alm has identified the problem.  From page 2-3 of the manual:

The amplifier for P6046 is an optional P6046 probe accessory consisting of an amplifier unit and a power supply unit.  It makes the probe compatible with any oscilloscope and plug-in combination which has 10mV/div deflection capability and appropriate bandwidth.  Operating it into a less sensitive device (higher mV/div) will provide unreliable results because of overdriving the probe and amplifier.

[crazyhog]

Hi David and Alm,

Thanks for the feedback.  Actually I did originally set the scope up for 10mv/div.  I got tire of having to mentally calculate the actual voltage instead of just reading the value off the screen.  Then I realize there is a simple shortcut! 

The probe multiplier (i.e. 1x, 10x, 20x, etc.) is a simple scaling factor that doesn't actually affect the oscilloscope signal acquisition!  It's only use to calculate the voltage to present on the screen.  The reason is simple.  Let's say we are using a 20x probe.
 The 20x probe resistor is at the probe, outside of the oscilloscope.  By the time the voltage get to the scope input, the voltage has already dropped!  So the oscilloscope just use whatever probe multiplier you tell it to multiple the voltage reading before presenting the result on the screen.

Here is how I verified my understanding. 
 
1.  setup my scope in single shot mode
2.  setup 1x probe multiplier
3.  setup 10mv/div
4.  do a force a single trace capture
5.  take a picture of the screen
6.  change probe multiplier to 20x
7.  take a picture of the screen (notice this is same data from step 4)

When I change the probe multiplier (without redoing the single-shot trace), the oscilloscope automatically change volt per division and Max voltage reading using the multiplier I gave it.  Alm, you are right in that if I change the volt/div at the amplifier (not the scope), I need change the probe multiplier at the scope. 

As for the manual's statement "Operating it into a less sensitive device (higher mV/div) will provide unreliable results because of overdriving the probe and amplifier".  Actually, when I read that in the manual, I don't see how that can happen.  The scope has 1M ohm input resistance vs. the 50 ohm resistance of the terminator.  To the probe amplifier, the 50 ohm dominates regardless of what the volt/div the scope is setup at.


Hi Tautech, mtdoc, alm, and david,

Thank you all for your input!  You gave me the confident that my setup is correct and that I should try to calibrate the probe using a 50 ohm terminator.  I will make it so.



 

[David Hess]

As for the manual's statement "Operating it into a less sensitive device (higher mV/div) will provide unreliable results because of overdriving the probe and amplifier".  Actually, when I read that in the manual, I don't see how that can happen.  The scope has 1M ohm input resistance vs. the 50 ohm resistance of the terminator.  To the probe amplifier, the 50 ohm dominates regardless of what the volt/div the scope is setup at.

The schematic shows that the emitter follower output stage driving the oscilloscope has an operating current of about 5 milliamps worst case so it literally cannot reliably produce more than 250 millivolts at the 50 ohm oscilloscope input through its double termination and at that point, it would have become very nonlinear.  10mV/div yields +/-40 millivolts which is much more reasonable.  So the P6046 amplifier really was designed to drive a 10mV/div input.

[crazyhog]

Hi David,

The important bit is the probe multiplier(i.e. 1x, 10,x, 20x, etc.).  The voltage/div and all other voltage readings on the oscilloscope screen is after the oscilloscope has applied the probe multiplier. 

The actual voltage at the scope input in my 2 previous screen capture are exactly the same even though the oscilloscope screen is showing different voltage level and different volt/div settings.