Measure signal level on oscilloscope, issues and correct procedure

[neazoi]

Hi, I am trying to measure a signal level on a HP 54520A scope and I want to know if I follow the correct procedure.
The probe is 10:1 at 1Meg, I think it is passive.

I set the probe from the menu of the scope to 10:1
I connect the probe onto the scope input
I set the input of the scope (connector) on 1Meg (the scope can be set for 50R also)
I then do the measurement.

The first measurement is done on a high impedance source.
Does the displayed signal level on the screen of the scope displayed in the correct level?

The second measurement is done on a 50R source, that has a 50R resistor shunted to the ground internally.
I am doing this second measurement without changing anything on the previous scope setup, i.e. the scope is in high impedance mode.
Does the displayed signal level on the screen of the scope displayed in the correct level (and measurements would refer to 50R)?

What about when switching to the FFT, do I need to switch the scope to 50R with the high impedance probe in or just leave it on 1Meg?
This last one is very confusing.

 

[KlausST]

Hi,

* use the correct setup for the voltage rate. eg. 10:1
* use input connector to 1Meg

**********
* "50 Ohm input setup" is for direct measurement of 50 Ohms signals, connected with 50 Ohm impedance cables...and when you want the signal to be terminated with 50 Ohms inside the scope.
* "FFT": no need to change setup when doing mathematical functions.

****

The first measurement is done on a high impedance source.
Does the displayed signal level on the screen of the scope displayed in the correct level?

Yes.
BUT... it shows the signal level at the probe input. It can´t show the unloaded source signal level.
Example: signal source: 5V with internal impedance of 4M. Probe 1M.
--> 4V will be dropped within the signal source. 1V will be at the probe input --> the scope will show this 1V.

Klaus

 

[neazoi]

Hi,
* "50 Ohm input setup" is for direct measurement of 50 Ohms signals, connected with 50 Ohm impedance cables...and when you want the signal to be terminated with 50 Ohms inside the scope.

But you can always measure a high impedance source by setting the scope to 50R and just a coaxial cable, provided that you want to see how this high impedance source will "behave" when loaded to 50R, is that right?

* "FFT": no need to change setup when doing mathematical functions.

So when I connect a 50R coaxial without any probe and set the scope input for 50R, the FFT measurements will refer to the loaded 50R?
Similarly, when I connect the 1Meg 10:1 probe and set the scope for 1Meg input and the 10:1 probe, the FFT measurements will refer to the loaded 1Meg?

Yes.
BUT... it shows the signal level at the probe input. It can´t show the unloaded source signal level.
Example: signal source: 5V with internal impedance of 4M. Probe 1M.
--> 4V will be dropped within the signal source. 1V will be at the probe input --> the scope will show this 1V.

If I understand correctly, you say that this 1Meg probe will load down the 4Meg source (an oscillator for example) and it will show the signal as loaded with this 1Meg probe, right?
Won't the 10:1 probe I use help on that, in the sense that the input impedance of the probe will be 10Megs and not 1Meg, or is this thought wrong?

 

[FvM]

The oscilloscope should always display the correct input level, no matter if 1 MOhm or 50 ohm input termination is selected. Scaling for probe attenuation may be applied automatically or need manual setting.


What however varies with oscilloscope input impedance (= load impedance) is the output voltage the 50 ohm source.

 

[neazoi]

The oscilloscope should always display the correct input level, no matter if 1 MOhm or 50 ohm input termination is selected. Scaling for probe attenuation may be applied automatically or need manual setting.


What however varies with oscilloscope input impedance (= load impedance) is the output voltage the 50 ohm source.

LEt's talk with an example to understand that better.
Say I have an oscillator that when measured on the 1Meg probe it is 10Vpp.
When I put a 10:1 probe and set the scope for a 10:1 probe, it will still show 10Vpp?
Finally, when I remove all probes and I connect the oscillator with a 50R coaxial directly to the scope and also set the scope input to 50R, then the scope will display the oscillator signal as it appears when loaded with 50R, which will be much smaller than 10Vpp.

For all three cases above, when switching to the FFT, the FFT would display the results based on the current loading case (1Meg, 10:1 1Meg and 50R) and these results would be different with each loading case of course.

Are these points correct?

 

[KlausST]

Hi,

The scope will allways show the signal at the probe input. (provided correct gain setup).

Now it depends on your load .. how the ouput value will vary with 10M load, 1M load or 50 Ohms load.

Klaus

 

[neazoi]

Hi,

The scope will allways show the signal at the probe input. (provided correct gain setup).

Now it depends on your load .. how the ouput value will vary with 10M load, 1M load or 50 Ohms load.

Klaus

So my considerations on the example above are correct?

 

[KlausST]

Hi,

So my considerations on the example above are correct?

Nobody can tell how your oscillator behaves.
It´s not unusual that the oscillator stops completely when (over) loaded with 50 Ohms.

Klaus

 

[neazoi]

Hi,


Nobody can tell how your oscillator behaves.
It´s not unusual that the oscillator stops completely when (over) loaded with 50 Ohms.

Klaus

Assuming the oscillator does not stop when loaded with 50R, are these correct?
I am trying to figure out if I have understood the whole think ok, so that I do not read error readings.

 

[betwixt]

For now, forget the 50 Ohms setting, it is only for situations where the cable is also 50 Ohms and the source expects a 50 Ohms load. For almost all purposes you want the probe to be high impedance so it doesn't influence the signal you are measuring. If you really want to see the output in a loaded condition, still use the high impedance mode and add the load directly across the source.

Whether the scope shows the correct voltage depends primarily on whether it has automatic attenuation sensing. Some probes 'tell' the scope whether they are set to x1 or x10 mode, they have extra pins as well as the BNC plug to tell it to change the scale on the screen. If the probe or scope do not support automatic scaling, you have to take it into account yourself. Using x10 will make the trace 10 times smaller.

Don't forget to set the compensation on a x10 probe or your readings could be wildly out.

Brian.

 

[neazoi]

Some probes 'tell' the scope whether they are set to x1 or x10 mode, they have extra pins as well as the BNC plug to tell it to change the scale on the screen. If the probe or scope do not support automatic scaling, you have to take it into account yourself. Using x10 will make the trace 10 times smaller.

Don't forget to set the compensation on a x10 probe or your readings could be wildly out.

Brian.

This scope has a 10:1 probe option in the menu. So as long as I set from the menu the 10:1 probe to be used, I guess that will be enough and the scope will automatically do the leveling?

I did not understand the thing about compensation, shall I do it myself?

 

[betwixt]

I'm not familiar with that scope but I would guess the option is the one to use. It almost certainly does nothing except change the graticule scale, not the waveform height but the number of volts per division the markings represent. Using a x10 probe actually reduces the signal by ten times but the scope can compensate by magnifying the scale ten times. The advantage of a x10 probe is it also isolates the scope's cable and input impedance from the circuit you are investigating, thereby minimizing it's influence on that circuit. The drawback is that for very tiny signals (probably in the uV region) the result is too small compared to system noise to be usefully measured.

Compensation is absolutely essential. Inside the x10 probe there will be a low inductance resistor in series with the tip to drop the signal level but in doing so it makes a low pass filter from the cable and scope input capacitance to ground. It follows that the probe will have a frequency response that rapidly attenuates higher frequencies - not very useful! The fix is to add a small variable capacitor across the resistor such that it compensates for the HF roll-off. If you look at any x10 probe you will see it has an adjustment slot on it, usually the shaft of a tiny pre-set variable capacitor. It needs adjusting to match the probe to the scope while keeping a flat frequency response.

Thankfully, adjusting the compensation is very easy and I think there is a built-in tool in the scope to help you. From photographs I can find, in the bottom right corner of the front panel it looks like there is a ground point and a calibrator output. What you do is this:

1. connect the probe in x10 mode (there is nothing to do in x1 mode if the probe has a switch)
2. connect the probe ground clip to the calibrator ground post
3. hook the probe tip on the calibrator output
4. Adjust the Y gain and sweep speed until you see a few cycles of waveform.
5. CAREFULLY, with a non-metalic screwdriver, adjust the compensation until you get a square wave.

If the waveform has rounded edges, it is under-compensated, if you see pointed overshoot on the edges, it is over compensated. You should try to get the top and bottom flat with square corners. Once you have done it, there is no need to adjust it again for that oscilloscope.

In step 5, I can't stress enough how careful you have to be, the capacitors are usually very fragile so make sure the screwdriver fits the slot well and never force it. If you break the capacitor the whole probe is trashed!

Brian.

 

[neazoi]

I'm not familiar with that scope but I would guess the option is the one to use. It almost certainly does nothing except change the graticule scale, not the waveform height but the number of volts per division the markings represent. Using a x10 probe actually reduces the signal by ten times but the scope can compensate by magnifying the scale ten times. The advantage of a x10 probe is it also isolates the scope's cable and input impedance from the circuit you are investigating, thereby minimizing it's influence on that circuit. The drawback is that for very tiny signals (probably in the uV region) the result is too small compared to system noise to be usefully measured.

Compensation is absolutely essential. Inside the x10 probe there will be a low inductance resistor in series with the tip to drop the signal level but in doing so it makes a low pass filter from the cable and scope input capacitance to ground. It follows that the probe will have a frequency response that rapidly attenuates higher frequencies - not very useful! The fix is to add a small variable capacitor across the resistor such that it compensates for the HF roll-off. If you look at any x10 probe you will see it has an adjustment slot on it, usually the shaft of a tiny pre-set variable capacitor. It needs adjusting to match the probe to the scope while keeping a flat frequency response.

Thankfully, adjusting the compensation is very easy and I think there is a built-in tool in the scope to help you. From photographs I can find, in the bottom right corner of the front panel it looks like there is a ground point and a calibrator output. What you do is this:

1. connect the probe in x10 mode (there is nothing to do in x1 mode if the probe has a switch)
2. connect the probe ground clip to the calibrator ground post
3. hook the probe tip on the calibrator output
4. Adjust the Y gain and sweep speed until you see a few cycles of waveform.
5. CAREFULLY, with a non-metalic screwdriver, adjust the compensation until you get a square wave.

If the waveform has rounded edges, it is under-compensated, if you see pointed overshoot on the edges, it is over compensated. You should try to get the top and bottom flat with square corners. Once you have done it, there is no need to adjust it again for that oscilloscope.

In step 5, I can't stress enough how careful you have to be, the capacitors are usually very fragile so make sure the screwdriver fits the slot well and never force it. If you break the capacitor the whole probe is trashed!

Brian.

All right, I always wondered what this little adjustment on the probe was. Thanks! The probe is the original that came with the scope, probably it is already calibrated but I will check this out.
Ok so I always measure with the probe at 1M and if I want to measure the 50R loaded behaviour I then place a shunt resistor on the probe at the circuit side. Then these readings will refer to the loaded 50R.
In the FFT the same applies as well.
I hope I get this right.
So if I read 10vpp ona circuit at 1Meg and then I read 10vpp on another circuit when loaded at 50R, what does it mean, that the two circuits have different powers or the same? This is tricky.

 

[betwixt]

So if I read 10vpp ona circuit at 1Meg and then I read 10vpp on another circuit when loaded at 50R, what does it mean, that the two circuits have different powers or the same? This is tricky.

Almost!
Measuring 10Vp-p with a high impedance load is telling you the voltage while virtually unloaded. It doesn't tell you that the circuit can't provide the same voltage while under load. A circuit producing 10Vp-p into a 50 Ohm load is obviously proven to be capable of delivering higher power.
Note that the 50 Ohm input impedance of the scope does not mean you can use it as a dummy load! There will be a specification that tells you how much the scope input can withstand. For example, my Tek scope says "1M, 11.5pF, <300V RMS. 50R <5V RMS" on it's front panel.

FFT is just a mathematical way of converting time domain signals (normal X-Y scope display) into frequency domain (frequency spectrum) display by sampling the waveform and analyzing it's periodic contents. It doesn't change the input scaling at all.

Brian.

 

[neazoi]

FFT is just a mathematical way of converting time domain signals (normal X-Y scope display) into frequency domain (frequency spectrum) display by sampling the waveform and analyzing it's periodic contents. It doesn't change the input scaling at all.

Brian.

So if you read a waveform on the scope on the 1M setup, you read the values on the FFT and they refer to the 1M loading. And similarly if you read a waveform on the scope on the 50R setup, you read the values on the FFT and they refer to the 50R loading.

In other words if you read on the FFT +10dbm and you have setup the scope for 1M input, but your circuit you test has a 50R shunt resistor at it's output, then on the FFT, the readings represent the signal when loaded with 50R, am I getting this right?

 

[betwixt]

FFT has nothing whatsoever to do with the input switching, it just shows a calculated spectrum of frequencies based on the amplitude samples of the input signal. If it reads +10dbm then that is what was at the scope input at that frequency, regardless of how the termination is set.

Brian.