Rigol MSO5000 Generator <--> Scope Amplitude Drop

[gawlas]

Hi,

I've just bought my first oscilloscope, and after some initial tests I would like to ask you some question.

I've tested build-in generator, and I noticed that when I use BNC cables, after some frequency, amplitude starts dropping a bit (no matter if I use 50ohm terminator or not), as you can see on attached images when I use BNC cables, at 25MHz amplitude drops around 5-7%.

I'm aware of transmission line theory, but thought that with 1m cable at 25MHz with 50 Ohm termination there shouldn't be any drops.

I would like to ask you if it is normal or I need to buy better cables/connectors.

BNC cable, connector and terminator are from Amphenol, so there are not the cheapest.

[JeremyC]

When you make the test with 50 ohm terminator, you should change the generator output from "HighZ" to "50 ohm".

[gawlas]

As I understand it only changes what is displayed on screen (divide actual amplitude by 2), so doesn't really matter.

I know that, when using 50ohm terminator amplitude on scope should be half of generated (1V), I'm asking about difference (drop) between 1kHz vs 25MHz

[JeremyC]

As I understand it only changes what is displayed on screen (divide actual amplitude by 2), so doesn't really matter.

I know that, when using 50ohm terminator amplitude on scope should be half of generated (1V), I'm asking about difference (drop) between 1kHz vs 25MHz

It may matter, but I'm not sure. It doesn't hurt to change it and repeat the test. BTW, how it looks when frequency is 20MHz? 25MHz it's the max.

[gawlas]

I've tried with '50Ohm' option in generator and results are the same.

Here some more screens from tests with BNC cables - 10MHz an 20Mhz, I didn't attach them before as there is limit per post.

As you can see amplitude is dropping with frequency:

BNC 0.5 meter (direct):
1kHz - 2.03 V
1MHz - 1.97 V
10Mhz - 1.97 V
20Mhz - 1.95 V
25Mhz - 1.93 V

BNC 0.5 meter (50 Ohm terminated):
1kHz - 0.99 V
1MHz - 0.98 V
10Mhz - 0.96 V
20Mhz - 0.94 V
25Mhz - 0.92 V

[thm_w]

10.1 dBm to 9.69 dBm thats not much of a drop. https://www.aetechron.com/calculator-Volt2dB.shtml

You can find the spec in the datasheet:
Frequency Range 100 mHz to 25 MHz
Flatness ±0.5 dB (relative to 1 kHz)

So its within spec.
If you need something better, get a dedicated function generator. The built in AWG is kind of a toy (although may be useful for certain automation applications).

[rstofer]

Almost everything rolls off with frequency.  A 100 MHz scope will likely show the trace -3dB down at 100 MHz.  You can expect yours to do this very thing.  At -3dB down, the voltage is 70.7% of the low frequency value.

The question is: How much should the scope roll off at 25 MHz.  See Summary point 1 and perhaps go back to the equations.

https://edadocs.software.keysight.com/kkbopen/how-does-gaussian-rolloff-factor-into-oscilloscope-bandwidth-selection-criteria-part-1-614517535.html

Then there is the issue of the AWG rolling off at 100% of bandwidth...

https://beyondmeasure.rigoltech.com/acton/attachment/1579/f-0907/1/-/-/-/-/MSO5000_datasheet.pdf

[xrunner]

The manual looks like the waveform generator output is +/- (2% of setting + 1 mV) (freq. = 1 kHz)

So 2V setting would give +/- .041 V accuracy, from 1.959 to 2.041 V . Does that make more sense?

[gawlas]

Thanks for answers, I think I understand this a little bit better now, but have two more questions:
 
- 10x probe shows almost exactly 2V over entire range (drop only with BNC coaxial cable), so I don't think generator accurracy/flatness is an issue, am I right?
- If amplitude drops because I'm getting closer to my scope bandwidth, then why probe shows correct values? Does this 5 times rule does not apply to probe?

[xrunner]

This isn't the same kit you are using, but for comparison I used my most "inferior" scope and a brand new scope to check a 2Vpp sine wave using a 1 m BNC cable. The signal came from a Rigol DG1032Z.

The poorest scope I have is the TDS-210 (60 MHz BW) and the best is a new Siglent SDS2202X-E opened to 400 MHz BW.

I checked the signal on both and got these Vpp measurements -

TDS-210
1 kHz   2.06
25 MHz   1.96

Siglent SDS2202X-E
1 kHz   2.06
25 MHz   2.00

They seem reasonable to me.

[thm_w]

Thanks for answers, I think I understand this a little bit better now, but have two more questions:
 
- 10x probe shows almost exactly 2V over entire range (drop only with BNC coaxial cable), so I don't think generator accurracy/flatness is an issue, am I right?
- If amplitude drops because I'm getting closer to my scope bandwidth, then why probe shows correct values? Does this 5 times rule does not apply to probe?

Look up the input impedance of the oscilloscope probe vs the oscilloscope itself.
You are essentially putting a large resistor in series with the input when you use the oscilloscope probe, so the loading will be less.