Please excuse my ignorance but I've been trying to run a simple bode plot of a lowpass filter (R=16Ohm, C=01.uF) with an Fc=100kHz.
I am using the 1104x-e with a sdg1032x and I'm having the following problem...
If I use only out1 of the AWG with a T-bar the results look fine. However if I try to use both channels of the AWG (with the tracking mode ON) the Fc is around 25kHz instead of the correct 100kHz.
Anyone knows what am I doing wrong? I've spent a lot of time yesterday and today fighting with my lack of knowledge and so far I'm loosing this battle...
Is it fun that 50 ohm (generator source impedance) + 16 ohm is 66 ohm. With this R and 0.1uF C it give bit over 24Hz... at least it’s a really funny coincidence, is it. Think about it, at least. 
I don’t want to take a further stand because I haven’t seen (whole setup) the whole reality you have in every factor involved.
Of course , I feel stupid I missed that... Thanks for pointing it out!
However, I still don't understand why with. single channel and a T-bar, the results were correct.
PS. I just tried with some more reasonable values (R=1kOhm, C0.1uF) and in both cases I got the correct results, Fc=1kHz.
Think about it. Think where from come signal what is reference.
Think this generator internal R1 and then your external RC circuit R2.
Bode Plot reference channel come from this point between R1 and R2 and is also input to your RC what is just this R2 and then C.
In this single channel method, the internal resistor R1 certainly acts but since the reference point is after it, its effect is eliminated because BP compares only DUT in and DUT out.
This does not happen with 2 channel operation because your RC will not load this other channel what is used only as a reference.
The two-channel method has some important advantages, but also some disadvantages, and one disadvantage is exactly what happens in this case.
That makes sense. Thank you very much!
What would you consider as advantages in the two-channel method that the single channel doesn't have?
Here's a little seeds from which can to grow thoughts:
• When using low frequencies and simple objects such as simple RLC type LP, BP and HP filters and so on, just nothing or nearly nothing.
• But when we go to higher frequencies, things may concern impedance matching and signal travel time matching.
Example At 100 MHz, depending on the cable type, the travel time can be, for example, around 0.5 ns 10 cm.
With this 100MHz, this alone would mean, for example, a phase shift of about 18 degrees and naturally angle change when frequency change. Then if T-split to oscilloscope 1M 20pF input and to DUT example 50ohm input this branch going into the channel of the oscilloscope can cause quite a problem. At low frequencies it can be forgotten, but at higher frequencies it can look cluttered. Of course, there are ways to reduce problems but it also requires the right supplies, knowledge and experience.
The signal travel time of the reference channel should be the same as the travel time of the cables of the branch that goes through the DUT (If user is interested about phase).
When we split the signal into the DUT input and the BodePlot reference channel (Dut In), the T split can be also a bit problematic in impedance matching and for travel time.
• But then comes a thing that can’t be done at all other than using the generator’s 2-channel. Suppose a DUT causes a constant frequency shift (example as RF mixer). With the single channel method, its analysis would be impossible. (It need remember that this Siglent BodePlot is as frequency selective level meter and receiver frequency stepping sweep follow this frequency what it command from generator during frequency stepping sweep) With the two-channel method, it is possible because you can put the generator in the tracking mode and use frequency offset for that slave channel. FRA do not know this slave channel frequency. It only listen this frequency what generator master channel is commanded to produce. (all three Dut Out channels follow this for measure signal and compare these to Dut In signal what can also think as reference signal (in default it is oscilloscope Ch1 but if user want it can change).
• We can also meet test situations where we want the voltage level of the reference signal to remain suitably high (example for less noise for better phase detect) or constant when, for example, DUT have high amplification and / or when we are analyzing something using variable level (level profile) during frequency sweep (Sweep Type / Vari-level).
Also in some times it is perhaps good to use different levels to Reference channel (dut in) for reduce channels Cross Talk effect. (example if we are looking some rf frequency filter with high stop band attenuation so that reference channel cross talk to DutOut channel does not obscure it from view.
Also it is important to select:
Vout/Vin or
Vout in Display Amplitude menu and in Stimulus menu select right Load value (50, 75, 600, Hi-Z or free variable starting from 50ohm to 100kOhm).
In more complex tests and measurements than RLC filters for audio frequencies, there may be quite a few things to consider when setting up and how to connect the DUT to the system.
Home
Help
Search
Login
Register
