Oscilloscope Probe coupling

[FreshmanNewbie]

I am using this Buck Converter - https://fscdn.rohm.com/en/products/.../power/switching_regulator/bd9g201efj-m-e.pdf

Schematic:


Buck Converter Specifications :

Input Voltage - 18V to 32V

Switching Frequency - 300kHz

Output Voltage 9V

Load Current - 0mA to 200mA Maximum.

While I am measuring the switching frequency using an oscilloscope, I am placing the probe directly on the TP1102 and also measuring the output voltage on the 47uF capacitor pad.

I got the below waveforms :

Input Voltage of 14V:



However, with the same input voltage, I am just removing the probe which was placed at the switching frequency node. Now, I didn't observe the ringing.
Input Voltage of 14V without probing the switching frequency :


As you can see, the ringing vanished once I disconnected the probe.

Can someone tell me how this ringing at the switching frequency node is happening even though I am probing at the TP1102 and how this is coupling at the other probes also?

Please also provide a solution on how to avoid this ringing while measuring the values.

Below is the ringing frequency:

 

[barry]

There is no “switching frequency node”. Do you mean pin 1? If so, it’s not surprising that the inductance and capacitance of the probe are upsetting things. What are you trying to measure that makes this an issue?

 

[Easy peasy]

separate ground connections can cause this effect - also just CM pick up by the scope channels - the extra wiggle you see are artifact ( mostly ),

@barry, TP1102, otherwise known as pin 1 - is indeed a switching node - and it switches at the switching frequency - it os possibly the easiest and most convenient point to measure said frequency ...

 

[FreshmanNewbie]

separate ground connections can cause this effect - also just CM pick up by the scope channels - the extra wiggle you see are artifact ( mostly ),

@barry, TP1102, otherwise known as pin 1 - is indeed a switching node - and it switches at the switching frequency - it os possibly the easiest and most convenient point to measure said frequency ...

Thank you. Could you please let me know why this is happening and how to eliminate this effect?

 

[danadakk]

Probing -

ABCs of Probes Primer | Tektronix

Precision Measurements Start at the Probe Tip Probes are vital to oscilloscope measurements. To understand how vital, disconnect the probes from an oscilloscope and try to make a measurement. It can’t be done. There has to be some kind of electrical connection, a probe of some sort between...

www.tek.com

Not all capacitors, for same C value, equal in ESR performance -

Regards, Dana.

 

[FvM]

I presume you have standard passive 10:1 probes. Can you show a photo how you connected the probes?

Problem is probe input capacitance loading the switching node TP1102. It causes a common mode transient running down the ground clip and probe cable. Part of it couples into the ground node and also the other probe cables. Don't know if the circuit has a clean ground suitable as probe reference.

Several ways to reduce crosstalk:
Use very short ground connection for probes, optionally toroid cores as common mode chokes for sensitive probes. Use active probes or passive resistive probes for fast signals.

 

[dick_freebird]

A problem is, that the grounds are common
although the tips are not.

So your SW node's edge rate couples through
probe tip capacitance to all-probes-ground,
retains some energy against the scope-
ground-to-signal-ground inductances, and
back from ground to all the other channels.

I like to "degenerate" the probe tip with a series
1K resistor (making a 10pF*1K or 100MHz pole,
but more importantly limiting how hard one
tip can possibly kick the channels-ground).
If you're really caring about edge-kicked tanks'
ringing attributes then a HF FET probe would
probably be better, but in general these little
tone-bursts are (1) not relevant to top-line
converter performance, (2) often an artifact
of the measurement method itself.

Ringing, if real, can lead to things like pulse
width jitter in current mode controlled buck
converters, leads directly to comparator chatter
/ jitter. Determining just what -is- real, is a task.

Those bursts, if present on an un-poked product,
can present a nasty little EMI tone-bundle out
in the higher frequencies, one which is unrelated
to fSW. This is a "handle" for determining the EMI
source - what does it follow, who is its master?

 

[FreshmanNewbie]

A problem is, that the grounds are common
although the tips are not.

So your SW node's edge rate couples through
probe tip capacitance to all-probes-ground,
retains some energy against the scope-
ground-to-signal-ground inductances, and
back from ground to all the other channels.

I like to "degenerate" the probe tip with a series
1K resistor (making a 10pF*1K or 100MHz pole,
but more importantly limiting how hard one
tip can possibly kick the channels-ground).
If you're really caring about edge-kicked tanks'
ringing attributes then a HF FET probe would
probably be better, but in general these little
tone-bursts are (1) not relevant to top-line
converter performance, (2) often an artifact
of the measurement method itself.

Ringing, if real, can lead to things like pulse
width jitter in current mode controlled buck
converters, leads directly to comparator chatter
/ jitter. Determining just what -is- real, is a task.

Those bursts, if present on an un-poked product,
can present a nasty little EMI tone-bundle out
in the higher frequencies, one which is unrelated
to fSW. This is a "handle" for determining the EMI
source - what does it follow, who is its master?

Thank you for the answer. Could you please tell me how you arrived at the solution to create a pole at 100MHz? Why that 100MHz value and how will it solve the ringing artifacts?

 

[Easy peasy]

put all ground clips on exactly the same point for a start and see if this improves things ....

 

[std_match]

I like to "degenerate" the probe tip with a series
1K resistor (making a 10pF*1K or 100MHz pole,
but more importantly limiting how hard one
tip can possibly kick the channels-ground).

I also like a series resistor. It makes the probe's ground connection less important.
I often don't connect the probe's ground clip at all when using a series resistor.

Edit:
Some ground connection is probably needed between the scope and the "device". The ground clip on another channel may be connected to the "device".

 

[danadakk]

A problem is, that the grounds are common
although the tips are not.

So your SW node's edge rate couples through
probe tip capacitance to all-probes-ground,
retains some energy against the scope-
ground-to-signal-ground inductances, and
back from ground to all the other channels.

I like to "degenerate" the probe tip with a series
1K resistor (making a 10pF*1K or 100MHz pole,
but more importantly limiting how hard one
tip can possibly kick the channels-ground).
If you're really caring about edge-kicked tanks'
ringing attributes then a HF FET probe would
probably be better, but in general these little
tone-bursts are (1) not relevant to top-line
converter performance, (2) often an artifact
of the measurement method itself.

Ringing, if real, can lead to things like pulse
width jitter in current mode controlled buck
converters, leads directly to comparator chatter
/ jitter. Determining just what -is- real, is a task.

Those bursts, if present on an un-poked product,
can present a nasty little EMI tone-bundle out
in the higher frequencies, one which is unrelated
to fSW. This is a "handle" for determining the EMI
source - what does it follow, who is its master?

Shouldn't that be 100 ohms, 10 pF for 100 (actual 159) Mhz ?

Regards, Dana.