How to Probe in a Noisy Environment

Thread Starter

TechWise

Joined Aug 24, 2018
151
I am interested to know how members go about probing relevant nodes in power electronic circuits. Conventional scope probes suffer from very long ground leads which pick up a lot of noise and add a lot of inductance to the path.

I recall a technical article on here about placing SMA connectors on the PCB and then using an SMA to BNC adaptor to plug that straight into the scope. I was rather nervous about this as the lead won't have the 10x attenuation and I don't want to blow up the input of my scope.

I have also tried the ordinary little through-hole test loops which are very convenient for DC signals and low frequency stuff but again you're stuck with a conventional probe and long ground leads.
 

tindel

Joined Sep 16, 2012
936
Measuring power electronic circuits is not trivial. High-speed and high-voltage edges are common. Common-mode voltages can also make the measurement difficult. Each measurement also requires different techniques. Ex: Switching node measurement techniques will be much different than the techniques used to measure the output switching noise.

Generally, the best measurements are via terminated coax (usually 50ohm) or active differential probes. Single ended 10x probes can be used as well with good technique.

Here are several high-speed appnotes by the late Jim Williams that goes into most of the measurement techniques that can be used. My personal favorite to show young engineers how to make measurements is App Note 70.
 

Thread Starter

TechWise

Joined Aug 24, 2018
151
Measuring power electronic circuits is not trivial. High-speed and high-voltage edges are common. Common-mode voltages can also make the measurement difficult. Each measurement also requires different techniques. Ex: Switching node measurement techniques will be much different than the techniques used to measure the output switching noise.

Generally, the best measurements are via terminated coax (usually 50ohm) or active differential probes. Single ended 10x probes can be used as well with good technique.

Here are several high-speed appnotes by the late Jim Williams that goes into most of the measurement techniques that can be used. My personal favorite to show young engineers how to make measurements is App Note 70.
In the lab we do have a top of the line Tektronix scope with the fantastic IsoVu probes which are the ultimate solution to the problem. The setup was eyewateringly expensive though as you can imagine! I'm looking for a setup that's less high end but marginally better than using a standard 10x scope probe with its associated ground lead inductance and pickup.

I will have a look through the app notes you have linked to. I have found the great Jim William's notes very useful in the past.

Would you mind expanding a bit on the "terminated coax" you mentioned? Would this involve designing an SMA connector into the power converter then screwing on an SMA to BNC and straight into the scope?
 

tindel

Joined Sep 16, 2012
936
If you're using IsoVu probes you're likely designing SiC or GaN based power supplies or motor drivers and need the isolation (safety), high-voltage, and high-slew rates associated with them. Tektronix is making a KILLING on these probes right now for a reason. They are the only thing available on the market that will do the job.

I would talk to the person that specified the IsoVu probes and ask them why they specified them and didn't look for a cheaper solution. There likely isn't a cheaper solution.

For perspective, we have six IsoVu probes at our office. They are expensive, but what's the cost of being second to market? Don't be penny wise and pound foolish.
 

Thread Starter

TechWise

Joined Aug 24, 2018
151
If you're using IsoVu probes you're likely designing SiC or GaN based power supplies or motor drivers and need the isolation (safety), high-voltage, and high-slew rates associated with them. Tektronix is making a KILLING on these probes right now for a reason. They are the only thing available on the market that will do the job.

I would talk to the person that specified the IsoVu probes and ask them why they specified them and didn't look for a cheaper solution. There likely isn't a cheaper solution.

For perspective, we have six IsoVu probes at our office. They are expensive, but what's the cost of being second to market? Don't be penny wise and pound foolish.
Yes indeed, the system was procured for looking very closely at SiC devices. I know why they were specified and bought - as you say they are the only thing on the market.

I still think there is no harm in having a halfway house between a six-figure setup and a bog standard probe that came with the scope. Particularly if it's only to get a look at signals that aren't in the presence of huge common mode voltages, like the bottom side gate drvie and even the voltage at the switch node when running at low voltage.
 
Top