# What's a safe method for determining the switching frequency of a blackbox ACDC converter?

My project uses an off-the-shelf ACDC converter that conceals design details from such as switching frequency (or SF range). It's 24 V output rated to 5 Amps. I want to measure it's switching frequency but was advised against probing the device with an oscilloscope since it can be dangerous. Any advice how I can accomplish this safely?

• SAFELY probe it with an oscilloscope – JonRB Mar 21 at 19:53
• But seriously you can use an oscilloscope, just know the capability of the probe and capability of the scope... I regularly hook >2kV to a scope, via an appropriate probe. – JonRB Mar 21 at 19:54
• Would connecting the probe ground clip to the earth ground tab be appropriate? It's the same earth tab as the incoming AC. – Big Al Mar 21 at 20:01
• See... I would use a high voltage diff probe, not a GND reference x10 probe BECAUSE that can be unsafe... – JonRB Mar 21 at 20:02
• What makes the x10 probe unsafe? – Big Al Mar 21 at 20:50

## 2 Answers

simulate this circuit – Schematic created using CircuitLab

This is one way; with a load on the dc power supply output you can examine the ac ripple on a scope. Use a high-pass filter (HPF) that presents a load of 50 to 100 $$\\Omega\$$ (as shown in the schematic) with a 1:1 oscilloscope probe. You should expect to see 50mV or so.

• Are you implying that I should measure the ripple on the 24 V, which we of course expect will be the same as the switching frequency? Could I safely measure the switching frequency "directly"? – Big Al Mar 21 at 20:03
• Yes and yes but you get cleaner actual ripple measurements this way and not with a resonant 10:1 probe with long ground leads as long as it is isolated !! and not some offline grid connected supply – Tony Stewart Sunnyskyguy EE75 Mar 21 at 20:04
• And to be clear, you mean with a load on the DC supply in addition to the HPF network? Because I figure C1 would block the DC. – Big Al Mar 21 at 20:32
• Yes and yes but high impedance may cause false noise ringing. Some DSO's have DC coupled 50 Ohm options. where DC coupled would blow the terminator and without terminator can cause 10: 1 probe to ring but one can even put a 1:1 scope probe shorted to its long ground clip in a loop antenna and hold it over the power supply and pick up MMF (magnetic motive force or current loop noise.) Why is ripple frequency important to you? it will have harmonics if not a square wave so a 50 Ohm AC load is best for accurate spectrum analysis while having a dummy DC load – Tony Stewart Sunnyskyguy EE75 Mar 21 at 20:38
• or whatever you plan to use it for like a BLDC motor which has its "own" ripple frequency – Tony Stewart Sunnyskyguy EE75 Mar 21 at 20:40
1. A pure resistive load on the power supply that will draw at least 50% of its rated output current. If you have an option, do not use a resistor described as "non-inductive".

2. Follow that with a single-pole, high-pass filter network. I have an old Topaz UPS with a switching frequency of 6 kHz. Assuming your supply is not that, any network above 20 kHz will help isolate the signal you are looking for. 7500 ohms and 1 nF gets you 21 kHz; that seems like a good starting point.

3. An alternate (or confirming) approach: Something that is a problem when measuring circuit noise can work to your advantage here. A long ground lead and clip on a scope probe can form a loop antenna and mess up conducted noise measurements. That is exactly what you want. Clip a 1K resistor between the probe tip and its ground lead and wave the loop around the supply. Depending on how RFI "tight" are the design and packaging, you might get a usable signal.

4. Note - All switching power supplies are not based on PWM techniques. With a constant resistive load you probably will find a useable signal and get the operating frequency at those conditions. Pulse Position Modulation and hysteretic control are just two alternative topologies that have operating frequencies that vary with the load.