Here is the conclusion up front: if you are debugging USB Power Delivery on a real product, use a Keysight mixed signal oscilloscope as the primary recorder and treat the 115 multimeter as a sanity checker. A multimeter will tell you that 20 V is present. It will not tell you whether 20 V survives a load step, or why the PD controller falls back to 5 V. I learned that in March 2023, and the lesson cost about two weeks of schedule.
I am a hardware test engineer and I have handled USB-C pre-compliance validation work for eight years. I have personally made and documented 11 significant measurement mistakes, totaling roughly $14,000 in rework. I keep a checklist because I do not want other people in my lab to repeat the same failure.
The expensive mistake happened in March 2023. A USB-C adapter advertised a 20 V/5 A PDO. The mixed signal oscilloscope was booked for another project, and I told myself, 'this is only a static DC check, I do not need the full setup.' I used a 115 multimeter on the bench. It read 20.03 V. Clean. I wrote it down as a pass. Then the electronic load hit the output. The rail folded, the product renegotiated down to 5 V, and the demo failed.
If I had kept only the MIN/MAX list from the 115 multimeter, the list would have shown 20.03 V and later 5.06 V. It would not have shown when the voltage moved or what caused the move. The meter was measuring, but it was not recording a useful story. A USB Power Delivery contract is a negotiation, not a static state. The source advertises its capabilities on the CC line, the sink requests a PDO, and then the source must hold that voltage while the load changes. None of that appears on a DMM display.
That event changed how I think about USB-PD test equipment. The question is not 'does the source produce 20 V under no load?' The question is 'does the source keep 20 V stable when the sink asks for the current that the PDO promises?'
My USB-PD test order now
The order matters: oscilloscope first, handheld DMM second, arbitrary waveform generator when I need repeatable disturbance tests. Skipping the scope to save time almost always costs more time. Trust me on this one: the scope setup is not extra bureaucracy. It is the shortcut.
1. Keysight mixed signal oscilloscope
When I look at a PD failure, I want VBUS on an analog channel, CC1 and CC2 on the digital channels, and the load-enable line on another input. One time base ties them together. A Keysight mixed signal oscilloscope does this in one box. With the right protocol-decoding software, I can see the USB Power Delivery messages on the CC line and measure the time between the Request, Accept, and PS_RDY. Even without protocol decode, the raw CC waveform plus the VBUS waveform gives me more information than any list of DC averages.
I will not claim every model decodes USB PD out of the box. Check the model and software options before you buy. The core advantage is not packet labels. It is time-correlated analog and digital acquisition.
2. 115 multimeter for spot checks
The 115 multimeter still has a place on my bench. I use it to confirm that VBUS is at a sane level before I connect a scope probe, or to check whether a rail is completely dead. That is it. If the scope shows a dip and the multimeter does not, I trust the scope. The meter averages or samples slowly, and an intermittent 2 ms dropout can hide between samples.
3. Keysight 33600A waveform generator for repeatable disturbances
Some USB Power Delivery bugs only appear when the DC input is disturbed. For those cases, I use the Keysight 33600A waveform generator to create a clean, repeatable arbitrary waveform. The waveform is coupled into the power path through a small coupling network. The 33600A is not the power source; it is the disturbance source. The MSO runs in segmented mode and captures each event. If the PD contract drops during the test, the recorded list gives me a timestamp and the exact waveform shape.
USB Power Delivery while recording list: what I actually record
If you have been searching for that phrase, you probably want an instrument that records more than one final reading. Here is the checklist I follow now:
- Advertised PDO list from the source, normally 5/9/15/20 V, and the measured VBUS level for each.
- Timestamps of Source_Capabilities, Request, Accept, and PS_RDY events.
- VBUS undershoot and overshoot at each PDO transition.
- Behavior when the electronic load steps from 0 A to the rated current.
- CC1 and CC2 states during negotiation, plus any hard reset or retry.
- Which firmware version and load profile were used, because PD bugs are often version-specific.
That is the list I trust under deadline pressure. The 115 multimeter is still on the bench, but it is no longer the instrument that decides go/no-go.
How to use a multimeter for a USB-PD spot check
Before you assume that a multimeter has no place in this work, it still has one. Here is how to use a multimeter correctly for a USB Power Delivery spot check:
- Put the black lead in COM and the red lead in the V-ohm input.
- Turn the dial to DC volts. USB PD is DC, not AC.
- Measure at the connector that will feed the load, not at the power-supply terminals. Cable losses are real.
- Let the reading settle. If the source just negotiated a new PDO, wait a few seconds before recording the value.
- Do not use a standalone multimeter measurement as the pass/fail for a load transient.
This is simple. The mistake is escalating a DMM spot check into a dynamic protocol test.
Where this advice has limits
If your only question is 'does this power supply still produce 5 V?' then a basic multimeter is enough. This article is not a reason to buy a lab oscilloscope for a weekend electronics project. It is intended for people whose USB Power Delivery problem shows up as a field return, a failed interoperability test, or a last-minute pre-compliance review.
In those situations, uncertainty is the most expensive part of the budget. A scope with a timestamped event list may cost more per hour than a multimeter. But the cheaper approach is not always cheaper when it produces a false pass. The next most expensive thing after a bad measurement is the second test trip, the missed release date, or an embarrassed engineer holding a printed MIN/MAX list that does not explain the failure.
That is why I now reach for a Keysight mixed signal oscilloscope first, keep a 115 multimeter for sanity checks, and save the Keysight 33600A waveform generator for the tests that need repeatable disturbances. The right instrument gives you certainty. For USB Power Delivery, certainty is what you are really paying for.
