← Back to Home

How to Read USB Power Meter Specs: Five Numbers to Doubt

USB power meterUSB testerbuying guide

The charger box says 140 W. The battery icon in your laptop corner says you are taking 60 W. To find out which link in the chain is losing the rest, you put a USB power meter - also sold as a USB tester or USB multimeter - between the charger and the cable.

The listings for these things all look the same: a four-digit voltage and current figure, a "supports PD3.0/QC4+" line, and a colour screen. The trouble is that about half of those numbers only hold under conditions the vendor never puts in the title.

On 2026-10-31 I pulled the FNIRSI FNB58 product page, the POWER-Z KM003C product page, the Klein Tools ET920 instruction-manual PDF and its official product page, plus three Amazon listings, and checked them against each other line by line. This article is not about which one to buy. It is about the five numbers on the spec sheet that each need a condition attached before they mean anything.

TL;DR: three shortcuts. First, if you want to read the PD messages themselves and will pay for protocol analysis, take the ChargerLAB POWER-Z KM003C: ChargerLAB's own page says it "integrates ARM Cortex-M4 processor and 20 bit ADC" and calls it a "USB PD protocol analyzer", rated "up to 50V 6A (240W in PD)". Second, if you want ripple and cable resistance on top of volts, amps and watts, take the FNIRSI FNB58: its official table gives 0.00001 V resolution and "up to 4 M sps" ripple sampling, at the cost of a 140 W power ceiling instead of the 28 V x 7 A envelope. Third, if you only want to know whether a port is actually charging, take the Klein Tools ET920, whose manual states the boundary in one line: "This meter will NOT display current below 50 milliamperes".

Number one: volts times amps is not the power this meter can measure

This is the easiest one to trip over, because the three meters give three different ratios.

FNIRSI's own specification table for the FNB58 reads: Monitor Voltage 4-28 V, Monitor Current 0-7 A, Monitor Power 0-140 W. Twenty-eight times seven is 196 W, and the official power field stops at 140 W - 71.4% of the envelope.

POWER-Z's own key-features line for the KM003C reads: "It can support up to 50V 6A (240W in PD)". Fifty times six is 300 W; the company circles 240 W - 80%.

The YOJOCK KWS-2303C listing says 4-30 V and 0-12 A, and puts 360 W right in the title. Thirty times twelve is exactly 360 W - a 100% envelope.

The Klein Tools ET920 page, meanwhile, never states a power ceiling at all. It gives Voltage 3 to 20V DC and Current 0.05 to 5A (Type C). Twenty times five is 100 W, but that is your arithmetic, not Klein's claim.

Put 71.4%, 80% and 100% on one line and the conclusion is hard to avoid: the voltage range and the current range are independent ceilings, and power is a third, separate limit, usually set by what the internal sense resistor and the pass MOSFET can dissipate. Before you trust the "240 W" or "360 W" in a title, check whether the spec table lists power as its own row at all.

It is the same ruler this site used on bench power supplies, where the OWON official table for six models ran from 1.00 all the way to 3.00 (60 V 10 A with only 200 W delivered). See Bench Power Supply Buying Guide: Rated Watts Is Not V x A.

Number two: the extra digits of resolution are not covered by the accuracy

FNIRSI's table gives a resolution of 0.00001 for voltage, current and power, and the copy says "The highest six-digit display of voltage, current and power". Six digits looks impressive.

Resolution and accuracy are different rows. Klein's manual is far more honest about it:

On the 20 V range, +/- 1% is +/- 0.2 V, and 0.2 V is twenty thousand times 0.00001 V. Of those six decimals, only the first couple are actually guaranteed. The digits resolution adds are not inside the accuracy figure. (Incidentally, "+/- 2 Bytes" is verbatim from the official manual, in both places it appears - almost certainly a slip for "digits".)

The YOJOCK KWS-2303C listing says "+/- (1%+5)" and never explains what the 5 counts - display counts, last digit, or something else. Without that one clause you cannot turn it into an absolute error at all.

The line worth copying out is this one from the Klein warnings: "This meter will NOT display current below 50 milliamperes, even if it is allowing this to pass through. The display will show zero current." In other words, point it at a sleeping phone or a standby drain and it will give you an honest, useless 0. That is design, not failure.

This is the same lesson as Digital Caliper Specs: Resolution, E_MPE/S_MPE and IP67 and Clamp Meters: Why a 400 A Range Cannot Read 1 A: resolution is what the display can show, permitted error is what the instrument can measure.

Number three: there are two sampling rates, forty thousand times apart

FNIRSI gives both in the same paragraph: "Support low-speed waveform drawing, 2 sps-100 sps sampling rate" and "Support high-speed ripple drawing, up to 4 M sps sampling rate".

Four million divided by one hundred is 40,000x. These are not two ways of saying one thing; they are two different paths. The slow one draws the charging curve and records how voltage and current move over minutes. The 4 Msps one exists to look at switching-supply ripple and noise. Advertising "4M sampling" off the ripple path, or "100 sps" off the slow path, is telling half the story either way.

The KM003C has a worse version of this problem: the English POWER-Z product page gives no sampling rate at all, while an official post on the Chinese site power-z.cn says "最高8M高频纹波检测" (up to 8M high-frequency ripple detection). Same model, two official sites in two languages, two different conventions - so buy against the description on the page you are actually ordering from.

Number four: "supports" splits into compatible, recognize, trigger and analyze

Every listing loves the line "supports PD3.0". Across four meters that sentence means four different things:

All four get sold as "supports". If you only need to confirm that charging is happening, the first tier is enough. If you need to know why a power bank will not give you 20 V, only the fourth can answer.

Number five: what the meter itself consumes, and what breaks it

The meter is also a load. Klein's page says the ET920 "requires NO batteries as it draws power from the power source to function", and the manual carries an easy-to-miss note: "USB-A and USB-C sides CANNOT be used simultaneously. The first device connected... will be the monitored source."

The same manual carries a hard boundary in the warnings: "Voltages above 24V DC will damage the product. The display will read 'OVERLOAD' and the output port will be disabled." Note that the range stops at 20 V while the damage threshold is written as 24 V - the four volts in between are headroom, not usable range.

FNIRSI writes a different kind of boundary into its usage notes: operating temperature -5 to 40 C, relative humidity not above 80%, and "When the screen on the LCD does not light up, it indicates that the voltage is too low".

One official contradiction worth recording: Klein's product page says "Continuously monitor up to 1000 hours", while the specification block in the manual says "Elapsed Time: Up to 999 hours, 59 min, 59 sec". The two differ by less than a thousandth, but they are not the same number, and both came from Klein.

Five meters, five real weaknesses

ChargerLAB POWER-Z KM003C

Listed at $109.99 on power-z.com as of this writing. A 1.54-inch display with gravity rotation, a CNC aluminium shell, two USB-C ports for input and output, plus a third USB-C HID port for PC connection or auxiliary power. Officially supports PD3.2, PD3.1, QC, SCP, FCP, AFC, VFCP, SVOOC, MTK, EPR and UFCS.

Real weaknesses: POWER-Z itself writes "Please go to our official website to get the manual and PC software" - both the manual and the PC software are downloads you have to go and find, not things in the box. The page states support for English and Chinese only. The English site gives no sampling rate while the Chinese site's blog says 8M, so the two official language versions disagree. And the official 50 V x 6 A = 300 W envelope is only warranted for the "240W in PD" figure.

Check current stock and price on the POWER-Z KM003C listing.

FNIRSI FNB58

Listed at $98.99 on fnirsi.com as of this writing. A 2.0-inch TFT with gravity sensing, USB-A, Micro-USB and Type-C connectors, an external 16-bit ADC and a PD protocol physical chip. It reads E-Marker chips and DASH cable data, measures cable resistance by the differential-pressure method, and logs onboard temperature and startup time. Supports QC2.0 to PD2.0 conversion.

Real weaknesses: the 2-year warranty is limited to direct purchases - the official wording is "2-year warranty (only available for orders purchased directly from fnirsi.com)", so buying anywhere else does not get it. There is no printed manual; firmware and the app come via a QR code on the box. The 140 W power ceiling is well inside the 28 V x 7 A envelope. Operating range is -5 to 40 C at no more than 80% humidity.

Check it on the FNIRSI FNB58 listing.

Klein Tools ET920

A US tool-brand meter: 2 m drop protection, IP20, 48 g, 75 x 50 x 15 mm, no batteries needed. The official page lists voltage, current, capacity, energy and resistance, with ceilings of 99,999 mAh, 999 Wh and 400 ohms, and storage for 10 readings.

Real weaknesses: anything below 50 mA reads as 0, in the official warning's own words. USB-A and USB-C cannot be used at the same time. Above 24 V DC it takes damage and the output port is disabled outright. The product page itself labels resistance "Resistance (Calculated)" - derived, not measured. No power ceiling is published. And the page and the manual disagree on run time (1000 hours versus 999:59:59).

Check it on the Klein Tools ET920 listing.

YOJOCK KWS-2303C

4-30 V, 0-12 A, with 360 W in the title - the only one of the five where voltage, current and power line up exactly. It has a 12 A high-current terminal, supports PD2.0/3.0, QC2.0/3.0, FCP, SCP, AFC, PE, DASH VOOC and Super VOOC, sets its curve window to 5 minutes, 15 minutes, 60 minutes or 2 hours, and retains data through power loss.

Real weaknesses: accuracy is stated as "+/- (1%+5)" with no explanation anywhere of what the 5 counts or how many display counts it corresponds to, so it cannot be converted to an absolute error. The curve record tops out at 2 hours, short for ageing runs. And there is no brand specification page at all, so every number here rests on the listing alone with no second source.

Check it on the YOJOCK KWS-2303C listing.

AiLi UM25C

A 1.44-inch colour screen with Bluetooth. The listing says it works with a phone app or PC software, supports 14 languages, measures voltage, current, power, capacity and energy, measures wire resistance by the differential-pressure method, and uses gold-plated USB contacts. It sits at the cheap end of the five.

Real weaknesses: the listing reads "PC Software Support PC Win 7 and Above, Phone Only Support Android (Android 5.0 and Above)" - the PC software is Windows-only and the phone app is Android-only, so an iPhone owner gets a screen with no app. The title says "USB 2.0 Type-C", so it tops out at USB 2.0 speeds and is not the tool for high-speed data cables or USB4. And its protocol recognition works off D+/D- voltages, so PD negotiation is invisible to it.

Check it on the AiLi UM25C listing.

FAQ

Q: I want to know why my laptop only takes 60 W. Which one do I buy?

A: That needs PDO reading - which voltages the charger advertises and which one your device asked for. Of these five, only the KM003C is officially listed with PDO test and a "USB PD protocol analyzer". The other four will tell you "19.8 V, 3.1 A right now" and stop there.

Q: I only want to measure how many mAh a power bank actually delivers. Do I need the expensive one?

A: No. Capacity is an energy integral; you only need the meter to stay inside its current range. Two things matter: keep the discharge current under the meter's ceiling, and watch the current floor - the ET920 officially shows 0 below 50 mA, so the low-current tail gets dropped.

Q: Is "supports PD3.0" the same claim everywhere?

A: No. Section four lists four tiers: compatible (it will not break), recognize (guessing from D+/D-), trigger (actively requesting a voltage), and analyze (reading the CC messages). Decide which tier you need before you pay.

Q: Does 4 Msps sampling actually matter?

A: It matters for ripple and nothing else. For a charging curve you want the 2-100 sps path; 4 Msps is there for switching-supply noise. Both numbers sit in the same official paragraph, and nobody tells you they are not the same path.

Q: Why do I see different numbers for the same meter on different pages?

A: Because the vendors contradict themselves. This article records three: Klein's 1000 hours versus the manual's 999:59:59, POWER-Z giving no sampling rate in English while the Chinese site says 8M, and FNIRSI's 140 W power row against its own 28 V x 7 A envelope. Buy against the page you are ordering from.

Where this article stops

👉 Join Xiaomi MiMo Platform: Leading AI model platform with cost-effective inference

👉 Join Aliyun AI: Top AI products with exclusive coupons for business innovation

📌 This article was AI-assisted generated and human-reviewed | TechPassive — An AI-driven content testing site focused on real tool reviews

🔗 Recommended Tools

These are carefully selected tools. Using our affiliate links supports us to keep producing quality content:

As an Amazon Associate, I earn from qualifying purchases at no extra cost to you.

☁️ DigitalOcean Cloud ⚡ Vultr VPS 🤖 QoderWork CN (Refer & Earn) ☁️ Aliyun AI Products 📚 WordPress Books 🧠 AI & LLM Books 🏗️ System Design Books 🐳 Docker & K8s Books 🐧 Linux Admin Books 🐍 Python Books 🗄️ Database Books 🚀 DevOps & SRE Books 🤖 Xiaomi MiMo Platform
← Back to Home