USB-C Power Delivery
USB-C Power Delivery: The 240 W Port That Doesn't Fry the Board
Every USB-C plug fits every USB-C socket, and almost nothing else about a cable is guaranteed. The CC resistors that set the first 15 W, the four-message negotiation, the 240 W EPR ladder, the e-marker that gates it, and the EU rule that now makes it law.
Picture a product that charges over USB-C. The charger says 240 W on its side. The cable looks identical to every other cable in the drawer.
The product plugs in, and the charge takes twelve hours instead of forty-seven minutes, or the port negotiates 48 V into a board rated for 20 V, and something smells. Every USB-C plug fits every USB-C socket, and almost nothing else about a cable is guaranteed. The connector is the easy part.
The power is a negotiation, and the negotiation has two layers, a cable that takes part in the decision, and a law that now makes the whole thing mandatory.
USB-C is the connector; USB Power Delivery, PD, is the negotiation protocol that decides what voltage and current a charger (the source) will provide to a device (the sink). This article is about the port that does it right: the resistors that set the first 15 W before any message is sent, the four-message contract that raises the voltage, the 240 W ladder that raises the power by voltage instead of current, the e-marker chip that gates the whole thing, and the EU rule that turned the standard into law.
The two layers: resistors first, messages second
USB-C power works in two layers, and the first layer is older and simpler than the second. Before a single PD message is sent, the CC wire does three jobs: it detects attachment, it sets which side is the power source, and it advertises how much current is available at 5 V. The mechanism is a resistor divider, and it is worth understanding, because it is the layer that decides whether the board gets power at all.
A USB-C receptacle has two configuration channel pins, CC1 and CC2. A cable connects only one of them end to end, which is how each side learns the plug orientation, the pin that sees the other side's resistor is the active CC, and the other CC pin can carry VCONN, the supply that powers the e-marker chip in the cable. The source asserts Rp, a pull-up, on CC, and the sink asserts Rd, a pull-down.
The sink side is simple: Rd is 5.1 kΩ to ground, and a receptacle sink needs its own Rd on CC1 and on CC2. The source side encodes its offer in the strength of the pull-up. The Type-C specification (Table 4-27, Release 2.5) allows three implementations, and the resistor-to-5 V column is the one most boards use: 56 kΩ for default USB power, 22 kΩ for 1.5 A at 5 V, 10 kΩ for 3.0 A at 5 V.
The current-source column is the same three advertisements as currents: 80 µA, 180 µA, 330 µA.
The sink reads the voltage on the CC pin against two thresholds, 0.66 V and 1.23 V, and learns which advertisement the source is making. Below 0.66 V is default, between 0.66 V and 1.23 V is 1.5 A, above 1.23 V is 3.0 A. The default, 5 V at 500 mA for USB 2.0, 900 mA for USB 3.0, is what a sink gets when the source offers nothing more, and it is the safe baseline the whole protocol starts from.
The source only switches on its 5 V supply once it sees Rd, which is why a board with no Rd gets no power at all.
The classic mistake is sharing one resistor. With CC1 and CC2 tied to a single 5.1 kΩ Rd, the cable's CC wire is joined to the far plug's VCONN contact, and in an e-marked cable that contact has Ra, the ~1 kΩ VCONN load of the e-marker, so the source sees 5.1 kΩ in parallel with 1 kΩ, about 0.84 kΩ, which reads as Ra. With Ra in both plugs it looks like an audio accessory.
Either way, VBUS stays off, and the board never powers up. The fix is two resistors, one per CC pin, and it costs nothing, which is why it is a layout-review item, not a debug item.
The four-message contract
Once 5 V is present, the second layer starts, and it is a conversation on the same CC wire. PD messages travel on the CC wire, not on VBUS and not on the USB data pairs, using Bi-phase Mark Coding, a version of Manchester coding, at a nominal bit rate of 300 kbps (270–330 kbps allowed). Each packet starts with a 64-bit preamble of alternating ones and zeros, then a start-of-packet marker, a header, up to seven 32-bit data objects in a normal message, a CRC, and an end-of-packet marker.
The receiver confirms every message with a GoodCRC control message, and a message that is not acknowledged is retried. The source can resend its capabilities every 100–200 milliseconds while it waits for the device's response.
The contract itself is four messages, and the order matters. Source_Capabilities, the charger sends a list of Power Data Objects (PDOs), each 32 bits, whose top two bits give its type: fixed supply, battery, variable supply, or an Augmented PDO (APDO) for PPS and AVS. The first PDO is always 5 V.
Request, the sink evaluates the PDOs and sends a Request, specifying the exact PDO by index and the operating and maximum current it wants. The sink chooses; the source supplies. Accept, the source confirms it can fulfil the request.
PS_RDY, the source ramps its DC-DC converter to the new voltage, and once the voltage is stable it sends PS_RDY, power supply ready, and the sink may now draw the negotiated current. The whole exchange takes a fraction of a second, and it can be repeated at any time, when a laptop battery is full and needs less power, for example.
The specification describes three contract states. A Default Contract is the 5 V power set by Rp at attach. An Explicit Contract is what exists after a successful negotiation.
An Implicit Contract is a short 5 V state after a power role swap. The negotiation from Default to Explicit is an Atomic Message Sequence, and the sink must be safe at 5 V before the PD contract, the power path has to survive the default before it can ask for more. That is the design rule the power piece built: the converter the PD port feeds must be safe at 5 V, and the protection must be rated for the highest voltage the port can negotiate, not the voltage it usually negotiates.
The ladder: SPR to 100 W, EPR to 240 W
The voltage ladder is the heart of the standard, and it has two ranges. Standard Power Range (SPR), from PD 3.0, covers contracts up to 100 W, with fixed rails at 5 V, 9 V, 15 V, and 20 V, at 3 A or 5 A, 20 V at 5 A is the 100 W ceiling. Extended Power Range (EPR), added by PD 3.1 in May 2021, extends the maximum to 240 W with three new fixed rails: 28 V (140 W), 36 V (180 W), and 48 V (240 W), all at 5 A.
The power rises by voltage, not by current, and the reason is the cable loss, which the demo works: loss is I²R, so moving 100 W from 20 V to 48 V cuts the current from 5 A to about 2.08 A and the cable loss by roughly a factor of six.
The ladder is gated by charger size, and the gates are part of the specification's power rules. A source may only advertise 28 V once it exceeds 100 W, may only advertise 36 V at 140 W or more, and may only advertise 48 V at 180 W or more. The power rules also require a source of a given PDP to offer 5 V up to 3 A, 9 V above 15 W, 15 V above 27 W, and 20 V above 45 W, each at PDP ÷ V up to 3 A (5 A for 20 V above 60 W, with a 5 A cable).
The first four voltages, 5, 9, 15, and 20 V, are the normative SPR rails. 12 V is extremely common in the field but is an optional voltage, not a required one, a perfectly compliant charger may go straight from 9 V to 15 V, so a product that designs around 12 V has designed around a rail that might not be there.
EPR is not a separate negotiation. It requires an initial SPR explicit contract to be in place before EPR can even be considered: the sink sends an EPR_Mode (Enter) request over the CC pin, then an EPR_Request for the 48 V / 5 A profile, and the source replies with Accept, transitions VBUS to 48 V, and sends PS_RDY. EPR mode can still be used for a lower-power contract, the important distinction is the capability and the negotiation mode, not a guarantee that every active contract exceeds 100 W.
The current published revision is PD 3.3 Version 1.0, posted 5 November 2026, but no public changelog was available, so the figures in this article describe PD 3.1 and the PD 3.2 restatements, and the article says so rather than assuming what changed in 3.3.
The cable: the third party in the negotiation
The cable takes part in the decision, and it is the part most designs get wrong. The standard cable assembly is rated for 3 A, which caps any charger, 65 W, 100 W, 140 W, at 60 W at 20 V until the cable is electronically marked for 5 A. The Type-C receptacle is rated for 5 A; the standard cable assembly is rated for 3 A.
A 100 W charger through a 3 A cable gives at most 60 W, because the source must see a 5 A e-marker before offering more than 3 A. Every certified USB-C to USB-C cable now carries a 60 W or 240 W logo, plus a data rate unless it is USB 2.0, and the logo is the only thing that tells the user what the cable can do, because the cables all look identical.
A cable above 3 A must contain an electronically marked cable assembly, a small chip, the e-marker, that answers with the cable's current rating, its data capability, and its vendor. The source supplies VCONN on the unused CC pin to power the e-marker chip in the plug, and the chip answers PD messages addressed to the cable, which is how the charger learns whether the cable is rated for 3 A or 5 A, and whether it is EPR capable.
The e-marker reports exactly one of two current ratings, 3 A or 5 A, and the cable classes are three: a plain 3 A cable, a 5 A e-marked cable certified to 100 W, and a 5 A EPR cable certified to 240 W.
EPR entry checks three e-marker fields: 50 V, 5 A, and EPR Capable. A 5 A cable that is not EPR-capable fails EPR entry, and the charger stays at 100 W or less. The original 20 V / 5 A cables are being deprecated, eventually they will no longer be certified by the USB-IF, so the EPR cable is the 5 A cable going forward.
The design rule is simple: never assume a random USB-C cable can carry 5 A, and design the system so it still works, slower, at lower power. The product that needs 100 W and gets 60 W is a product that charges slowly; the product that needs 100 W and tries to draw 5 A through a 3 A cable is a product that fries the cable.
PPS and AVS: the programmable rails
Alongside the fixed rails, a source may advertise Augmented PDOs, the programmable rails. PPS (Programmable Power Supply), from PD 3.0, is optional, and it lets the sink request voltage in 20 mV steps with a current limit in 50 mA steps, over a window of 3.3–21 V (typical advertised windows are 3.3–5.9 V, 3.3–11 V, 3.3–16 V, and 3.3–21 V). It is what allows a phone to drive its battery almost directly and skip a conversion stage, the sink walks the voltage down as the battery fills, instead of the charger holding a fixed rail and the phone burning the difference as heat.
PPS lives inside the Standard Power Range, so it is bounded by 5 A and by 100 W.
AVS (Adjustable Voltage Supply) is the EPR equivalent, in 100 mV steps from 15 V up to the advertised 28 V, 36 V, or 48 V maximum. It is a simpler variable voltage source than PPS, it has no programmable current limit in the way PPS does, and it is intended to provide a steady system input closer to the device's needs, including higher-power EPR devices. The difference that matters for the design: AVS is required in EPR, unlike PPS in SPR. A source that offers EPR must offer AVS; a source that offers SPR may or may not offer PPS. PD 3.2 added SPR AVS as well, 9–15 V on ports over 27 W to 45 W, 9–20 V above 45 W, at 100 mV effective resolution, so the adjustable rail now exists in both ranges, with different floors.
The two are not interchangeable, and the demo shows why. PPS tops out at 21 V and AVS tops out at 48 V, and neither can track a 48 V battery pack that runs from 39 V empty to 54.6 V full. The programmable rail is for the battery that fits inside it; the pack that does not fit needs a converter between the PD port and the battery, and the PD port is the input, not the charger.
The EU rule: the standard is now law
The European Union forced the first half of the convergence. Directive (EU) 2022/2380, signed in Strasbourg on 23 November 2022, amends the Radio Equipment Directive 2014/53/EU and adds Annex Ia. It entered into force on 28 December 2022, and it applies from 28 December 2024 for twelve categories, handheld mobile phones, tablets, digital cameras, headphones, headsets, handheld videogame consoles, portable speakers, e-readers, keyboards, mice, portable navigation systems, and earbuds, and from 28 April 2026 for laptops, defined broadly as portable computers including notebooks, convertibles, and netbooks.
Member states had until 28 December 2023 to adopt national laws.
For devices that charge by cable, Annex Ia requires three things. A USB Type-C receptacle as described in EN IEC 62680-1-3:2021. Charging with compliant cables, meeting the same standard. And USB Power Delivery above the default: in so far as a device charges at voltages higher than 5 V, currents higher than 3 A, or powers higher than 15 W, it must incorporate USB PD as described in EN IEC 62680-1-2:2021, and any additional charging protocol must allow the full functionality of USB PD.
The directive leaves resistor-only charging alone, 5 V at 3 A is 15 W, the ceiling of Type-C Current without any PD message, and requires USB PD only for power above that ceiling. And it requires unbundling: the device must be sold without a charger as an option, which makes a no-charger SKU standard for EU lines. Devices drawing more than 240 W can retain proprietary high-power connectors but must also support USB-C charging up to 240 W.
The Commission has to report by 28 December 2026 on whether the unbundling rule should extend to cables.
The rule matters even for products outside the thirteen categories, because it sets the design baseline: USB-C is the default starting point for any EU-bound portable design, and the charger the product plugs into is a compliant charger. A product that negotiates badly with a compliant charger is a product that fails in the field, whether or not the directive binds it directly.
The build: the port, the path, and the protection
The PD port is a system of four parts, and each one has to be rated for the highest voltage the port can negotiate, not the voltage it usually negotiates.
The sink controller. The PD sink controller is the chip that speaks BMC on the CC wire, reads the Source_Capabilities, sends the Request, and drives the power path. It is the part that decides which PDO to ask for, and it is the part that has to be safe at 5 V before the contract, the controller must not ask for 48 V until the contract exists, and the power path must not pass 48 V until PS_RDY arrives. Pick a controller that explicitly supports the target range, SPR, EPR, PPS, AVS, and validate with EPR-capable sources and cables, not just a random charger.
The power path. The FETs, the eFuse, the over-voltage protection, the discharge circuit, and the connector must all have explicit voltage margin at 48 V. The caps, the TVS, and the OVP have to be rated for the EPR ceiling with margin, 48 V is not 20 V, and a 25 V capacitor on a 48 V rail is a failure waiting for the first EPR negotiation. The discharge circuit matters because the sink must bring VBUS down safely when the contract ends or the cable is unplugged, and the increased voltage levels come with a risk of arcing, especially when plugging in or unplugging a USB-C cable.
The CC lines and the connector. The CC lines are the negotiation wire, and the PCB DFM piece built the rules for them: the 5.1 kΩ Rd on each CC pin, separately, the routing that keeps the CC lines short and away from the switching noise of the converter, and the connector footprint that takes the 24-pin receptacle with its four power and ground contacts that aggregate to 5 A. The connector is rated for 5 A; the cable is the limit, not the connector.
The converter and the battery. The PD port feeds the converter from the power piece, and the converter feeds the battery the BMS from the BMS piece protects. The BMS rules still apply: over-voltage protection at 4.20–4.25 V per cell, under-voltage at 2.8–3.0 V, no charging below 0 °C, and the cell balancing that keeps the pack from losing 5% of its capacity to the weakest cell. The PD port is the input; the BMS is the charger; the converter is the stage between them. The demo shows the gap: the pack needs 54.6 V at full charge, the EPR ceiling is 48 V, so the converter must boost, and the AVS rail, which tops out at 48 V, cannot track the pack directly.
Example Labs is a made-up company we use for worked examples.
DEMO
The pack is the one the bench arm runs on, from Articles #20 and #23: the shoulder drive needs 199 W mechanical, 221 W electrical, from a 48 V bus. The pack is 13S1P NMC, 3.5 Ah per cell, 48.1 V nominal, 54.6 V at full charge, 39.0 V empty, 168 Wh. The PD port charges it, and the arithmetic was run in Python before writing.
Step 1, the charge time, by contract. The same pack, the same charger, four different contracts. At 48 V × 5 A EPR (240 W), the ideal charge time is 168 Wh ÷ 240 W = 42 minutes; at 90% converter efficiency (stated), 47 minutes. At 20 V × 5 A SPR (100 W), the ideal is 101 minutes, realistic 112 minutes. At 20 V × 3 A, the plain 3 A cable, the practical ceiling without an e-marker, the ideal is 168 minutes, realistic 187 minutes. At 5 V × 3 A with no PD contract at all, the ideal is 672 minutes, realistic 747 minutes, 12.4 hours. The cable decides the charge time: 47 minutes against 12.4 hours, from the same charger and the same pack, because the contract is what the cable allows.
Step 2, why the power rises by voltage, not current. The cable loss is I²R, and the demo takes a cable loop resistance of 0.1 Ω (stated). At 100 W delivered at 20 V, the current is 5 A and the loss is 25 × 0.1 = 2.5 W, 2.5% of the delivered power. At 100 W delivered at 48 V, the current is 2.08 A and the loss is 4.33 × 0.1 = 0.43 W, 0.43%. At 240 W delivered at 48 V, the current is 5 A and the loss is 2.5 W, 1.04%. Moving 100 W from 20 V to 48 V cuts the current by a factor of 2.4 and the cable loss by a factor of 5.8, roughly six, which is the figure the guides give. The power rises by voltage because the loss rises with the square of the current, and the current is the thing that melts cables.
Step 3, the CC voltages. The sink presents Rd = 5.1 kΩ on CC1 and CC2, and the source's Rp sets the default. The sink reads against 0.66 V and 1.23 V: below 0.66 V is default, between 0.66 V and 1.23 V is 1.5 A, above 1.23 V is 3.0 A. The three numbers land in the three bands, with margin, which is why the ±5% tolerance on the 22 kΩ and 10 kΩ resistors matters, and why the ±20% tolerance on the 56 kΩ default resistor is acceptable.
Step 4, the voltage gap. The pack's full-charge voltage is 54.6 V and the EPR ceiling is 48 V, a gap of 6.6 V. The PD port cannot charge the pack directly, and neither programmable rail can track it: PPS tops out at 21 V, AVS tops out at 48 V, and the pack runs from 39.0 V empty to 54.6 V full. The on-board charger must boost 48 V to 54.6 V, and the BMS from #24 is the charger, the PD port is the input, the converter is the stage, and the BMS is the part that decides when the pack is full. A design that connects the PD port straight to the pack is a design that undercharges it by 6.6 V, which is 12% of its capacity, and a design that asks the PD port for 54.6 V is a design that asks for a voltage the standard does not offer.
Step 5, the EPR gate. The source reads the cable's e-marker before it offers EPR, and it must see 50 V, 5 A, EPR Capable. A 5 A cable that is not EPR-capable fails EPR entry, and the charger stays at 100 W or less, the 112-minute charge instead of the 47-minute one. A plain 3 A cable never offers more than 3 A, so the sink falls back to 20 V × 3 A = 60 W and the 187-minute charge. The e-marker is the gate, and the gate is the reason the cable matters more than the charger.
The traps
Designing around 12 V. The 12 V rail is extremely common in the field, and it is an optional voltage. A perfectly compliant charger may go straight from 9 V to 15 V, and a product that needs 12 V has to ask for 15 V and convert down, or ask for 9 V and convert up. The rail that might not be there is not a rail.
Assuming the cable is 5 A. The cable that came in the box is 3 A unless its e-marker says otherwise, and the charger will not offer more than 3 A until it has read the marker. A product that assumes 5 A and draws it through a 3 A cable is a product that fries the cable, and the cable is the cheapest part in the system, which is why it is the one that fails first.
Sharing one Rd across CC1 and CC2. The classic mistake: one 5.1 kΩ resistor shared by both CC pins, and with an e-marked cable the cable's Ra appears in parallel, the source no longer sees a plain sink, and VBUS stays off. Two resistors, one per pin, and the layout review checks it.
Requesting EPR without the e-marker check. The sink that enters EPR mode without the source having read the cable's e-marker is a sink that has skipped the safety check the standard exists to enforce. The source owns the check, and the sink owns the fallback: if the e-marker does not say 50 V, 5 A, EPR Capable, the sink takes the SPR contract and charges slower.
Trusting the charger's label wattage. The label says 240 W, and the cable says 3 A, and the delivered power is 60 W. The wattage on the label is the charger's capability; the contract is the charger's capability, the cable's rating, and the sink's request, and the delivered power is the smallest of the three. The product that reports "charging at 240 W" because the charger says so is a product that lies to its user.
Using PPS as a battery charger without the current limit. PPS lets the sink set the current in 50 mA steps, and the current limit is the part that makes it a charger, without it, the sink is asking for a voltage and hoping the battery takes what it needs. AVS has no programmable current limit at all, which is why it is a system input, not a battery charger.
The decision test: five questions
A team designing a USB-C PD port can answer five questions in an afternoon. If it cannot answer three of them, it has not designed a port, it has drawn a connector.
- What is the highest voltage the port can negotiate, and is every part on the VBUS path rated for it with margin? If the caps, the FETs, the TVS, and the OVP are rated for 20 V and the port can negotiate 48 V, the first EPR charger that plugs in is the failure.
- What does the sink do at 5 V, before any PD message? If the answer is "it draws its full load", the board will brown out on the default contract, and the charger will see a fault instead of a sink.
- What cable does the design assume, and what happens with the wrong cable? If the design assumes a 5 A EPR cable and the user plugs in a plain 3 A cable, the charge time triples, and the product has to still work, slower, not fail.
- What are the CC resistor values, and is there one Rd per CC pin? If the answer is "one shared resistor", the board will not power up with an e-marked cable, and the fix is a layout change, not a firmware change.
- What does the EU rule require of this product, and when? If the product is in one of the thirteen categories and it charges above 15 W, it must have a USB-C receptacle and USB PD, and the rule has applied since 28 December 2024, or 28 April 2026 for laptops.
Objections
"Our product only needs 15 W, we don't need PD." Correct, and the directive agrees: 5 V at 3 A is 15 W, the ceiling of Type-C Current without any PD message, and the EU rule leaves resistor-only charging alone. The product needs two 5.1 kΩ Rd resistors and nothing else, but it needs them, separately, on CC1 and CC2, or it gets no power at all.
"Our charger says 240 W, so the product charges at 240 W." The charger's label is its capability. The contract is the charger's capability, the cable's rating, and the sink's request, and the delivered power is the smallest of the three. A 240 W charger through a 3 A cable gives 60 W, and a 240 W charger through a 5 A cable that is not EPR-capable gives 100 W. The label is not the contract.
"We'll just use a barrel jack, it's simpler." It is simpler, and for a product outside the thirteen categories it is still legal. But the charger the customer already owns is a USB-C PD charger, the cable they already own is a USB-C cable, and the product that needs its own barrel jack is a product that needs its own charger in the box, which the EU rule is unbundling, and which the customer will lose. The barrel jack is the simpler port and the harder product.
"PPS can charge the battery directly, so we don't need a charger IC." PPS can drive a battery almost directly, because the sink walks the voltage in 20 mV steps and sets the current in 50 mA steps, but the current limit is the charger, and without it the sink is asking for a voltage and hoping. PPS is a programmable supply, not a battery charger, and the BMS still has to decide when the pack is full.
"EPR is only for laptops." EPR is for anything that needs more than 100 W, and the Example Labs bench pack is the example: 168 Wh, 48 V, and a 47-minute charge against 12.4 hours without PD. The laptop is the most visible EPR product; the bench, the dock, and the tool battery are the ones that need it next.
Sources
- Safeguard, "USB-C Power Delivery: PD 3.2, EPR, Negotiation and Charger Authentication"
What we used
Numbers from USB PD Revision 3.2 Version 1.2, USB Type-C Release 2.5, and the USB Type-C Authentication and Security Foundation specifications; CC wire first (Rp vs 5.1 kΩ Rd sets default / 1.5 A / 3.0 A at 5 V); four-message contract (Source_Capabilities, Request, Accept, PS_RDY, each acknowledged with GoodCRC, ~300 kbps on CC); EPR adds 28, 36, 48 V for 140, 180, 240 W, and the charger must first read the cable's e-marker and see 50 V, 5 A, EPR Capable; PD 3.2 v1.2 cover 19 August 2026, posted 14 September 2026 (its own revision table lists January 2026); Type-C Release 2.5 posted 8 April 2026; EU requires USB-C and USB PD above 15 W, phones and eleven other categories since 28 December 2024, laptops since 28 April 2026; Ra ≈ 1 kΩ VCONN load of an e-marker; Linux exposes the negotiation at /sys/class/typec and /sys/class/usb_power_delivery
- Plugable, "What Is 240W USB Extended Power Range (EPR)"
What we used
PD 3.1 enables up to 240 W; SPR relabelled (the previous 100 W range); EPR 28 V, 36 V, 48 V at 5 A; AVS enables fine-tuning between 15 V and 48 V in 100 mV steps; AVS required in EPR, unlike PPS in SPR; EPR requires an initial SPR explicit contract; original 20 V/5 A cables being deprecated
- Diodes, "USB PD EPR and SPR: Unraveling the Differences"
What we used
SPR (PD 3.0) max 100 W, fixed 5/9/15/20 V at 3 A or 5 A; EPR (PD 3.1) adds 28/36/48 V at 5 A, max 240 W; PPS 20 mV steps, primarily for battery charging below 100 W; AVS fine regulation in 100 mV steps between 15 V and 48 V
- AgentCalc, "USB-C Power Delivery Calculator"
What we used
Fixed PDOs 5/9/15/20 V normative SPR, 28/36/48 V EPR; 12 V optional, not required; source may only advertise 28 V above 100 W, 36 V at 140 W+, 48 V at 180 W+; PPS 20 mV voltage steps and 50 mA current steps, windows 3.3–5.9 / 3.3–11 / 3.3–16 / 3.3–21 V, bounded by 5 A and 100 W; moving 100 W from 20 V to 48 V cuts current from 5 A to ~2.08 A and cable loss by ~6×; e-marker reports exactly one of 3 A or 5 A; three cable classes (plain 3 A, 5 A e-marked to 100 W, 5 A EPR to 240 W)
- RSQ, "USB-C Cables: What 240W, 20Gbps and a 0.8 m Limit Actually Get You"
What we used
Standard cable rated 3 A, 60 W at 20 V the ceiling without an e-marker; 5 A cable must be electronically marked; 240 W needs a 240 W-rated EPR cable, charger, and device; Type-C receptacle rated 5 A, standard cable assembly rated 3 A; current published revision PD 3.3 (posted November 2026), Type-C Release 2.5 (posted April 2026)
- Mbedded.ninja, "USB Charging and Power Delivery"
What we used
PD 3.1 released May 2021, max power 100 W → 240 W; EPR fixed 28 V (140 W), 36 V (180 W), 48 V (240 W); AVS 100 mV resolution between 15 V and max supported voltage; power direction flexible; EU mandated Type-C from 2024, laptops from 2026
- Eronix, "USB Power Delivery and iPhone: How the Handshake Works"
What we used
Rp pull-up vs Rd pull-down; Rp value carries the first message (default, 1.5 A, 3 A at 5 V); four messages with GoodCRC receipts; fixed PDO table (5 V / 3.0 A / 15 W; 9 V / 2.22 A / 20 W); EPR_Mode (Enter) then EPR_Request for 48 V/5 A, Accept, transition to 48 V, PS_RDY
- Salivity, "How USB Power Delivery Negotiates 240W Over CC Pins"
What we used
BMC single-wire half-duplex at 300 kbps; packet structure (preamble, SOP, payload, 32-bit CRC, EOP); Source_Capabilities → GoodCRC → Request → Accept → PS_RDY; EPR steps (EPR_Mode Enter, EPR_Request, Accept, transition, PS_RDY)
- Acutena, "USB PD 3.1 Protocol Testing Guide"
What we used
BMC base rate 300 kbaud; message structure (preamble, SOP, header, data objects, CRC-32); typical negotiation (Source_Capabilities → Request → Accept → PS_RDY); EPR PDOs listed separately and require the sink to explicitly enter EPR mode
- Nodeloop, "USB-C PD Hardware Design Guide: CC, PPS, EPR and AVS" (verified July 2026)
What we used
Current specs Type-C 2.5 and PD 3.2 v1.2; Rp values (56 kΩ default, 22 kΩ 1.5 A, 10 kΩ 3.0 A, all to 5 V); Rd 5.1 kΩ to GND, one per CC pin, shared-resistor mistake (e-marker's Ra in parallel, source no longer sees a plain sink, VBUS may not enable); EPR AVS 15 V to max in 100 mV steps; component ratings need explicit voltage margin; EPR cable electronically marked for 5 A and 50 V, reports EPR Mode capability
- GeekyChris, "USB-C and Power Delivery"
What we used
CC voltage thresholds (sink reads against 0.66 V and 1.23 V); Ra below vRa (0.2/0.4/0.8 V), Rd up to vOPEN (1.6/2.6 V); shared-resistor mistake (5.1 kΩ ∥ 1 kΩ ≈ 0.84 kΩ reads as Ra); power rules (PD §10: 5 V up to 3 A, 9 V above 15 W, 15 V above 27 W, 20 V above 45 W); sources Type-C Release 2.3 and PD Revision 3.2
- EDN, "USB Type C Power Delivery and Hobby Power"
What we used
24 contacts, four power and ground contacts aggregate to 5 A; 5.1 kΩ ±10% the only acceptable Rd for 1.5 A/3.0 A at 5 V; 56 kΩ ±20% the recommended DFP Rp for default USB power
- PCBWay, "USB Type-C Pinout: Why Are 5.1KΩ Pull-Down Resistors Needed?"
What we used
CC1 and CC2 each require a separate 5.1 kΩ pull-down; Vcc = Vbus × (Rd / (Rp + Rd)); Rp can also be a current source
- Informatique Phone, "USB-C PD: Fixed PDO, PPS & E-Marker Cables"
What we used
PD 3.1 enables 240 W; SPR 3 A at 5/9/15/20 V covers 15–60 W, 5 A at 20 V covers >60–100 W (5 A cable required), EPR 5 A at 28/36/48 V covers >100–240 W; PPS 20 mV/step, 3.3–21 V, optional, directly charges a battery by bypassing the battery charger; AVS 100 mV steps, 15–48 V (one source: 9–48 V), no constant current; e-marker required for >3 A, >20 V, and all EPR
- Done.land, "USB Triggering"
What we used
PPS optional in PD 3.0 and 3.1; AVS 15–48 V, 100 mV steps, no constant current, PD 3.1+ (optional); e-marker required for >3 A or >20 V and all EPR
- Hackaday (via SmartHomeOnline), "Powering Up With USB"
What we used
R3.1 increases voltage to 48 V while keeping 5 A, because current is the defining factor in whether cables melt; PPS 3.3 V to SPR voltage + 1–2 V in 20 mV steps; AVS from 15 V up to any EPR voltage in 100 mV steps; increased voltages come with arcing risk on plug/unplug
- MEF Mobile, "USB-C Power Delivery Explained"
What we used
PD 3.3 v1.0 posted 5 November 2026, no public changelog; SPR AVS 9–15 V on ports over 27–45 W, 9–20 V above 45 W, 100 mV effective resolution (added in PD 3.2); EPR AVS 15 V to charger maximum, 100 mV resolution, no programmable current limit; PPS 20 mV steps, optional; figures describe PD 3.1 and PD 3.2 restatements
- EUR-Lex, Commission Notice 2024/C 2997 (guidance for the Common Charger Directive)
What we used
The harmonised charging solution is the USB-C receptacle and the harmonised charging technology (USB-C power supply options up to 15 W and USB PD above 15 W); rules apply to handheld mobile phones, tablets, digital cameras, headphones, headsets, handheld videogame consoles, portable speakers, e-readers, keyboards, mice, portable navigation systems, and earbuds as of 28 December 2024, and to laptops as of 28 April 2026
- Altium, "After the EU Laptop Charger Deadline, USB-C Becomes the Design Baseline"
What we used
Laptops added 28 April 2026, completing the phase-in; USB PD required for covered devices charging above 15 W; devices drawing more than 240 W can retain proprietary connectors but must also support USB-C charging up to 240 W; Commission to report by 28 December 2026 on extending unbundling to cables
- Example Labs figures from the cluster: the 48 V bench pack (13S1P NMC, 3.5 Ah/cell, 48.1 V nominal, 54.6 V full charge, 168 Wh
What we used
#20/#23); the BMS rules (OVP 4.20–4.25 V/cell, UVP 2.8–3.0 V, no charging below 0 °C; 5% capacity spread = 175 mAh, 8.4 Wh of 168 Wh lost without balancing, #24); the field sensor hub (100 × 60 × 40 mm, CR2032 230 mAh, #6/#7); the STM32H7 (512 kB SRAM / 2 MB flash, 48 MHz, #9); the BLE radio at 2.4 GHz (#12)