News | April 21, 1999

Meeting USB Over-Current Protection Requirements - Part 2

Reliable, resettable over-current protection is a must in the plug-and-play and hot-plug environments.

By Paul Wiener, Raychem Corp.


•Design Considerations
• Other Overcurrent Protection Technology Solutions

Design Considerations

•Host/Self-Powered Hub
•Bus-Powered Hub
•Voltage Drop
•Voltage Droop

Host/Self-Powered Hub

Output filter capacitor

A capacitor should be used on the Vbus output of each downstream port. Tests were carried out to determine the effects of "capacitor output" geometry versus "ferrite output" with respect to meeting voltage droop and EMI requirements. Far better droop performance was obtained with the output capacitor (150 µF) downstream of the ferrite bead due to the lowered effective impedance as seen by the output connector. The "ferrite output" configuration, with the ferrite bead "downstream" of the output capacitor, is more commonly seen in EMI protection geometry, but measurements indicated that EMI protection was excellent in either geometry.

Ferrite beads

A ferrite bead in series with the Vbus pin is usually necessary to lower EMI radiation down the USB cable. The series resistance of the bead and its RF impedance are the important criteria. A "bead-on-a-wire" configuration is selected primarily for its low series DC resistance. These devices are available in both through-hole and SMD (bead-on-a-strap) configurations. Any ferrite bead with 50 ohms or more impedance at 100 mHz will work. The "Capacitor output" geometry, where the output capacitor follows the ferrite bead, is the preferred geometry for excellent droop characteristics at the downstream port.

Overcurrent protection

The USB specification 1.1 requires overcurrent protection in either an individual or ganged protection scheme. A UL-recognized safety device facilitates agency testing and provides high reliability. Along with meeting safety requirements, the protection device should protect the equipment from damage (for example, protect PCB traces from burning). In addition, the protection device must not nuisance trip (such as during a hot-plug event). Using PolySwitch devices in an individual port protection scheme provides an optimal design for a downstream power connection. When a fault occurs on a port and that port's PolySwitch device trips, the adjacent ports remain functional. Individual port protection also allows the designer to select a smaller and faster-acting device.

Bus-Powered Hub

Power switch

The hub controller supplies a software-controlled on/off signal from the host. Power to downstream ports must have on/off signal-switching capability to power off all downstream ports when the hub comes out of power-up, or when it receives a reset on its root port. A low RdsOn MOSFET must be selected for power switching. Selection of a MOSFET should include a 50-milliohm-or-less RdsOn resistance, logic-level gate, and a 3-A or higher current rating. A P-Channel MOSFET is usually employed as the initial logic level of most control microprocessors' output pins and is weakly pulled high. The P-Channel MOSFET is turned on by driving its gate low; the default state of the MOSFET should be off. A weak external pull-up may be used if desired.

Overcurrent protection

As with any powered port, short circuits and equipment damage can occur and should be a concern. During a short-circuit event, current can get high enough to cause damage to the Bus-Powered Hub. PolySwitch devices are appropriate in individual or ganged port protection schemes and provide Bus-Powered Hub designs protection at the port to protect the MOSFET during a fault.

Soft start

Section of the USB specification requires inrush-current limiting when switching downstream power. This applies to Bus-Powered Hubs, where the Vbus voltage could droop significantly if measures are not taken to "soft start" the downstream power. A simple resistor capacitor (RC) network will provide an inexpensive solution to limiting inrush-current. Figure 7 shows Vbus rising sharply without soft start, whereas Figure 8 shows a much more gradual rise with soft start implemented.

Pulse-width modulation of the PWR_ON- pin may be used in place of the RC network to provide the necessary soft start. The control microprocessor is programmed to drive the pin to zero in a stepwise fashion over a few milliseconds. Figure 9 illustrates the behavior of the pulses.


In order to meet the voltage drop, droop, and EMI requirements in Specification 1.1, careful PCB layout is necessary. The following guidelines must be considered:

  • Keep all Vbus traces as short as possible and use at least 50-mil, 1-oz. copper for all Vbus traces.
  • Avoid vias as much as possible. If vias are necessary, make them as large as feasible.
  • Place the output capacitor and ferrite beads as close to the USB connectors as possible.
  • If using a Vbus switch, place it close to the USB connectors.
  • Use a separate ground and power planes if possible.
  • If using separate planes, separate the power plane into switched (downstream) and unswitched (logic or upstream) sections.

Voltage Drop

The USB specification states a minimum port-output voltage (VOUT) in two locations on the bus, 4.75 V out of a Self-Powered Hub Port and 4.40 V out of a Bus-Powered Hub Port. Careful PCB design layout and propor device selection can easily achieve these requirements.

Voltage Droop

The following voltage droop analysis is divided into two parts, Standard USB voltage droop and short- circuit voltage droop.

Standard USB voltage droop

Voltage droop occurs during a hot-plug event as the result of the connection of a peripheral and its uncharged-input bulk capacitance to a USB port. For a few tens of µs the host must supply high current into the peripheral until its bulk capacitance is charged to Vbus, creating a capacitive divider between the host's bulk capacitance and that of the just-plugged-in USB peripheral. A standard test has been developed by the USB Implementor's Forum (USB-IF) to ensure compliance with the USB specification's limitation of voltage droop to 0.330 V maximum on a hub. Six different PolySwitch devices were tested. Worst case results for each device are shown in Table 1.

Table 1. Standard Voltage Droop test results.

PolySwitch Device

Vdroop (V)













The test data in Table 1 demonstrate voltage droop values lower than the USB specification maximum (0.330 V). Designs using PolySwitch devices therefore meet the USB voltage droop requirement.

Short-circuit voltage droop

If one USB hub downstream port is accidentally short-circuited, it is desirable that the other downstream ports keep functioning normally. While this behavior is not required by USB specification 1.1, it is achievable in a reliable, economical manner using resettable fuses.

When a port is short-circuited, a very large current load is imposed on the power supply that is providing the bus power. This large load causes the remaining ports to experience a transitory droop in bus supply voltage until the overcurrent device on the shorted port reacts. Once the device trips, the remaining ports are isolated from the shorted port. Several design configurations were tested for port protection and short-circuit power supply droop. Four PolySwitch devices commonly used for USB port protection (miniSMDC075, miniSMDC110, RUSB120, and RUSB075) were tested. All four devices meet the voltage droop requirements of >4.00 V. Worst case results are shown in Table 2.

Table 2. Short-Circuit Voltage Droop Test Results.

PolySwitch device

Vo (min.)




45 ms



50 ms



20 ms



45 ms


· Vo (min.) is the minimum droop voltage measured on one port when the second port is shorted.

· t-Trip is the time it takes the PolySwitch device to go into the high-impedance state.

The test data shows that the miniSMDC075 device is the best-performing device. Considering voltage drop requirements in addition to the voltage droop requirements, the miniSMDC110 device is the recom-mended device and meets both requirements. For through-hole designs, the PolySwitch RUSB120 device meets the design criteria and is the recommended device. In summary, all five devices meet the design criteria of >4.00 V. With the design configuration used in these tests, excellent port short-circuit protection is provided at a very low cost per port.


Other Overcurrent Protection Technology Solutions

Resettable fuse devices are the best solution for overcurrent protection. Other approaches to overcurrent protection design include:

  • Fuses (fusible-link type)
  • Active circuitry (silicon-based current limiters)

Fusible-link fuses are a low-cost alternative but not a viable option as the only protection for which short-circuit operation is expected. A blown fuse renders the device or USB port inoperable until the fuse is replaced. Most low-cost fuse solutions are board-mounted, and are not field-replaceable, which means the unit has to be sent out for warranty service. The cost to both the customer and the manufacturer in the fuse-only case is usually prohibitive. Fuses are often employed in combination with active-circuit protection devices (description follows). Additionally, "one-use" fuses can weaken in the Plug-and-Play/Hot Plug environment, causing them to blow unnecessarily. Most important is that they are not resettable, a violation of the 1.1 version of the USB Specification.

Active-circuitry solutions offer current-limiting USB power switches based on active circuitry in single-chip form. These current-limiting "high-side MOSFET switches" allow both foldback current limiting and power switching in a single package. The foldback-current-limiting feature of this type of chip allows a form of short-circuit protection. These types of devices do have significant drawbacks. Silicon switches have in-rush current problems—they trip too fast. This is particularly troublesome with highly capacitive, high- current loads like CCD cameras and scanners. Furthermore, multiple hot plugs of a USB peripheral can also result in the silicon switch nuisance-tripping. In multiport silicon switches, a fault in one port can cause the adjacent ports to shut down because of internal heating in the silicon package. Thus, only expensive, one-port (not multiple-port) protection devices could be used for providing individual port protection.

The most important disadvantage of silicon switches may be their failure mode. In short-circuit conditions this kind of device cycles in a "thermal-rise shutdown cycle" mode to prevent switch junction meltdown. Depending on the thermal properties of the package and mounting, such rapid under-power thermal cycling can lead to device failure, typically in short-circuit mode, leaving the system with no protection. Active-circuit short-circuit protection devices are therefore usually protected themselves with an additional fuse-device of some kind. This oscillation during faults, also creates noise and EMI problems. Moreover, silicon devices are not UL-recognized safety devices. Finally, another important consideration is cost—silicon switches generally are a much more expensive solution.


Used by leading computer and peripheral manufacturers to protect powered ports, resettable fuses like Raychem's PolySwitch devices have become the standard solution for overcurrent protection. The devices comply with Windows 95 and PC98 standards and meet the UL1950/IEC60950 safety requirements. The variety of part numbers available allow the designer to select the specific device needed to meet USB requirements and the specific requirements of the circuit design.

Contributed by: Raychem Corporation, Electronics Division, 300 Constitution Drive, Menlo Park, CA 94025-1164. Tel: 800-227-7040.