Quick Answer: A mouse or keyboard moves well under 1 Mbit/s, so USB 2.0 at 480 Mbit/s is never the bottleneck; power and port sharing are. A wired mouse polling at 1000 Hz reports every 1 ms, and that timing is what a hub can damage, not the data rate. Keep passive extension to 5 m in total, add a powered hub beyond that, and leave high-polling devices off a hub that also carries a drive. Bluetooth adds its own delay, so use a wired link or a dedicated 2.4 GHz receiver for timing critical work.
The polling rate decides what an input device needs: 125 Hz means a report every 8 ms, 1000 Hz means every 1 ms, and neither fills a USB 2.0 link of 480 Mbit/s. What does break is power and timing. A hub that renegotiates its upstream link, or a port sharing a controller with a busy drive, adds delay a user feels long before any bandwidth figure moves. Length, power and port sharing decide the rest.
Bandwidth is not the bottleneck
What a mouse actually sends
A 1000 Hz mouse sends a report every 1 ms, and each report is a handful of bytes: buttons, two movement axes and a wheel step. That is a few tens of kbit/s. A keyboard interrupts far less often. Even with several devices on one port, the traffic stays below the capacity of a full-speed 12 Mbit/s link, let alone 480 Mbit/s.
Where the 5 m figure comes from
The 5 m limit for passive USB 2.0 cable is a signal-integrity rule, not a power rule: the host must see a device reply within a fixed time after it asks. Long, thin cable attenuates the edges and the reply arrives late, so the device drops off or re-enumerates. Active cable or a hub in the middle resets the clock.
Power and port sharing
A keyboard with a backlight, a mouse with a charging dock and a bus-powered drive all draw from the same 5 V rail. A port rated at 0.5 A for USB 2.0 cannot feed all three, so the hub browns out and devices vanish mid-session. Count the current, not the ports.
When a powered hub earns its place
A hub with its own 5 V supply holds the upstream link while it feeds each downstream port, which is what a powered USB hub adds over a passive splitter. Put receivers and keyboards on it, keep drives on a separate port, and check that the hub does not sit between the receiver and the host in a way that adds a renegotiation step.
Interfaces and limits side by side
The figures below are the ones that matter when you plan a desk. Passive length is the total of every cable and adaptor between host and device, not the length of one lead.
| Interface | Data rate | Passive limit | Typical device |
|---|
| USB 1.1 low speed | 1.5 Mbit/s | 3 m | keyboard with hub |
| USB 2.0 full speed | 12 Mbit/s | 5 m | keyboard, mouse |
| USB 2.0 high speed | 480 Mbit/s | 5 m | mouse at 1000 Hz |
| USB 3.2 Gen 1 | 5 Gbit/s | 3 m | hub uplink, drive |
| USB 2.0 through a powered hub | 480 Mbit/s | 5 m plus 5 m | mouse and keyboard |
Lag you can measure and lag you cannot
Total input delay is polling interval plus transport plus display. At 125 Hz the interval alone is 8 ms, and no cable shortens it. A good USB extension cable adds nothing measurable; a chain of two passive leads adds dropouts rather than delay. Where two machines share one keyboard, a USB KVM switch adds a switching step, so test it before rolling it out to a whole desk.
Practical scenarios
- Situation: a mouse stutters only while a portable drive is copying. Move the drive to its own port or to a powered hub, because the shared controller is stalling the 1 ms reports.
- Situation: the desk is 6 m from the machine. Use a single 5 m passive lead plus a powered hub, so the far segment stays short and the hub refreshes the link.
- Situation: a backlit keyboard drops out at boot. It draws more than the 0.5 A a USB 2.0 port guarantees, so fit a hub with its own 5 V supply and enough current for the load.
- Situation: two machines share one keyboard and mouse. A KVM switch is the tidy answer, provided it passes 1000 Hz reports and does not bury the receiver inside a metal enclosure.
FAQ
Does a USB hub add lag to a mouse?
A hub adds a fraction of a millisecond, which sits below the polling interval of any mouse. Delay becomes visible when the hub is overloaded or shares a controller with a drive, because reports then queue. A powered hub with current headroom is the safe choice.
Can I join two USB extension cables?
Not for anything timing sensitive. Each joint adds a reflection, and the total passes the 5 m passive limit for USB 2.0, so a device may enumerate and then drop. Use one active extension or a powered hub instead of two passive leads.
Is wireless worse than wired for a keyboard?
A 2.4 GHz receiver on a dedicated link is close to wired for typing, with delay in the low millisecond range. Bluetooth shares its band and its polling schedule with other devices, so it suits office work and is less predictable for timing critical input.
