MajorLanGod wrote:
Back to the original question: Back when USB 2.0 was what was available,
some devices like external drives needed more power than a 2.0 port could
provide. Y-cables were sold that would plug into two USB ports, one a full
commection with data and power, and the other power only. If one connected
the power only plug to the USB port, and the other end to the device
needing power, you would accomplish what you wanted.

USB2 ports come in "stacks of two". They're numbered in pairs.

----- fuse ---+--- USB2 #5
|
+--- USB2 #6

In the "config space" of the USB2 device, are some declarations.
A device may seek to define itself as a "high power device".
The declaration says "500mA".

The Alcatel "Frog Modem" was an early example of a violator.
While performing Internet functions, it drew 530mA.

The electronic identifier, has nothing to do with "enforcement".

Intel wrote a white paper on the topic, telling designers
to "not be policemen". The purpose of the fusing, is
to prevent copper tracks or electrical connector pins from
being burned.

Apparently they can take more of a load than the 500mA.

A typical Asus brand retail motherboard is like this.

----- fuse ---+--- USB2 #5
1.1A |
+--- USB2 #6

What this means is, if a Toshiba 2.5" hard drive that runs
off 5V and the USB bus, starts to spin, it draws 1.0A for
ten seconds. The fuse holds. It holds as long as two drives
aren't plugged in and spinning up at the same time.

Hard drive power consumption (for the motor) drops once
accelerated up to speed.

However, not all hardware works this way.

In laptops, Polyfuses would overheat, because there is no airflow.
In laptops, a "silicon policeman" is used. It's a dual channel
protection device, in an 8-pin DIP package. It isolates VBUS
with a MOSFET. If more than 500mA is drawn (say, 501mA), the
fuse opens instantly. It isn't a pokey fuse like the Polyfuse.
This makes laptops "brittle" when it comes to USB2 drives.

The Y cable was invented, so the second head would carry power.
On a laptop, they were hoping the current flow allocations would
add nicely. But that's not exactly what happens electrically.

The situation for 2.5" drives is hardly ideal. Available power
can be marginal in some cases.

they could easily have used a 1.1 for *each* port if they
wanted to. But you know how cheapskates are. They're cheap.

One dick-head designer, put *all* the 5VSB loads on a motherboard,
on a single fuse. PS/2 connectors, Parallel Port, USB ports,
all sitting off *one* fuse. Just to save some pennies. Imagine
how happy the customers of that board would be... His fellow
designers do not agree with that approach. That's why nobody
else does that.

The hard drives could all have been given barrel power
connectors, with a proper circuit for switchover. But it's
the usual story: "the other guy should fix this, why should
I fix it?".

There are occasional boutique designs, which go overboard.
For some of this style of problem, some enclosures had
a "booster circuit" which stored up power in advance before
spinup, then there are more "local amperes" to finish the
spinup phase. Seagate did this for something a couple years ago.

It's a topic of constant amusement. Bad engineering, that never
got fixed over the years. It doesn't have to be this bad, really...

Why should customers have to be exposed to these topics ???

Paul