I seem to recall a discussion thread regarding speeding up
recognition/enumeration of new devices. The problem that needs to be solved
is that it's taking from 7 to 20 seconds for a new device to become
available for communication after power on connection. The folks at our
production facility want this to happen more quickly.

The port that we will be using can be dedicated to this function. The device
VID, PID, and Version will be constant but the USB serial number will
probably be changing. My inital reaction is that the amount of time needed
for enumeration is determined by how much work the PnP manager has to do to
figure out the correct driver.

As I mentioned, I seem to remember a similar issue that was on the list, but
a search of current articles doesn't seem to turn anything up. Does anyone
remember this one?

Sam Robinson

Re: USB enumeration time in a production test environment by Ray

Ray
Thu Jun 30 16:05:15 CDT 2005

There's a fairly good discussion of the USB serial number handling on
Raymond Chen's blog at
http://blogs.msdn.com/oldnewthing/archive/2004/11/10/255047.aspx

I haven't tried this, but there's a registry key mentioned in there that
is supposed to get Windows to ignore the serial number for a particular
HWID:
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\UsbFlags\IgnoreHWSerNumVVVVPPPP

Samuel Robinson wrote:
> I seem to recall a discussion thread regarding speeding up
> recognition/enumeration of new devices. The problem that needs to be solved
> is that it's taking from 7 to 20 seconds for a new device to become
> available for communication after power on connection. The folks at our
> production facility want this to happen more quickly.
>
> The port that we will be using can be dedicated to this function. The device
> VID, PID, and Version will be constant but the USB serial number will
> probably be changing. My inital reaction is that the amount of time needed
> for enumeration is determined by how much work the PnP manager has to do to
> figure out the correct driver.
>
> As I mentioned, I seem to remember a similar issue that was on the list, but
> a search of current articles doesn't seem to turn anything up. Does anyone
> remember this one?
>
> Sam Robinson
>
>

--
../ray\..

Re: USB enumeration time in a production test environment by Robert

Robert
Fri Jul 01 04:40:28 CDT 2005

Samuel Robinson wrote:

> The port that we will be using can be dedicated to this function. The device
> VID, PID, and Version will be constant but the USB serial number will
> probably be changing. My inital reaction is that the amount of time needed
> for enumeration is determined by how much work the PnP manager has to do to
> figure out the correct driver.

Changing the serial number is a bad idea.
This is where Windows takes its time and clogs up the registry.