Ignore Suspend Current in USB Bus Pwr application?

It certainly won't pass if you draw more than 100mA before it has been configured.

You can, I believe, generally get a dispensation for the suspend mode current - the normal limit is 500uA but 2.5mA is permitted. If your device can manage it you can rely on the fact that the current is averaged over one second (see 7.2.3, USB 2.0).

Whether you need USB certification may depend upon your customers.

The only way that I've been able to get near the suspend requirements in the past is to shut everything down and have the USB device (effectively) control the MCU. In some circumstances this isn't ideal and self-powered is the way to go.

Andrew

Andrew Jackson wrote in news: snipped-for-privacy@eclipse.net.uk:

So if we assume that the current can never be actually suspended (not even close to a few mA), then the device will still work but it will never be USB compliant in Bus Power.

The situation I have in mind is an instrument tied to a notebook computer. It is really inconvenient to have an external supply in this case. The instrument would probably still have an external power input to allow optional self power operation and therefore avoid the issue.

I too have used this method, but it doesn't play well with my current design constraints.

Thanks

Al

Yes. In normal use, I've never had a problem with high suspend currents. Clearly overcurrent (>500mA) is a different matter and all sorts of alarms go off then!

I've a number of USB instruments that aren't certified (though the controller chips are) but perform perfectly well. For my projects, I normally run through all the USBCV tests anyway (irrespective of whether I am actually trying for certification).

The suspend current has always seems way too low to me. I can understand the reasons for it but it you want any sort of intelligence in your (slave) device then it is a real pain and requires a lot of effort to achieve.

Andrew

Also my experience.

Me too.

The change from 500uA to 2.5mA did the difference between "completely impossible" to "should be possible, so must be achieved". All my recent devices are compliant, but I cannot remember the last time a client actually wanted to do certification. But suspend sure does cost a lot of software development time. Suspend is fine for the few devices where remote wakeup is relevant, but for all the rest, im really missing a bmAttributes bit to say "suspend is useless to me, just cut the damm power".

Leo Havmøller.

Even though the spec only defines 2.5W power (500mA @ 5V) there are USB hubs that provide 5W. We have used these in the past. And have used non-suspend supporting designs without issue. Never certified.

Chris Felton

Christopher Felton wrote in news:ipqnpe$d0a$1@dont- email.me:

not

even

This may not exist and further show my USB ignorance, but is there some kind of simple hub-like device that could act as a USB power management device that would be just smart enough act as a pass thru extender and manage power during enumeration and suspend.

It seems to me that this would be pretty handy for a lot of designs.

Al

Hopefully, I didn't mis-speak. Doing a quick google I can't find a similar USB hub that I thought existed. I will need to go through my emails and see if I can find the hub that we used in the past. We used this for a board that was normally self-powered but when we were testing the board it was simpler to run off USB power.

Chris

An alternative is to have two USB connectors, as found on some of the portable hard drives, one of them only provides power.

Andrew

Here is the USB hub, I believe we used, it has two "high" powered ports.

Chris