XCORE-200 VBUS

Technical discussions around xCORE processors (e.g. xcore-200 & xcore.ai).
henk
Respected Member
Posts: 347
Joined: Wed Jan 27, 2016 5:21 pm

Post by henk »

Hi Caleb,

That sounds like quite a long time; can you remove a cap to shorten that?
Also - instead of removing the #define SELF_POWERED, you may want to replace it with

Code: Select all

#define SELF_POWERED       0
Cheers,
Henk


User avatar
ccrome
Active Member
Posts: 62
Joined: Wed Sep 23, 2015 1:15 am

Post by ccrome »

Will do, thanks.

BTW, I just got a Beagle USB 480, so now I should have better visibility into what's happening. Plowing through the manual now...

-Caleb
User avatar
ccrome
Active Member
Posts: 62
Joined: Wed Sep 23, 2015 1:15 am

Post by ccrome »

Hi all,
Just to clear up everything for posterity:

The real problem on this one was not to do so much with startup timing, but rather the VBUS damaging the USB port on the MXOS part. It's definitely a good idea (really necessary I'd say) to put the RC filter in there.

We have new boareds with the RC filter, and we can't get them to fail.
dweeb4
Active Member
Posts: 52
Joined: Mon Jan 19, 2015 12:47 pm

Post by dweeb4 »

What ripple & Freq was on the Vbus line that damaged the XMOS chip?
User avatar
ccrome
Active Member
Posts: 62
Joined: Wed Sep 23, 2015 1:15 am

Post by ccrome »

dweeb4 wrote:What ripple & Freq was on the Vbus line that damaged the XMOS chip?
I beleive the larger issue is that it was a bus powered device that took a long time to power up. Therefore the xmos part had no power for a while (a bunch of milliseconds), yet a full 5V with no RC filter or anything to limit current. So... reality is probably 5V above VCC, plus whatever inductive kickback happened during plug insertion.

The new boards fix both the startup time and the RC filter. Now that I think of it though, there should be some additional clamp as well after the RC filter to prevent any meaningful amount of current into VBUS before the power is stable.

-Caleb
User avatar
mon2
XCore Legend
Posts: 1913
Joined: Thu Jun 10, 2010 11:43 am
Contact:

Post by mon2 »

USB load switch with built-in inrush current protection is recommended here to protect downstream devices. Many are available with 0.5A as max limit. ESD protection on top of this to prevent Electro syndrome.

Image

Also, Vbus should have a max of 10 uf capacitance for this reason of inrush currents.

http://www.testusb.com/inrush_issue.htm
RitchRock
XCore Addict
Posts: 186
Joined: Tue Jan 17, 2017 9:25 pm

Post by RitchRock »

Do you know if a single channel USB load switch with soft-start, set to 500mA is OK, or do we need to first limit at 100mA prior to enumeration? Here is an example where the current is first limited to 100mA, however I'm wondering if this is needed: http://ww1.microchip.com/downloads/en/A ... ations.pdf
User avatar
akp
XCore Expert
Posts: 578
Joined: Thu Nov 26, 2015 11:47 pm

Post by akp »

I don't have a lot of experience with XMOS USB but I have used USB device for years. To conform to the spec you might need something like the dual current limit. However, it really shouldn't affect the performance of the PC or the XMOS chip or cause any USB issues; I've never done it and I've never had a problem.
User avatar
mon2
XCore Legend
Posts: 1913
Joined: Thu Jun 10, 2010 11:43 am
Contact:

Post by mon2 »

@RichRock, that is an interesting find:

http://ww1.microchip.com/downloads/en/A ... ations.pdf

but the article in our opinion is dated and is now 17 years old. Strongly believe there are single device solutions that will achieve the same support at a lower cost and smaller footprint with the soft start features.

Consider for example,

AP2331 = 200 mA limit with soft start to limit in-rush currents

https://www.diodes.com/assets/Datasheets/AP2331.pdf

http://www.mouser.com/search/refine.asp ... =183556915

or

one of the many other choices (upto 500mA) - select a suitable USB 2.0 compliant load switch with a soft start feature and you should be fine:

https://www.diodes.com/products/power-m ... tab-finder

From a rough cost estimate, the article you cited will amount to around $2 USD + assembly fees for the mix of external parts = too costly for this required protection. A single device can replace all of these discretes. The front of the referenced article is doing just that, performing a delayed (soft start) ramp up and current limit of in-rush currents. Best to wire up a device for testing before moving to PCB layout.
RitchRock
XCore Addict
Posts: 186
Joined: Tue Jan 17, 2017 9:25 pm

Post by RitchRock »

Thank you very much for the insight - the cost / component count was my main concern with having to conform to the spec in the article I referenced.

One more question in this regard - if the device is to be either be powered from Vbus OR external +5v supply (or for that matter a battery), how should the Vbus pin on the XMOS chip be connected?
Post Reply