Analog Discovery 2 Current Exceeds USB 2.0 Specification

[clf]

I've measured the Analog Discovery 2 drawing a current of 510 mA over USB.  The USB 2.0 specification [1] dictates that devices shall not draw an absolute maximum current greater than 100 mA before configuration and shall not draw an absolute maximum current greater than 500 mA once configured.  It appears to me that the AD2 is not compliant with the USB 2.0 specification.

This could explain the Waveforms 2015 errors messages that I reported in another thread on this forum [2].

Measurements were taken on a "warm" AD2 that had already been operating; I anticipate peak current on an AD2 that has been disconnected for some period will be higher (capacitances discharged).  Measurements were performed with the AD2 running solely from USB power.  Measurements were taken while connecting the AD2 to a PC on which the Waveforms 2015 software was already running.  Measurements were not taking during AD2 acquisition or waveform generation operations.  Didn't get a photo of the 510.1 mA measurement, but a photo of a 507.7 mA measurement is attached.
 
The measurements were made with a Fluke 87V in mA DC mode with the range manually fixed to "600".  The large currents were detected using peak-min-max mode (250 microsecond response time).  The large currents were not detected using regular min-max mode (100 millisecond response time).  

The USB 2.0 spec makes the peripheral responsible for limiting its in-rush current and implementing soft-start as needed.

-- Chris

[1] http://www.usb.org/developers/docs/usb20_docs/usb_20_120516.zip --> usb_20.pdf, sections 7.2.1 and 7.2.4.1

[2] https://forum.digilentinc.com/topic/3226-waveforms-2015-jtscinitscanchain-failed-error-dptiio-failed-error-ubuntu-16041-x86-64-analog-discovery-2-analog-discovery-2/#comment-11671

 

[attila]

Hello,

The device needs about 2W to work, with 5V supply about 400mA. You can check the integrated hot swap controller and power monitor readings in the WaveForms application status bar. https://reference.digilentinc.com/reference/instrumentation/analog-discovery-2/reference-manual#usb_power_control

The current increases as the voltage drops due to cabling resistance (USB plugs connections, your breadboard, ammeter shunt resistance...)
Above 500mA a warning is shown in the application (Status: OC) and at 1A the device is stopped.
Nowadays computers/laptops can provide much more power, like an USB powered HDD takes as much as 2A.

 

[clf]

Hi @attila,

Thanks for your reply.  I don't quite follow your explanation.

1)  Regarding your screenshot of the "USB Current" reported in the Waveforms software:

1a)  Is that value measured by the host or by the Analog Discovery device?  If this value is self-reported by the Analog Discovery device, then measurements could only commence once the device is communicating.  If the device is communicating, then it has already fully powered up and initialized.  If it is already fully powered up, then the inrush current spike I'm reporting has already passed.

1b)  Is this value a sample at a point in time, an average over an interval, or a maximum over an interval?

1c)  What is the period of the interval from 1b?  As mentioned in my original post, I only detected the spike when I enhanced my meter's interval to 250 microseconds, as the spike wasn't detectable on the 100 millisecond interval.

2) Regarding "The current increases as the voltage drops due to cabling resistance (USB plugs connections, your breadboard, ammeter shunt resistance...)":

All of those factors contribute a resistance in series with the Analog Discovery 2.  The USB port's voltage is fixed.  By Ohm's Law, introducing these resistances will decrease (not increase) the current.  Therefore, the inrush current I am measuring actually underepresents the true inrush current of the Analog Discovery 2 product in a normal configuration.

3) Regarding "Nowadays computers/laptops can provide much more power, like an USB powered HDD takes as much as 2A.":

Digilent's literature [1] indicates that the Analog Discovery 2 is a USB 2.0 device.  As mentioned in my original post, the USB 2.0 specification [2] does not permit devices to draw more than 500 mA; drawing more than 100 mA requires the host's consent.

I went through the trouble of taking these measurements because (a) my Analog Discovery 2 device + Waveforms was malfunctioning, and (b) you'd indicated that the malfunction was a power problem [4].  The measurements confirm that there is a power problem.  These measurements characterize the power problem as being specific to the Analog Discovery 2 device and not to the host.

For what it's worth, the host is an enterprise-grade Dell.

Has the Analog Discovery 2 been tested for USB 2.0 compliance by a third-party certification agency?  I haven't been able to find any USB-IF trademarked markings [3] on the packaging or literature.

-- Chris

[1] https://reference.digilentinc.com/reference/instrumentation/analog-discovery-2/reference-manual?redirect=1

[2] http://www.usb.org/developers/docs/usb20_docs/usb_20_120516.zip --> usb_20.pdf, sections 7.2.1 and 7.2.4.1

[3] http://www.usb.org/developers/compliance/

[4] https://forum.digilentinc.com/topic/3226-waveforms-2015-jtscinitscanchain-failed-error-dptiio-failed-error-ubuntu-16041-x86-64-analog-discovery-2-analog-discovery-2/#comment-11671

 

[sbobrowicz]

I asked about the USB-IF testing, I'll let you know when I hear back.

As for what you are measuring, after poking around the USB-IF documentation a bit, it seems that for the inrush spike they are only concerned about charge and not current. In other words, they don't care about how high the current spike is but rather its integral over the "inrush region". This is largely defined by the capacitance of VBUS on the device, and they spec out a maximum of 10uF.

Another interesting note: not having an inrush spike greater than at least 100 mA will cause you to fail USB-IF. This is because it is used for connection detection.

So if you wanted to run a test to see if the AD2 meets inrush requirements you will need to use a scope to measure the current instead, so you can see the waveform and integrate it to find the charge (though it seems unclear exactly what charge is acceptable to pass). Also, you will need to make sure your test setup doesn't introduce too much additional capacitance

[sbobrowicz]

I heard back from RnD... We did not pay for USB-IF certification, however we did use inrush limiting circuitry (ADM1270) and ran internal tests to gain confidence that we are compliant.

As attila mentioned, our software warns you when you get over the 500 mA limit, but doesn't shutdown until you get to 1 A. This is because many computers will provide this additional power without issue, and customers often like to take advantage of that to power more hungry circuits. Strictly speaking this probably isn't USB 2.0 compliant, but the violation can easily be avoided by changing how the AD2 is being used.

Attila, is it possible for customers to configure the software so that the AD2 will shutdown when 500 mA is exceeded so that a customer can feel confident that they are always operating within USB 2.0 spec?

[attila]

The earlier WaveForms (v2) application had such feature. This was kind of annoying and result lot of support questions.

[clf]

I wouldn't reintroduce that feature because it wouldn't solve this: the transients in question occur before the AD2 and Waveforms are communicating in that mode.

Also, there is an interesting transient upon "Select"-ing the AD2 in the Waveforms Device Manager.