USB - Details on electrical requirements — Oculus
New to the forums? Click here to read the "How To" Guide.

Developer? Click here to go to the Developer Forums.

USB - Details on electrical requirements

MerlinCMerlinC Posts: 51
Hiro Protagonist
After recently upgrading my graphics card to GTX1080 (worked well with my GTX780 but with performance limitations) and not paying attention to what USB 3.0 port I used for the specific devices I had recurring sensor and HDMI failures. The only way to get around the issue I was able to find by trial and error is to plug the sensor to a USB 2.0 port.

Reading all the issues with regard to USB connection of the rift I assume that my USB 3.0 Ports are not able to provide the power/current required by the sensor and/or headset. According to the motherboard  manufacturer the USB ports are according to USB 3.0 standard.

Due to not having had any issue with other USB devices I assume that the current drawn by my rift is either n the edge or over the rated power for USB 3.0 ports. Was not able to find any information what maximum current is specified for the sensor and headset.

Could you please provide this information?
«1

Comments

  • MerlinCMerlinC Posts: 51
    Hiro Protagonist
    edited January 2017
    So far I did not get any feedback / information Oculus. The only information that I received by eMail was that this information is beyond the scope of support they are able to provide.

    Not sure why Oculus support is that reluctant to share the information because disclosing the power consumed by the  device(s) should not cause competitive disadvantage. There must be other reasons?

    Having ongoing issues with sensor / HMD disconnects using the USB Host Adapter as recommended by Oculus I did some research on the USB standard. Furthermore I ordered a device to be able to measure the current-consumption of the sensor / HMD. To get a first indication how much current the sensor / HMD are drawing from the USB port they are attached. 

    Although most vendors of USB Host Adapters do not show there the maximum current per USB-Port (although they provide the information when you ask) most of them support a current up to 0.9A (=4.5W). Thus my current understanding (not reading all the available developer docs so far) what current/power is supported by USB 3.0 is: 
    • Typical USB 3.0 adapters supports power supply for devices consuming up to 4.5W (0.9A@5V). This very likely is the case for all USB host adapters independent whether they have a seperate power supply by a MOLEX or SATA connector. 
    • If the USB BC 1.2 (USB battery charging standard) is supported devices can consume power up to 7.5 W (1.5A@5V)
    • Power supply for devices consuming more than 7.5 W is possible as well in case USB Power Delivery standard is supported. But most USB 3.0 host adapter do not support this standard currently. Furthermore a PD-aware cable needs to be used with power consumption is in the range of 7.5W - 100W.
    Here is one document summarizing the situation from USB.org.
    http://www.usb.org/developers/powerdelivery/PD_1.0_Introduction.pdf

    Question is why do have that many people issues with sensor/HMD disconnects.My hypothesis is that the sensor and/or HDM requires (at least sometimes) more than .9A (or 4.5W) which is the maximum power/current supported by USB 3.0. At the moment there are multiple potential root causes for the issues people having:

    (1) USB ports used not complient with USB 3.0 standard. That's the root cause some people are claiming for - but on my computer it does not work using the USB interface from my motherboard (gigabyte) and using the dedicated USB host adapter as recommended by Oculus. Thus this is not very likely.

    (2) Oculus Sensor comsuming more power than specified for USB 3.0 devices (standard mode 900 mA) or modes like "battery charge" or "power distribution" not activited accordingly. This hypothesis made perfect sense to me until I recognized that the same sensor / cable combination is working connected to a USB 2.0 interface for many people. But the supported power with USB 2.0 is less than at USB 3.0 (USB 2.0: 500 mA / UBS 3.0: 900 mA). Could be answered very easy if Oculus disclosed the information how much power is being drawn by the sensor

    (3) Another potential root cause is that the cabeling of the sensor is not compliant with USB 3.0 requirements. Oculus is not using a standard USB 3.0 cable from a USB certified supplier - instead there is an oculus propriatery cable delivered with the sensor. Higher current and bandwith with USB 3.0 and the cable length  used on the sensor are leading to some challenges with regard to the electrical properties of the cable. Not sure wheter Oculus ever tested the compliance of the used cable as requested in "USB 3.0 Connectors and cable Assemplies compliance document"

    http://www.usb.org/developers/docs/devclass_doc/CabConn_3_0_Compliance_Document_v1.02_2011-10-04.pdf




    From my point of view it is not acceptable that Oculus is forcing their customers to do trial and error to figure out which host adapter / cable configuration is working. They should either provide a working combination or disclose the necessary information that allows their customers to order the right hardware.

    P.S.: This would include the information whether the included USB cableing for sensor and HMD are PD-aware cables in case more than 7.5 W are necessary.


    Remarks: Eddited my previews post due to new learnings - all changes in italic / deletions strikethrough

  • andyfangandyfang Posts: 10
    NerveGear
    The trial and error debugging sucks... I absolutely won't be recommending Oculus to my friends because support doesn't take the time to offer solutions (besides the obvious unplug).
  • MerlinCMerlinC Posts: 51
    Hiro Protagonist
    See changes/updated in my second post
  • TickTockTickTock Posts: 200
    Nexus 6
    I am able to power two sensors and one USB3.0 hub on one port and they work just fine so I am pretty sure it isn't a power issue.  I'm guessing they are still sending full video frames to let the PC do the processing and that is why we end up with BW issues.  Hopefully CV2 will have some smarts in the sensors and only send XY coordinates.  Doesn't take much horsepower - could easily do it with a $2 PIC.
    "I'm not crazy about reality, but it's still the only place to get a decent meal."
     -Groucho Marx
  • kevink808kevink808 Posts: 793 Poster of the Week
    be·ta test
    noun
    1. 1.
      a trial of machinery, software, or other products, in the final stages of its development, carried out by a party unconnected with its development.
    verb
    1. 1.
      subject (a product) to a beta test.
    Gear: Liquid cooled MSI Z170A PC Mate | Intel i7-6700K | 32GB DDR4 SDRAM 3000 Mhz | GeForce GTX 1080, Acer Predator XB241H G-Sync, Rift/Touch/3 cams, Oculus GO 64GB.
  • Nick_BbNick_Bb Posts: 6
    NerveGear
    I used the program USBDeview, it shows the power: http://imgur.com/rrMJ5s8. Rift + Rift audio = Headset USB.

  • TwoHedWlfTwoHedWlf Posts: 2,233 Valuable Player
    Is that an instantaneous current draw or something like the max current mode?  Is that in use or just sitting in standby?  Because I'd be shocked if the HMD is only drawing 100ma.
  • kojackkojack Posts: 5,152 Volunteer Moderator
    What USBDeview is most likely showing is the bMaxPower value.
    USB devices give the host a configuration descriptor. This includes the bMaxPower value, it says how many mA the device could potentially draw.

    This is not an actual measure of the current draw, it's just what the device promises to not go over. The real value could well go over the bMaxPower if the device misbehaves.


  • MerlinCMerlinC Posts: 51
    Hiro Protagonist
    edited January 2017
    Yesterday I received an eMail from Oculus Support indicating that, the recommended USB 3.0 port is not necessary for the current design and that the recommendation for USB 3.0 is just for future incemental bandwith demands.

    But that brings up the question why is Oculus recommending a dedicated USB 3.0 card from a certain vendor when everything is fine with standard USB 2.0 port as well? Sound very strange and I'll continue to provide more information (as requested) to the Oculus support. 

    Here is what Oculus was telling me.

    I'd like to know if using USB 2.0 produces any issues. If not, then please feel free to continue to use USB 2.0 - we recommend USB 3.0 at this time in order to future-proof against increases in USB bandwidth demand by our devices, but those increases are unlikely to occur any time soon. USB 2.0 will work just fine in many cases.



  • Nick_BbNick_Bb Posts: 6
    NerveGear
    edited January 2017

    I found that using USB2 for sensors are very prone to poor tracking quality errors when there are sunlight in their way. My 3rd sensor is partially facing the window, works well during the evening, I never used the rift during the day so I never encountered this problem till now. The errors went away after I changed to a USB3 extension.

    As for the hmd's usb is only 200mA. Maybe the hmd's panels got its power from hdmi with some type of mhl.




  • MerlinCMerlinC Posts: 51
    Hiro Protagonist
    Nick_Bb said:

    ...The errors went away after I changed to a USB3 extension. ...

    What do you mean with USB3 extension? An active USB 3.0 extension cable?

    P.S.: Most people (including myself) do have the issues just connecting the original cable of the sensor to an USB 3.0 port.
  • Nick_BbNick_Bb Posts: 6
    NerveGear
    MerlinC said:
    Nick_Bb said:

    ...The errors went away after I changed to a USB3 extension. ...

    What do you mean with USB3 extension? An active USB 3.0 extension cable?

    P.S.: Most people (including myself) do have the issues just connecting the original cable of the sensor to an USB 3.0 port.
    I used 4 AmazonBasics USB 3.0 Extension Cable for all 3 sensors and the hmd.
    2 sensors are connected to a USB3 hub that is connected to the MSI G45 motherboard.
    3rd sensor and hmd is connected to Inateck Superspeed 4 Ports




  • TwoHedWlfTwoHedWlf Posts: 2,233 Valuable Player
    I tried turning on the hmd and then refreshing,  no change to the power mode.  Wonder if even if the power mode is 100ma it can still draw 1000ma or whatever the max the port can provide? 

    I could see this causing problems if some chipsets enforce the limit and some don't. 
  • kojackkojack Posts: 5,152 Volunteer Moderator
    The idea is that usb devices can define multiple configurations for different uses, the host can choose which one to use. For example a device could say it has both a low current mode that is self powered (uses a power adapter or batteries) or a high current one that is host powered. The host can then decide which mode the device should use (a laptop on batteries could say to use the low power mode, while a laptop on wall power could use the high power one).

    This is the interesting bit though, from http://www.beyondlogic.org/usbnutshell/usb5.shtml
    If a device loses external power, then it must not drain more than indicated in bMaxPower. It should fail any operation it cannot perform without external power.
    That talks about external power, but I think it's generally applicable.
    What if the sensors are reporting a bMaxPower lower than they need, then disconnecting/glitching rather than going over that value. Or maybe some motherboard chipsets are providing slightly less than the requested max current.

    I had a western digital portable hard drive. It only worked on really short usb cables, on a slightly longer one it just couldn't get enough current (usb hard drives were notorious for trying to pull more than usb was supposed to support).

    The other interesting thing is that USBPcap (usb packet sniffing) can see my rift sensors on my laptop (it's usb3 is fully rift friendly) and on my desktop's usb2 ports. But on the desktop's usb3 ports that are glitchy (frequent disconnects, when it decides to work), it can't see them at all. Everything else is there, just not the rift.



  • kojackkojack Posts: 5,152 Volunteer Moderator
    I saw a thread on reddit about tracking issues, and two people have both claimed that turning motherboard XMP on or off reliably controlled if their sensors glitched. XMP on gave frequent glitches, XMP off and everything worked fine. Tested multiple times with no other changes.

    XMP does auto configure of ram based on profiles embedded in the ram modules. It includes power settings.
    If XMP is cranking up the voltage more than default, that could negatively affect the usb host I guess?

    I can't even get into my bios at the moment, it has stopped accepting my K95 keyboard during bios start.


  • Hornet_F4CHornet_F4C Posts: 213
    Nexus 6
    If power supply is an issue I wonder if they could provide a powered break-out box for people with more than two sensors? On the PC side the break-out box could even have two USB connectors along the HDMI connector, thus could be connected to two indpendend USB controllers if desired. And of course it could get power from an ac adaptor. Further it would easily give another meter of cable length and could also provide some intelligent connector system to protect the ports from pulling the cable...
    Of course it should offer a rock steady 4 sensor setup to justify itself.
    Would certainly be a welcome addition, no?
  • steviehstevieh Posts: 10
    NerveGear
    I had my Oculus replaced for black screen issue just received the new one today and have the same issue, gutted!
  • MerlinCMerlinC Posts: 51
    Hiro Protagonist
    stevieh said:
    I had my Oculus replaced for black screen issue just received the new one today and have the same issue, gutted!
    Sorry to hear that a replaced device did not solve the issue.

    On the other hand it's not a big surprise to me because fixing an issue requires to accept that there is currently a design issue.

    All feedback from Oculus I received so far is more in the direction that they are assuming that the root cause is not at there device (e.g MS windows, USB cards although certified by USB.org not compatible, etc).

    But I have not seen that Oculus either acknowledged that there is an issue or provided evidence that there device can't be the root cause.

    Let's start with the USB interface and cabeling. Why is it that difficult to provide the current drawn from the devices sensor/HMD? If they are in the limits specified by USB standard I do have something in my hands to get either back to my motherboard or USB card vendor. 

    Next would be USB cabeling. Carefully reading USB 3.0 articles you can figure out that there is no maximum cable length - but due to the higher bandwidth - there must certain specification met that USB 3.0 works reliable. Thus cables a lot of high quality (higher priced) cables starting with a cable length are having an amplifier integrated to guarantee sufficient signal quality.

    Furthermore you can read that meeting this requirements will increase the cost and effort on the cable side if cable length is greater 3m - and we get a cable which 5 m and can't be replaced with a shorter standard cable. Thus Oculus please do us and yourself a favor and execute the USB 3.0 compliance tests and get it certified by USB.org (or provide the technical data to your customers).

    You might want to consider a revised design providing standard USB and HDMI connectors than people can choose there preference of cable length and supplier. And spare cables would be no issue anymore.
  • MerlinCMerlinC Posts: 51
    Hiro Protagonist
    Further it would easily give another meter of cable length and could also provide some intelligent connector system to protect the ports from pulling the cable...
    On the USB side I don' think that higher output current will simple enable higher cable length - enabler (and challenge) is meeting the USB electrical cable requirements independent of length.
  • TwoHedWlfTwoHedWlf Posts: 2,233 Valuable Player
    MerlinC said:
    stevieh said:
    I had my Oculus replaced for black screen issue just received the new one today and have the same issue, gutted!
      Why is it that difficult to provide the current drawn from the devices sensor/HMD? If they are in the limits specified by USB standard I do have something in my hands to get either back to my motherboard or USB card vendor.
    If you're that hugely stressed over the current drawn by them, why don't you measure it?  Shouldn't take anything more than a short USB extension and a cheap multimeter, cut it open and measure the current through the +5v wire.

    Should be red.

  • Hornet_F4CHornet_F4C Posts: 213
    Nexus 6
    TwoHedWlf said: If you're that hugely stressed over the current drawn by them, why don't you measure it?  Shouldn't take anything more than a short USB extension and a cheap multimeter, cut it open and measure the current through the +5v wire.

    This should work as well?

    https://www.amazon.com/Centech-USB-Power-Meter/dp/B00DAR4ITE

    Would be suprised if that wasn't already tested by somebody.

  • TickTockTickTock Posts: 200
    Nexus 6
    edited January 2017
    HMD + USB3 repeater reads 0.31-0.41A (bounced around).  Two sensors, a USB3 hub and a USB3 repeater read 0.61A. This is while in operation.
    "I'm not crazy about reality, but it's still the only place to get a decent meal."
     -Groucho Marx
  • TickTockTickTock Posts: 200
    Nexus 6
    The active USB repeaters appear to be negligible. I get about the same reading at the HMD as when measured at the PC side of the repeater (bounces around so hard to say exactly).  Unable to check the sensors independently since the hub is mounted to the ceiling in a way that prevents reading the meter when it is plugged in.
    "I'm not crazy about reality, but it's still the only place to get a decent meal."
     -Groucho Marx
  • Hornet_F4CHornet_F4C Posts: 213
    Nexus 6
    TickTock said:
    HMD + USB3 repeater reads 0.31-0.41A (bounced around).  Two sensors, a USB3 hub and a USB3 repeater read 0.61A. This is while in operation.

    That means around 0.9A for three sensors + 0.4A for the HMD.

    Is the USB specication meant per port or per controller? If every port can deliver 0.9A each, this should be fine, but if the 0.9A are shared between all ports of the same controller that could be a problem for some setups...

  • steviehstevieh Posts: 10
    NerveGear
    You guys are way more technical than I am, good luck though I appreciate it shame Oculus aren't involved in coming up with a solution or ideas with you as the  community, maybe they will eh?
  • MerlinCMerlinC Posts: 51
    Hiro Protagonist
    TickTock said:
    HMD + USB3 repeater reads 0.31-0.41A (bounced around).  Two sensors, a USB3 hub and a USB3 repeater read 0.61A. This is while in operation.

    That means around 0.9A for three sensors + 0.4A for the HMD.

    Is the USB specication meant per port or per controller? If every port can deliver 0.9A each, this should be fine, but if the 0.9A are shared between all ports of the same controller that could be a problem for some setups...

    USB specification is .9A per port and w/o specific modes like battery charging or power distribution. But to be sure you should have a look at the technical data provided by the vendor of your card or just ask him.
  • MerlinCMerlinC Posts: 51
    Hiro Protagonist

    TwoHedWlf said: If you're that hugely stressed over the current drawn by them, why don't you measure it?  Shouldn't take anything more than a short USB extension and a cheap multimeter, cut it open and measure the current through the +5v wire.

    This should work as well?

    https://www.amazon.com/Centech-USB-Power-Meter/dp/B00DAR4ITE

    Would be suprised if that wasn't already tested by somebody.

    You are both right.

    Yes I will do the measurement on my rift and sensor although this values should be available at Oculus. Challenging thing is that the really interesting currents are those drawn when the sensor failure and/or blackouts occur. Thus the cheap multimeter most likely will not do the job to get a trace of the current for a view 500 ms before and after the event.
  • Hicks613Hicks613 Posts: 297
    Art3mis
    Why did Oculus drop the AC adaptor?  This used to be an effective way to inject additional power to the HMD.
  • XlordBXlordB Posts: 289
    Nexus 6
    edited January 2017
    I so want to get into this thread and find out what the maximum power draw a sensor or the rift would draw when its pushed.. if we can find out the maximum the rift or sensors draw then we can theoretically measure our maximum power per usb port and then figure out whether usb 2 OR 3 are actually no good for the rift and MAYBE it might be worth looking into a pci card with usb 3.1 ports instead of usb 3.

    Im currently due to move house so im busy painting and stuff no time to look more into these pressing issues.

    HEY cyber since your like the inside man for us can you get any info on the power specs of the rift or at least the sensors.?

  • vrmowervrmower Posts: 181
    Art3mis
    Still think those having problems should try a ferrite on each Rift USB cable.

    Environment already a little intense with USB3 and HMD+Sensor transceivers all running @ 2.4ghz.  Does not take much fail in shielding/grounding dept to cause headaches.
Sign In or Register to comment.