Oak-D Lite insists on USB2.0 (slow)

[Pito Salas]

Hi all,

A little help from my friends. My Oak-D Lite happily works on my raspberry Pi. But the pi never recognizes it as high speed (USB3.0)

/:  Bus 001.Port 001: Dev 001, Class=root_hub, Driver=dwc2/1p, 480M

/:  Bus 002.Port 001: Dev 001, Class=root_hub, Driver=xhci-hcd/2p, 480M

/:  Bus 003.Port 001: Dev 001, Class=root_hub, Driver=xhci-hcd/1p, 5000M

/:  Bus 004.Port 001: Dev 001, Class=root_hub, Driver=xhci-hcd/2p, 480M

    |__ Port 001: Dev 003, If 0, Class=Hub, Driver=hub/4p, 480M

        |__ Port 003: Dev 005, If 0, Class=Vendor Specific Class, Driver=[none], 480M

/:  Bus 005.Port 001: Dev 001, Class=root_hub, Driver=xhci-hcd/1p, 5000M

    |__ Port 001: Dev 002, If 0, Class=Hub, Driver=hub/4p, 5000M

NOTE: the cable is tested as a USB3.0 cable. It is USBA on one side and USBC on the other.

NOTE: the above shows it with the powered usb hub but it does the same if I plug the camera directly into the Raspberry Pi 5:

/:  Bus 001.Port 001: Dev 001, Class=root_hub, Driver=dwc2/1p, 480M

/:  Bus 002.Port 001: Dev 001, Class=root_hub, Driver=xhci-hcd/2p, 480M

    |__ Port 001: Dev 002, If 0, Class=Vendor Specific Class, Driver=[none], 480M

/:  Bus 003.Port 001: Dev 001, Class=root_hub, Driver=xhci-hcd/1p, 5000M

/:  Bus 004.Port 001: Dev 001, Class=root_hub, Driver=xhci-hcd/2p, 480M

/:  Bus 005.Port 001: Dev 001, Class=root_hub, Driver=xhci-hcd/1p, 5000M

I have chatted for a while with Claude and it doesn’t believe what it’s seeing and insists that the Camera therefore must be defective.

I need some actual human assistance. If you’ve worked with the Oak I think you’ve been through the wringer and back. Any thoughts?

Best,

Pito

Boston Robot Hackers && 

Comp. Sci Faculty, Brandeis University (Emeritus)

[Nathan Lewis]

Hi Pito,

When the Oak-D cameras power up they boot into USB 2 mode. They won’t expose USB 3 until an application starts. Do you have software you’re trying to run that’s reporting a failure to use USB 3?

- Nathan

--

You received this message because you are subscribed to the Google Groups "HomeBrew Robotics Club" group.

To unsubscribe from this group and stop receiving emails from it, send an email to hbrobotics+...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/hbrobotics/D601A734-6139-4455-966F-5ACFB2D7E066%40gmail.com.

[Pito Salas]

Hi Nathan

Thanks… I did come across that fact along the way. Oddly I found that if I flip the USBC plug over it will work vs. not work the other way around. And this was reproducible (at least once). Have you ever heard of that? It’s pretty weird.

Best,

Pito

Boston Robot Hackers &&
Comp. Sci Faculty, Brandeis University (Emeritus)

> To view this discussion visit https://groups.google.com/d/msgid/hbrobotics/f03d9e60-8fdb-47c9-9c87-9a391e174f28%40app.fastmail.com.

[Ross Lunan]

Re the asymmetric connector behavior, attached is the display on my new handy-dandy Treedix Cable tester on the USB3 to USB2 data cable I use for my Oak-D-Lite, (that uses a Power/DataSplitter to separately feed DC from a separate PowerPack). Note the USB-C D+ D- are only connected to B6 & B7. Interestingly I have not encountered a connection problem running the ros_examples scripts. I'll investigate further. Ross 

[Pito Salas]

I have the same tester. But I didn’t know to notice B6 and B7 etc. So you think it’s the cable (which I tested to be good for usb3) not the camera. Interesting…..

Best,

Pito

Boston Robot Hackers &&
Comp. Sci Faculty, Brandeis University (Emeritus)

> To view this discussion visit https://groups.google.com/d/msgid/hbrobotics/91756e43-e64d-40db-8ab9-3dabd69feedcn%40googlegroups.com.

[Nathan Lewis]

USB-A to USB-C cables only attach the wires to a single side of the USB-C port and depend on the USB-C device to implement cable reverse correctly.

One does not make a fork in high-speed impedance controlled differential pairs :)

- Nathan

To view this discussion visit https://groups.google.com/d/msgid/hbrobotics/91756e43-e64d-40db-8ab9-3dabd69feedcn%40googlegroups.com.

[Pito Salas]

Fascinating! You know a lot about this stuff.

So if the cables always attach wires to one side, and the camera hardware is fixed, this implies that this effect should be universal across anyone trying to use the Oak-d Lite. Is that correct reasoning?

Best,

Pito

Boston Robot Hackers &&
Comp. Sci Faculty, Brandeis University (Emeritus)

> To view this discussion visit https://groups.google.com/d/msgid/hbrobotics/67a4f184-05d2-480c-9a8f-4737161c513b%40app.fastmail.com.

[Alan Timm]

in case it helps, oakd is *VERY* picky about the usb cables you use, and it was never obvious to my why one cable works in USB3 while another one only goes to USB2

To view this discussion visit https://groups.google.com/d/msgid/hbrobotics/86DA0613-654D-49B8-9928-665CC428188E%40gmail.com.

[Nathan Lewis]

Also, if your experience was with the Pi 4, those have TERRIBLE signal quality on the USB 3 ports. I could never get it to maintain a USB 3 connection with the camera, yet the same cable worked flawlessly with my laptop. The Pi 5 has seemingly worked thus far.

- Nathan

To view this discussion visit https://groups.google.com/d/msgid/hbrobotics/CADhdHjXibetk_EWTQ9B1nBH%3DDg4UUw8e5gAXnfE6e2beJiXGgA%40mail.gmail.com.

[Nathan Lewis]

Actually, it does seem plausible the Oak-D lite has issues with cable reverse.

I found this fun thread on the Luxonis forums:

https://discuss.luxonis.com/d/6443-usb-30-connection-issue-with-oak-d-lite/12

Sounds a lot like what you’re describing. 

- Nathan

To view this discussion visit https://groups.google.com/d/msgid/hbrobotics/86DA0613-654D-49B8-9928-665CC428188E%40gmail.com.

[Duane Nielsen]

Note you need to be careful about the USB bus bandwidth also.  If multiple cams are plugged into the same bus, you may not have enough bandwidth to support the stream.

1 controller 1 bus is a good idea.

Duane

To view this discussion visit https://groups.google.com/d/msgid/hbrobotics/CADhdHjXibetk_EWTQ9B1nBH%3DDg4UUw8e5gAXnfE6e2beJiXGgA%40mail.gmail.com.