Kinect 2, Electrical Specification

[Aurel W.]

Hi,

can we gather some data on the electrical specification and the pinout
of the connector? I have no access to a new Xbox One and even have
troubles finding any decent images of the plug on the web (just
annoying gadget bloggs).

What are the specs of the provided PSU for the k4w?
Any ideas on the pinout of the Connector?

It would be also interesting what kind of socket/plug they are using.
I am fairly sure someone can buy sockets for those to build an
adapter.

I will try to connect a Kinect provided with an xbox one to PC in a
view days so any information at this point would help a lot.

aurel

[Ryan Gordon]

Hi Aurel,

I don't have a K4W but I have a Xbox One Kinect. Me and a friend traced the pinouts as follows:

https://i.cloudup.com/0GXiqsIcpa-3000x3000.png

Here is a pictures of the side of the connector that goes into the Kinect 2 itself: https://i.cloudup.com/hjjAJLnN1d-3000x3000.jpeg

It's in the same Dimension as the USB 3.0 Powered Type B but it has one extra pin on the bottom

Also a picture of the back of female port on the back of the Kinect 2.0: https://i.cloudup.com/O1ubdmydWS-3000x3000.jpeg

On the other side of the cable is a custom/proprietary connector Microsoft used which doesn't fit anything I can find online

[Aurel W.]

Thanks a lot Ryon!

So judging from your pin-out it seems like the plug has 4 pins on the
top and bottom each on the old USB 2.0 type-B part? It looks like it
would connect into a 3.0 type B socket but has an incompatible number
of pins and placement.

I guess I will have to cut the entire cable and resolder it to a
type-A plug. I am still not sure about those pins:

Pin 10: GND - is that GND for VBUS?
Pin 11/12: Are those both GND for the 12V?
Pin 9: Any idea what this might be used for? It may be a good idea to
have a pull up or down resistor here.

aurel

> --
> You received this message because you are subscribed to the Google Groups
> "OpenKinect" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to openkinect+...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.

[Florian Echtler]

I've had a look inside the adapter box, and it seems there's also a 3.3V
supply around... so maybe that's the grey wire. See pic below:

https://plus.google.com/u/0/photos?pid=5954639771932774546&oid=115708687434865126342

BR, Florian

--
SENT FROM MY DEC VT50 TERMINAL

[Aurel W.]

Sounds interesting, but can you share pictures with something
different than Google+?

[Florian Echtler]

Yeah, sorry, here you go. Note also the "USB2 ENABLE" text near to
jumper J2, will perhaps try to bridge this later today and see what happens.

BR, Florian

[Aurel W.]

Ok, that looks way more complicated than expected.

Some good high resolution images and IC labels would be interesting.
It's hard to tell any details from that image. Also the backside of
the PCB would be interesting.

I guess we would need a closer look on the schematics to see what is
really going on. It's hard to come up with any good reason for all
that extra hardware. Maybe it provides additional in system debugging,
with an extra serial link down to kinect.

For the basic operation I hope just USB and power supply will be
sufficient. Another thing I can think of is some sort of hardware
level protection to protect the bare kinect ever working on usb host.

I worry a lot of damaging my device when I hook it up to any voltage
source because the pinout is not clear to me at this point. It would
be really helpful as a first step that someone with a k4w adapter
actually measures the pinout and voltage levels. It's obvious that
there is a 3.3V voltage regulator present because of all those extra
ICs. I doubt that those 3.3V are passed over the long cable, that
would be pretty unusual.

I really wish I would have access to a xbox one or a k4w adapter to
just hook up an oscilloscope to see what is going on for real. For now
I pretty stuck with just having the kinect.

aurel

[Florian Echtler]

On 09.12.2013 17:07, Aurel W. wrote:
> Ok, that looks way more complicated than expected.

Yes, I was a bit surprised, too.

> Some good high resolution images and IC labels would be interesting.
> It's hard to tell any details from that image. Also the backside of
> the PCB would be interesting.

I don't have a macro lens, but I wrote down the labels of the 3 major ICs.

top right: TPS2590 TI31I (Texas Instruments Hotplug Controller)
center: SMSC USX2072 (some kind of USB 2.0/3.0 hub device AFAICT)
bottom left: 502CP TI21K AJFF (couldn't find out what this is)

The rest of the ICs are op-amps or linear regulators AFAICT.

Backside is mostly large power supply traces, no components. I assume
it's a 4-layer PCB...

> I guess we would need a closer look on the schematics to see what is
> really going on. It's hard to come up with any good reason for all
> that extra hardware. Maybe it provides additional in system debugging,
> with an extra serial link down to kinect.
>
> For the basic operation I hope just USB and power supply will be
> sufficient. Another thing I can think of is some sort of hardware
> level protection to protect the bare kinect ever working on usb host.
>
> I worry a lot of damaging my device when I hook it up to any voltage
> source because the pinout is not clear to me at this point. It would
> be really helpful as a first step that someone with a k4w adapter
> actually measures the pinout and voltage levels. It's obvious that
> there is a 3.3V voltage regulator present because of all those extra
> ICs. I doubt that those 3.3V are passed over the long cable, that
> would be pretty unusual.

Nevertheless, there's a 3.3V test point right next to the Kinect
connector. Might just be there for the USB IC, though.

> I really wish I would have access to a xbox one or a k4w adapter to
> just hook up an oscilloscope to see what is going on for real. For now
> I pretty stuck with just having the kinect.

I'll see if I can figure out the power traces tomorrow.

Florian

[Florian Echtler]

BTW, I just figured out that the "502CP" chip is also from TI, and it's
an USB 3 signal conditioning IC (e.g. for longer cables). For the SMSC
IC, on the other hand, I simply can't find a datasheet or even a product
brief...

Florian

[Aurel W.]

Another thing we should consider is the length of the kinects cable
and the USB 3.0 cable attached to break out box. Of course it is
necessary to repeat the USB 3.0 signal to stay within specs.

I am pretty sure that the IC on the bottom left (502CP TI21K AJFF) is
some sort of redriver (repeater). The closest thing I found is the TI
SN65LVPE502CP Dual Channel USB3.0 Redriver/Equalizer. It's also a 24
PIN package. Imho that makes totaly sense also giving the position on
the PCB.

The TPS2590 TI31I could be there just for development and debug
measuring. It just acts as a circuit breaker when a given threshold
current is reached.

So far not much to worry about.

aurel

[Aurel W.]

> BTW, I just figured out that the "502CP" chip is also from TI, and it's
> an USB 3 signal conditioning IC (e.g. for longer cables). For the SMSC
> IC, on the other hand, I simply can't find a datasheet or even a product
> brief...

Ok, same conclusion here :)

About the SMSC USX2072. It would make sense to put an extra USB 2.0
host controller onto the development board. It could provide in system
programming or other low level access to the hardware. The header on
the top and the jumper "USB2 ENABLE" would be strong sings for that.

So far it seems like it could work to just hook it up directly to USB
and provide supply voltage. Given the correct pinout ;)

aurel

[Joshua Blake]

I bet if you open up an Xbox One you'd find similar chip setups in there to handle the longer cable.

One of the changes made between Kinect for Xbox 360 and Kinect for Windows v1 was making the cable shorter and adding an additional USB inline dongle of some type. These allowed it to work on all the various PCs and USB chipsets. 

I suspect that the final Kinect for Windows v2 sensor itself will have similar changes that eliminate the need for many of the breakout box chips. I doubt that anything in the breakout box is security related. That would just take way too much effort that could be used on developing features instead.

> Aurel wrote:

> So far it seems like it could work to just hook it up directly to USB

and provide supply voltage. Given the correct pinout ;)

Agreed, it would probably work with Xbox One Kinect in this way, although it might be slightly out-of-spec and only work with certain USB chipsets. Be careful not to fry your Kinect, though!

Thanks,

Josh

[Aurel W.]

Any news on the pinout or cable coloring? Florian, have you managed to
trace those pins?

aurel

[Aurel W.]

I opened up my kinect and did some measurements. Please note I HAVEN'T
tried this yet. I will solder an USB 3.0 plug to it this evening.

Brown wire: 12V
Gray: GND for 12V
Red: 5V
Black: GND for 5V (USB GND)
White/Black: very thin wire, this is definitively not GND, but any
auxiliary signal.

Gray and Black 5V and 12V GND are shortened. The metal sheets inside
the kinect are grounded as well.

Brown to Gray resistance: 495 Ohm.

aurel

[Ryan Gordon]

Hi Aurel,

That is great work! Just to confirm based on your analysis, it looks like the GND denotation for the White/Black wire in my diagram is wrong and I could clarify the denotation for the kinect specific wiring diagram based on your analysis. Do you have any idea what type of auxiliary signal the white/black wire might be for?

[Ryan Gordon]

I went ahead and updated the diagram with the information based on your analysis, does this look right? https://i.cloudup.com/gsEOYAqIDS.png

[Aurel W.]

>> That is great work! Just to confirm based on your analysis, it looks like
>> the GND denotation for the White/Black wire in my diagram is wrong and I
>> could clarify the denotation for the kinect specific wiring diagram based on
>> your analysis. Do you have any idea what type of auxiliary signal the
>> white/black wire might be for?

If Florian is tracing it in his breakout box, we will know. I guess
it's an ID pin.

> I went ahead and updated the diagram with the information based on your
> analysis, does this look right? https://i.cloudup.com/gsEOYAqIDS.png

I can confirm the pinout. About the GND, from it's pin placement it
would make more sense that Black is 12V GND and Gray is 5V GND, but
since they are shortened in the device (at least by measurement) it
won't matter much. At least in the standard USB color scheme, black is
5V GND. The wires have identical thickness as well.

aurel

[Florian Echtler]

On 19.12.2013 13:36, Aurel W. wrote:
>>> That is great work! Just to confirm based on your analysis, it looks like
>>> the GND denotation for the White/Black wire in my diagram is wrong and I
>>> could clarify the denotation for the kinect specific wiring diagram based on
>>> your analysis. Do you have any idea what type of auxiliary signal the
>>> white/black wire might be for?
> If Florian is tracing it in his breakout box, we will know. I guess
> it's an ID pin.

So I've finally managed to have another look at the breakout box, and it
appears this pin is somehow connected to the debug connector. All the
other pin connections are consistent with the exisiting diagram. I'd
suggest to just tie this wire to ground with a 10 kOhm resistor for the
moment.

Note that in the pictures below, I've used the pin labeling from the PCB
for consistency.

http://floe.butterbrot.org/external/k4w_breakout_pcb.jpg
http://floe.butterbrot.org/external/k4w_breakout_socket.jpg
http://floe.butterbrot.org/external/k4w_breakout_top.jpg

Note 2: I've attempted to upload these to the OpenKinect Wiki, but the
upload function seems to be broken. Can anyone with admin access have a
look, please?

Florian

>> I went ahead and updated the diagram with the information based on your
>> analysis, does this look right? https://i.cloudup.com/gsEOYAqIDS.png
> I can confirm the pinout. About the GND, from it's pin placement it
> would make more sense that Black is 12V GND and Gray is 5V GND, but
> since they are shortened in the device (at least by measurement) it
> won't matter much. At least in the standard USB color scheme, black is
> 5V GND. The wires have identical thickness as well.
>
> aurel
>
> On 19 December 2013 10:19, Ryan Gordon <rygo...@gmail.com> wrote:
>> I went ahead and updated the diagram with the information based on your
>> analysis, does this look right? https://i.cloudup.com/gsEOYAqIDS.png
>>
>>
>> On Thursday, December 19, 2013 1:06:50 AM UTC-8, Ryan Gordon wrote:
>>>
>>> Hi Aurel,
>>>
>>> That is great work! Just to confirm based on your analysis, it looks like
>>> the GND denotation for the White/Black wire in my diagram is wrong and I
>>> could clarify the denotation for the kinect specific wiring diagram based on
>>> your analysis. Do you have any idea what type of auxiliary signal the
>>> white/black wire might be for?
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "OpenKinect" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to openkinect+...@googlegroups.com.
>> For more options, visit https://groups.google.com/groups/opt_out.
>

[Aurel W.]

> http://floe.butterbrot.org/external/k4w_breakout_top.jpg
It looks like that pin8 is really for some auxiliary debug
communication, not just an ID pin. I guess 'EG' labels some sort of
capacitor? It is only used in the part of the circuit concerning the
test connector. From a brief look on the circuit it seems to be signal
level conversion for a com link. I guess the opamp connected to pin8
could be a MAX4470 or something alike. Is it labeled? Maybe it
provides in system programming/debugging, who knows.

The question remains, if there are any differences in the firmware of
the xbox and k4w devices.

aurel

[Florian Echtler]

On 20.12.2013 13:19, Aurel W. wrote:
>> http://floe.butterbrot.org/external/k4w_breakout_top.jpg
> It looks like that pin8 is really for some auxiliary debug
> communication, not just an ID pin. I guess 'EG' labels some sort of
> capacitor?

As far as I could tell, it can only be a capacitor, but I've never seen
'EG' as a label in circuits. R58 has about 25 Ohms, and I think C63 is a
decoupling cap between the +12V plane and GND, but I haven't yet
properly measured the rest of the connections.

> It is only used in the part of the circuit concerning the
> test connector. From a brief look on the circuit it seems to be signal
> level conversion for a com link. I guess the opamp connected to pin8
> could be a MAX4470 or something alike. Is it labeled?

I think it is, but I wasn't able to read anything.

> Maybe it provides in system programming/debugging, who knows.
> The question remains, if there are any differences in the firmware of
> the xbox and k4w devices.

Hm, one of the bulk transfers during initalization should probably
contain the firmware version?

BR, Florian

[Aurel W.]

> Hm, one of the bulk transfers during initalization should probably
> contain the firmware version?

Good idea. It seems like there hasn't been much investigation in that
initial bulk transfers yet. Dumping and comparing them for different
devices would be a first step.

aurel

[Joshua Blake]

Yes, the initial bulk transfer contains firmware version info. Don't
have it in front of me otherwise I would elaborate more.

Josh
From: Aurel W.
Sent: 12/20/2013 8:05 AM
To: openk...@googlegroups.com
Subject: Re: [OpenKinect] Re: Kinect 2, Electrical Specification

[Timon Paul]

Not sure if I just forgot to post or if the post got deleted (in that case sorry for posting it again).
These links might be helpful, quite a bit information about the used chipsets and nice high-res photos of the internals.
http://www.ifixit.com/Teardown/Xbox+One+Kinect+Teardown/19725
http://www.chipworks.com/en/technical-competitive-analysis/resources/blog/inside-the-xbox-one-kinect/