depthai-hardware
depthai-hardware copied to clipboard
luxonis
OAK-D Pro-PoE
Preorders available: OAK-D-Pro-POE
Start with the why:
OAK-D-PoE was our “make it exist” product. It served as a pathfinding design to see what is the community's reaction to it and what could be improved. It has been quite well received. With specific feedback on what would be nice in a Pro version.
The OAK-D-PRO-POE will be the next step and an upgrade to the original OAK-D-PoE, implementing user feedback and improvements through experience gained from previous design.
Overall folks are happy with OAK-D-PoE. But from the requests we do get, they are nearly all about illumination. And specifically the requests ask for:
- IR laser dot projection for active depth (allowing low-light and no-light depth sensing)
- Blanket IR LED illumination (allowing low-light and no-light computer vision).
Everyone loved that OAK-D-PoE is IP67 sealed (so we're keeping that) but many desired an M12 connector instead of the RJ45 IP67 gland, which is a bit large - and more importantly is the reason OAK-D-PoE is a bit bigger and heavier than some desired.
- M12 connector - smaller and if used, the whole device can be smaller and lighter
- Weight - Some found OAK-D-PoE too heavy
- Size - Some found OAK-D-PoE too big
And the final common request was some form of external IO for being able to trigger or be-triggered by external equipment:
- External IO connectivity is desired
Move to the how:
Connectivity and power:
Just like the OAK-D-POE design, the OAK-D-PRO-POE will have and ethernet connection for data and PoE for power. But that is where similarities end, as the OAK-D-PRO-POE will feature a more robust and industrial connector type called M12.
To afford external IO connectivity, we will implement an M8 or similar connector. Likely with USB host-support from the Myriad X so that external USB devices (like thermal cameras, or IO expanders) can be used.
Form factor:
Form factor will be heavily based on the OAK-D-PRO design (https://github.com/luxonis/depthai-hardware/issues/114) with it's sleek small case. Afforded by the M12 connector, it is now possible to make this design a lot smaller. And the design will still have the same IP67 rating as was with OAK-D-PoE.
Other features:
Implementing IR led and dot projector will be the same as in the OAK-D-PRO model (details in https://github.com/luxonis/depthai-hardware/issues/114), as it is already in the testing phase and shows a lot of promise.
Move to the what:
A small, lightweight, cost-reduced version of OAK-D-POE that is still IP67 sealed, that has IR laser dot projector and IR illumination LED on-board (Pro version), and IP67-sealed IO connectivity.
- Optics/Illumination/Active Depth: Same as https://github.com/luxonis/depthai-hardware/issues/114.
- M12 X-coded for Gigabit Power over Ethernet
- M8 connector for GPIO, hardware multi-device sync, and direct external power (instead of PoE).
We intended to put these on the M8:
- power input (5V) as an option instead of PoE, or 5V power output for some external circuitry
- USB2 (D+ and D-)
- camera IOs: FSIN (frame sync) and STROBE (for driving a flash)
- 2 other aux IOs, capable of UART, I2C or GPIO
We're going to do a wide FOV variant of OAK-D Pro-PoE as well. Same formfactor, just slightly different front cover.
Hi @artificiel ,
We'll likely have these OAK-D Pro-W-PoE in March. We would have them in January but likely these will be pushed by an extended (because of COVID) shutdowns in January/February in Asia where these are produced.
So March is a good guesstimate. We can add these to the Beta Store if you'd like to be able to pre-order to get some of the first prototype finished units. Just let us know. :-)
https://shop.luxonis.com/collections/beta-store
Thanks for the interest.
Hi ShawnSchaerer,
we have not yet done the thermal/stress testing on the design (we just recently got the boards), but we are expecting the unit to be able to have a temperature range from 0°C - 70°C at least.
It would be great If you could hit 80 or 85C
Hi ShawnSchaerer,
we have not yet done the thermal/stress testing on the design (we just recently got the boards), but we are expecting the unit to be able to have a temperature range from 0°C - 70°C at least.
What wavelength laser are you using? Are you keeping it near 832nm to mitigate blooming from the sun? Are you planning any options for something more powerful than a class 1? Maybe 3R?
Hello @robotwhispering , it's near 940nm (projector specs here). We don't have any plans for stronger lasers, as even with this one dots are quite visible at +15m. What would be the application for more powerful ones?
- M8 connector for GPIO, hardware multi-device sync, and direct external power (instead of PoE).
We intended to put these on the M8:
- power input (5V) as an option instead of PoE, or 5V power output for some external circuitry
- USB2 (D+ and D-)
- camera IOs: FSIN (frame sync) and STROBE (for driving a flash)
- 2 other aux IOs, capable of UART, I2C or GPIO
Is there additional information on the FSIN (frame sync) feature?
- What is the voltage/duration/type of signal for triggering the cameras?
- Will the frame sync be able to trigger both the rolling shutter RGB and global shutter IR cameras?
- What will be the tolerance of the frame sync signal? Does it have to be at a set hertz/framerate? Or can it be triggered arbitrarily as long as there is enough time between frames?
Hello sam598,
sorry for a late response.
What is the voltage/duration/type of signal for triggering the cameras?
- As we have a buffer on this signal, the maximum voltage tolerance of the pin is 5V.
- There is a minimum on time for FSIN signal and depends on the clock provided to the camera. The range of the min on time is between 78ns - 833ns (@alex-luxonis can correct me if wrong here)
- You need to apply a rising edge to the FSIN pin to start the image capture.
Will the frame sync be able to trigger both the rolling shutter RGB and global shutter IR cameras?
The FSYNC signal is routed to both Stereo cameras, RGB camera and also routed to the MyriadX VPU.
What will be the tolerance of the frame sync signal? Does it have to be at a set hertz/framerate? Or can it be triggered arbitrarily as long as there is enough time between frames?
@alex-luxonis would be able to answer this?
NOTE!!: I just want to point out that we will be actually changing the M8 pinout in the next boards revision, to have a more robust and standard interface, that the industry is using.
New pinout will be:
- power input (5V) as an option instead of PoE, or 5V power output for some external circuitry
- USB2 (D+ and D-)
- Opto isolated FSYNC input (input circuitry reference)
- Opto isolated Strobe output (output circuitry reference)
- GPIO
With regards to the above questions the major change will be the Input voltage tolerance which would now be 5V - 24V.
Let me know if you have any further questions.
Nejc
Hello @GNejc and @sam598, I'm really interested in this feature.
What will be the tolerance of the frame sync signal? Does it have to be at a set hertz/framerate? Or can it be triggered arbitrarily as long as there is enough time between frames?
Related to this topic, I need more information on the following two points.
- Is arbitrary timing trigger supported?
- Does OAK-D Pro-PoE provide trigger mode for burst multiple frame predefined frame rate with color and depth sync?
Any additional information is helpful, thanks!
In general two operation modes are possible where an external sync/trigger signal is used:
-
Continuous streaming with external frame sync input (also named in sensor documentation: "slave" mode) - the sync signal has strict timing requirements (usually in the range of few microseconds), its rate needs to follow closely the configured sensor frame rate, otherwise would cause the sensor to output a bad frame. We have a diagram of the sensor operations here: https://docs.luxonis.com/projects/hardware/en/latest/pages/guides/sync_frames.html?h#synchronizing-frames-externally For the global shutter OV9282(mono) and OV9782(color), it's possible to configure the period of the external signal also as a multiple of the sensor frame period, so for example: sensor 30Hz, external fsync 10Hz. But not clear if the same applies to IMX378, we'll have to test.
-
Snapshot trigger mode. Supported at least by OV9282/OV9782 (we need to confirm for IMX378, or try to emulate it from firmware), the sensor stays in idle, then when it receives a pulse on the pin, it starts streaming a burst of frames (configurable), at the preconfigured rate (
.setFps(..)). Note that for OV9282/OV9782, the first frame in the group has degraded quality, and depending on application's needs, it may have to be discarded.
@alex-luxonis Thank you for sharing great insights. I can't wait to try OAK-D Pro PoE! If you have update for IMX378 also support Snapshot mode, please share us! I have one more question. Would you mind to share us the recommended I/O cable for OAK-D Pro PoE? I assume M8 female cable(8 pin) is the one but I want to start with reliable option. So which model or product would be recommended? Thank you!
@YoshiyasuIzumi
sorry for the late reply,
In the latest board revision we are switching the M8 connector from Male to Female and thus would require a new cable as well. In addition we are also changing the pinout of the M8 connector (shown in the image below). The new pinout will have:
- STROBE output and FSYNC input be optically isolated (same type of IO as Basler)
- 1x GPIO with the capability of booting to USB with a pull up resistor. This IO is the same as it was with the previous revision.
As this is currently in the development/testing stages, we do not have any functional devices to attach on these yet. We have developed a testing board which has the signal pins exposed on a header for testing porpoises. We could arrange for a board to be sent along with your OAK-D-PRO-POE.
Let me know if you have any other questions.
@GNejc Thank you for your reply. Please let me confirm my understanding. I thought I only need to connect I/O cable from OAK-D-PRO-POE to signal source(5V) in order to provide a trigger to the camera. To provide the trigger to the camera, do I need to use this "testing board" for this purpose? And one more thing. I checked Basler's M8 I/O cable and it seems to have only 6 pins. Does this cable match to the pins on the camera? Thanks!
I'm so grateful I got to edit my response before you read it, because I obviously didn't pay attention to your comment when I started rattling off a bunch of stuff you already know.
When Basler moved forward with the Ace2 series, they changed to the 6 pin M8 connector instead of the "push-pull" Hirose connectors.
It's common to enlist a pull-up resistor (or pulldown) for opto-isolated io so that the signal didn't just "float' in a near constant state of charge. You either have to pull it up to high when you want it to go high, or sink it to 0 when you want a low signal. I recall the voltages were much lower compared with the GPIO (which I think was 3.3V} I can't remember what ohm resistor I used for my control signal, but they have documents for each camera that give you the pinout color code, as well as what resistors would keep the current low enough at each voltage step. But to be honest, the documentation from Lucid, for their Triton line of Gige cameras has even better documentation, and it's basically the same camera as the basler ace2. I was controlling heater resistors and wipers for the enclosures that the cameras were in.
I'm not sure if it'll be any use to you, but I'm attaching a PDF from Lucid explaining how to wire the IO up, including a chart for which resistors you might want to use for each voltage range to keep current from getting out of hand. It really helped me.
And you're right, Basler uses the 6 pin version of the M8. But Lucid uses an 8 pin version similar to Luxonis. I don't know if the pinouts will match up exactly, but they'll probably be pretty similar. Triton+Camera+Technical+Manual+v1.0.0 (1).pdf
The relevant information starts on page 19 and continues through page 23.
Oh, and a very common reason for using female connectors on the device, instead of male, is because it's much safer for the device and anything else that will come in contact with it. Male pins can much more easily be shorted inadvertantly and cause harm and damage.
Hi robotwhispering, Sorry for my late response and thank you for your insighs! In my configuration, original trigger logic level is 24V. So I need to convert it to 5V logic level. To do that, I'm thinking of using SSR or Isolated photocoupler. For this scenario, if I choose 6 pins M8 cable, I should pull-down(or pull-up) the trigger signal for avoiding channel float. Is my understanding correct? But I found M8 cable with 8 pins in Lucid vision labs from your reference document, so I guess I shall start with this option.
Please let me know if you have any other comment on this. Thank you.