Eeprom data corrupted?

Did you make on the board all the necessary changes for it to work with the RPi (https://wiki.analog.com/resources/eval/user-guides/ad-96tof1-ebz/ug_rpi#modifying_the_ad-96tof1-ebz_to_work_with_the_rpi)?

Which board revision do you have, B or C?

Hi, thank you for your answer,

I have a rev. C.
I did the changes that must be made on the camera board: switch S1 to 1 and for S5 only 1 to OFF.
I did the power on sequence steps too.

I have the following log now : 

pi@raspberrypi:~/Desktop $ ./aditof-demo.sh
I0505 22:13:47.918064 804 device_enumerator_raspberry.cpp:47] Looking for devices on the target
I0505 22:13:47.921128 804 device_enumerator_raspberry.cpp:80] Looking at: /dev/video0 for an eligible TOF camera
I0505 22:13:47.922816 804 system_impl.cpp:62] System initialized
I0505 22:13:47.922966 804 camera_96tof1.cpp:78] Initializing camera
I0505 22:13:47.923008 804 local_device.cpp:158] Opening device
I0505 22:13:48.700927 804 calibration.cpp:261] EEPROM calibration data size 64940 bytes
I0505 22:13:54.495831 804 camera_96tof1.cpp:95] Camera initialized
I0505 22:13:54.500403 804 local_device.cpp:253] Starting device
I0505 22:13:54.586835 804 camera_96tof1.cpp:129] Chosen mode: near
I0505 22:13:54.586958 804 camera_96tof1.cpp:170] Camera range for mode: near is: 250 mm and 800 mm
I0505 22:13:54.588925 804 camera_96tof1.cpp:180] Found firmware for mode: near
I0505 22:13:54.588973 804 camera_96tof1.cpp:183] Firmware size: 10440 bytes
I0505 22:13:55.875015 804 camera_96tof1.cpp:199] Camera calibration parameters for mode: near are gain: 1 offset: 0
W0505 22:14:04.854745 808 local_device.cpp:830] select timeout
W0505 22:14:04.855360 808 camera_96tof1.cpp:329] Failed to get frame from device

It is the same result I obtained using a jetson nano (I wrote the issue last week here : ://github.com/analogdevicesinc/aditof_sdk/issues/337)

The changes that you made to the board seem fine, there isn't anything else that needs to be done

The "select timeout" message shows there isn't any data coming from the camera. Do the 3 green LEDs on the top right corner of the camera board light up after you start the software?

Have you tried using the camera with a DragonBoard? That would help identify if there's an issue with the camera or there's something wrong with the connection to the RPi or the Jetson boards. 

ACozma said:

Do the 3 green LEDs on the top right corner of the camera board light up after you start the software?

Yes it does.

ACozma said:

Have you tried using the camera with a DragonBoard?

No but on a Jetson Nano.
I have the same log than on Raspberry Pi.
It seams that the select method always return 0 (timeout).

If the 3 LEDs turn on it means that the camera is programmed correctly. There might be a few things that could cause this issue:

 - the ribbon cable use to connect the camera is broken

 - the ToF camera might be damaged - could you please check visually the board to see if there's anything suspicious - broken components, any damage to the PBCs?

 - the ToF camera power supply could be faulty and it cannot withstand the current needed by the lasers

Are you using the SDK from the master branch or did you get one of the releases? As a last resort, if you're using the master branch, I'd suggest trying with the latest release.

Thank you very much for your reply.

- For the ribbon : I have tested it with a raspberry camera and it works

- There is nothing suspicious on the board : no scratch, no impacts on the PCB and components

- For the power supply : I have replaced it with another that I often use and the result is the same. 

- On jetson I use master branch, on the raspberry I use this sdk image camo.githubusercontent.com/.../68747470733a2f2f696d672e736869656c64732e696f2f62616467652f72656c656173652d6c61746573745f73645f636172645f696d6167652d626c75652e737667

I can't use the latest release (v1.3.0) on jetson (no jetson folder in sdk/src/target).

Any other ideas?

Thank you for your reply.

- For the ribbon : I have tested it with a raspberry camera and it works

- There is nothing suspicious on the board : no scratch, no impacts on the PCB and components
- For the power supply : I have replaced it with another that I often use and the result is the same. 

- On jetson I use master branch. I can't use the latest release (v1.3.0) because there is no jetson folder in sdk/src/target.

  On the raspberry I use the latest release image
  I have just try with the v1.3.0 branch, it is almost the same log :

pi@raspberrypi:~/Desktop $ ./aditof-demo.sh
I0507 17:19:03.336621 3655 device_enumerator_raspberry.cpp:47] Looking for devices on the target
I0507 17:19:03.338390 3655 device_enumerator_raspberry.cpp:80] Looking at: /dev/video0 for an eligible TOF camera
I0507 17:19:03.340169 3655 system_impl.cpp:62] System initialized
I0507 17:19:03.340320 3655 camera_96tof1.cpp:78] Initializing camera
I0507 17:19:03.340359 3655 local_device.cpp:156] Opening device
I0507 17:19:04.120066 3655 calibration_96tof1.cpp:163] EEPROM calibration data size 64940 bytes
I0507 17:19:09.914160 3655 camera_96tof1.cpp:95] Camera initialized
I0507 17:19:09.920964 3655 local_device.cpp:251] Starting device
I0507 17:19:09.992377 3655 camera_96tof1.cpp:129] Chosen mode: near
I0507 17:19:09.992539 3655 camera_96tof1.cpp:170] Camera range for mode: near is: 250 mm and 800 mm
I0507 17:19:09.994480 3655 camera_96tof1.cpp:180] Found firmware for mode: near
I0507 17:19:09.994532 3655 camera_96tof1.cpp:183] Firmware size: 10440 bytes
I0507 17:19:11.280812 3655 calibration_96tof1.cpp:376] Camera calibration parameters for mode: near are gain: 1 offset: 0
I0507 17:19:11.281221 3655 calibration_96tof1.cpp:388] Camera intrinsic parameters:
fx: 391.248
fy: 387.814
cx: 325.193
cy: 242.345
W0507 17:19:16.949777 3662 local_device.cpp:863] select timeout
W0507 17:19:16.950568 3662 camera_96tof1.cpp:316] Failed to get frame from device

Any other ideas?

The log is exactly as it should be up to the select timeout error. If nothing suspicions comes up in the dmesg output, as suggested here (https://github.com/analogdevicesinc/aditof_sdk/issues/337), the only remaining reason for this behavior would be a hardware issue and you should contact your sales representative to replace the system.

Hi,

Thanks for your reply.

I think it is not a hardware problem because when  I started the adi tof demo exemple
with dmseg --follow command in background, and I have the following output:

[ 6.467920] addi9036 0-0064: addi9036 detected at address 0x64
...
47.902774] addi9036 0-0064: could not write to register 0
[ 47.902783] addi9036 0-0064: could not write to register 0
[ 47.902788] addi9036 0-0064: could not write to register 0
[ 47.902793] addi9036 0-0064: could not write to register 0
[ 47.902797] addi9036 0-0064: could not write to register 0
[ 47.902802] addi9036 0-0064: could not write to register 0

Does it mean that something is blocking writing to the register 0?

Hi,

Thanks for your reply.

I think it is not a hardware problem because when  I started the adi tof demo exemple
with dmseg --follow command in background, and I have the following output:

[ 6.467920] addi9036 0-0064: addi9036 detected at address 0x64
...
47.902774] addi9036 0-0064: could not write to register 0
[ 47.902783] addi9036 0-0064: could not write to register 0
[ 47.902788] addi9036 0-0064: could not write to register 0
[ 47.902793] addi9036 0-0064: could not write to register 0
[ 47.902797] addi9036 0-0064: could not write to register 0
[ 47.902802] addi9036 0-0064: could not write to register 0

Does it mean that something is blocking writing to the register 0?

Register 0 is not a valid address, which means there's something suspicious with the EEPROM data.

Best would be if you could read back the EEPROM data and post it here so that we can analyze it.

These are the steps to read the data from the EEPROM:

 - modify the first_frame example by adding the following code at this line: https://github.com/analogdevicesinc/aditof_sdk/blob/master/misc/chicony_firmware.bin

uint8_t data_buffer[65000];

std::shared_ptr<aditof::DeviceInterface> device = camera->getDevice();

device->readEeprom(0, data_buffer, 65000);

- save the data_buffer to a file

Hi,

Here is the file containing the eeprom data.
Is the format ok?
I don't know how to analyze it.

eeprom_data.txt

The EEPROM content doesn't show any obvious problems. Most probably the issue  is hardware related.

Since we exhausted all the options, as a final attempt you can try to reprogram the EEPROM with a standard image from here: https://github.com/analogdevicesinc/aditof_sdk/blob/master/misc/ad96tof1_eeprom.bin

Can use the device->writeEeprom function to write the eeprom.

Thank you for your reply.

I don't know which parameters to pass to the method LocalDevice::writeEeprom(uint32_t address, const uint8_t *data, size_t length).  I guess that :

• address is 0
• data : the content of ad96tof1_eeprom.bin 
• length : the content length

So my method will looks like :

void WriteEeprom() {

    std::ifstream stream("../../misc/ad96tof1_eeprom.bin", std::ios::in | std::ios::binary);
    std::vector<uint8_t> data((std::istreambuf_iterator<char>(stream)), std::istreambuf_iterator<char>());
    device->writeEeprom((uint32_t)0, (uint8_t *)data, (size_t)data.size());
}

Is that right?

Yes, this is correct.

I have re-rewrite the eeprom as mentioned (code here).
I have still the same timeout error with select. 

Then the only remaining reason for this behavior would be a hardware issue and you should contact your sales representative to replace the system.