Probing a driver on JetPack 5.0.2 and JetPack 5.1.1

Hi,
I’m developing a camera driver on JetPack 5.1.1 for a Jetson Orin AGX 32GB devkit.

The sensor board needs:

  • reset signal,
  • a 1v8 regulator (provided by default by the Jetson).
  • a 2v8 regulator.
  • a 3v3 regulator (provided by default by the Jetson).

When I attached the sensor board to the Jetson platform using JetPack 5.1.1, I enable the reset GPIO, then enable the 2v8 regulator, and finally try to check the i2c bus, but I’m not able to detect the sensor board address. I’m just able to detect the EEPROM address (0x50), but randomly. Sometimes it shows up, other times it does not.


    0  1  2  3  4  5  6  7  8  9  a  b  c  d  e  f
00:          -- -- -- -- -- -- -- -- -- -- -- -- -- 
10: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
20: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
30: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
50: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
60: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- 
70: -- -- -- -- -- -- -- --

On the other hand, when I use JetPack 5.0.2 and do exactly the same process as above, I’m able to detect the EEPROM address and other components addresses from the sensor board. I’m also able to detect the sensor address itself (0x36).

   0 1 2 3 4 5 6 7 8 9 a b c d e f
00:     -- -- -- -- -- -- -- -- -- -- -- -- --
10: 10 -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
20: -- 21 22 23 -- -- -- -- -- -- -- -- -- -- -- --
30: -- -- -- -- -- -- UU -- -- -- -- -- -- -- -- --
40: -- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
50: 50 -- -- -- -- -- -- -- -- -- -- -- -- -- -- --
60: -- -- -- -- -- -- -- -- -- -- -- -- 6c 6d 6e 6f
70: -- -- -- -- -- -- -- --

Is the integrity of i2c signals different between these two JetPacks versions?

Best,

JDiego Delgado
Embedded SW Engineer at RidgeRun
Contact us: support@ridgerun.com
Developers wiki: https://developer.ridgerun.c om/
Website: www.ridgerun.com

hello jdiegodelgado,

it likely to be power issues for unable to detect the sensor board.
is it possible to check all those power were given correctly? is there other drivers affect them?

Hi,

We measured the power signals and all of them look fine.

Best,

JDiego Delgado
Embedded SW Engineer at RidgeRun
Contact us: support@ridgerun.com
Developers wiki: https://developer.ridgerun.c om/
Website: www.ridgerun.com

Are those power rails enabled in the same order? It’s only a slight possibility that it matters, but sometimes order of enabling rails matters.

@linuxdev ,

They have been enabled in the same order. We tried to avoid editing the JetPack, when we connect the sensor board using JetPack 5.0.2 to the Jetson platform, we are able to watch the addresses of some components of the sensor board (EEPROM, RETIMERS…), but when we do exactly the same thing using JetPack 5.1.1, we are not able to detect any of them.

So the question is what could’ve changed? What could we check on JetPack?

Best,

Best,

JDiego Delgado
Embedded SW Engineer at RidgeRun
Contact us: support@ridgerun.com
Developers wiki: https://developer.ridgerun.c om/
Website: www.ridgerun.com

I don’t know if device tree is read yet for this stage, but if it is, then is the device tree read at the same time or sequence at or near this power rail stage? I doubt device tree even takes part at this point, but maybe some other firmware does. If the setup needed for those power devices differ in any way, then that can be a problem. If not, then you’ve isolated this to hardware using the power states rather than the power states themselves.

There is no update from you for a period, assuming this is not an issue any more.
Hence we are closing this topic. If need further support, please open a new one.
Thanks

hello jdiegodelgado,

would like to confirm your steps in detail in order to understand the difference.