Hello, TI engineer:
I want to ask about the firmware version difference of DLPC3479.
We started mass production of DLPC4710 optical engine in 21 years, and the firmware version used has always been v8p0p1. The optical machine is designed based on EVM design.
This version has been given to a customer who is doing AOI and has been used stably without any problem. But since last September, we switched the firmware version to v8p1p0, and then this customer used it to test the new AOI project, and found some problems.
Work status of this project:
They will use four optical engines in one device, all of which use v8p1p0 firmware, and the four optical engines and a camera are integrated on a mechanical arm. Each optical opportunity receives a trigger command separately and then projects four 1D pictures. The trigger command signal is I2C signal sent by MCU, and the trigger command means "36 9E0000". The I2C timing is tested to be up to standard.
Four optical engines will receive trigger instructions to project pictures in sequence, and one camera will collect the projected pictures synchronously. After one area is projected, the mechanical arm will quickly move to the next area for projection. The time interval between the two areas is about 400 ~ 500 ms. That is, the time interval of receiving the trigger command for a single optical engine is about 400 ~ 500 ms.
Repeat the above process and run the aging test. Within 10 minutes to several hours after starting operation, any one of the optical machines may have the following problems.
Problem description:
The optical engine can receive the trigger command, but does not project the picture, and the pin M_GPIO5 is always low. Connect the USB cable and use the GUI to operate "RUN ONCE" in "Light Control". The phenomenon is the same. However, when operating "Test Pattern" or "Splash Image" in "display", there are projected pictures. But then "RUN ONCE" is operated, and there is still no projected picture.
Finally, restart the power supply of the optical machine and it will return to normal.
Problem analysis:
After investigation, it should not be the problem of power supply or equipment hardware or optical engine hardware, because the brightness of the light source used is the lowest, and switching to "display" can project after the problem occurs, which proves that the hardware of optical engine is still in normal state.
Then compared with other devices they have made before (which passed the aging test), this device is the most similar to the device with problems. I found that the biggest difference is the firmware version. The firmware version used by the optical engine on this device is v8p0p1.
Then we replaced all four optical engines on the faulty equipment with v8p0p1 firmware. Do not change the hardware on the device. Then I continued to run the aging test and found that the problem disappeared.
My question is as follows, please ask the engineer to help me analyze and give a reasonable explanation.
1. Compared with v8p1p0 and v8p0p1, what are the upgrade changes? Looking at DLPC 3479 Firmware V8.1.0 Release Notes, there is no mention of related issues.
2. Why does v8p1p0 firmware have this problem, that is, after running for a period of time, the "light control" cannot project, but the "display" can project. At present, only one customer has found this problem. This v8p1p0 firmware has passed our internal verification test and found no such problem. This problem has not occurred in other customer application scenarios.
3. The latest version released by TI a few days ago is v8p1p1, but our company has not switched. When posting, I saw that the latest version was v8p3p0. However, we still can't confirm whether the latest version of v8p3p0 0will have similar problems, so please ask the engineer to give the above analysis and countermeasures, or make it clear that the latest version of V8P3P0will not have similar problems.
Thanks!
