DLPDLCR230NPEVM: DLPDLCR230NPEVM resets consistently when displaying a fine checkerboard

Hello User,

Welcome to the E2E forum and thank you for you interest in DLP® technology!

Can you share what commands you are using to set up the checkerboard and to control the zoom? What resolution are you using to set the 100% zoom?

Regards,

Austin

Hi Austin,

Thank you for your quick response. I'm using the standard hdmi_timings which is:

hdmi_timings=1920 0 20 10 10 1080 0 10 10 10 0 0 0 58 0 125000000 3

I generate the image by using a small Python script:

import numpy as np
from PIL import Image

img = np.ones((1080, 1920, 3)).astype('uint8')*255

for i in range(0, 1080, 2):
    img[i] = 0
    
for i in range(0, 1920, 2):
    img[:,i] = 0

Image.fromarray(img).save('test.png')

When I open this image using the desktop GUI and mouse it opens it at a standard zoom level of 33%. If I then I double click on the image to go to fullscreen the EVM resets. I have found similar behavior with  slightly different images. All with pixelwise checkerboards. Some black and white, but also with colors.

Hello User,

Please allow us some time to try and recreate this behavior on our end. Thank you for your patience.

Regards,

John

User,

Thank you for this information.

Is it possible to check the properties of your created image before displaying? In particular, I would like to check your resolution. 

Regards,

Austin

Hi Austin,

I've checked my resolution settings using xrandr. It gave these outputs:

$ export DISPLAY=:0
$ xrandr
Screen 0: minimum 320 x 200, current 1920 x 1080, maximum 7680 x 7680
DSI-1 connected primary 1920x1080+0+0 (normal left inverted right x axis y axis) 0mm x 0mm
   FIXED_MODE    57.46*+


$ xrandr --listmonitors
Monitors: 1
 0: +*DSI-1 1920/508x1080/286+0+0  DSI-1

Hopefully this will help you.

Regards,

Koen

Hello Koen,

Thank you for sending this.

I am considering that the image may be an image processing issue. It will be a couple of days before I am able to test this on my own system, but in the meantime you could run a test to check.

Please try building an images with rows and columns with fully white pixels (i.e. make images with either of the for loops commented out). Also try this with varying incremental values, such as:

import numpy as np
from PIL import Image

img = np.ones((1080,1920,3)).astype('uint8')*255

for i in range (0,1080,4):
    img[i] = 0

# for i in range (0,1920,2):
#     img[:,i+1] = 0

Image.fromarray(img).save('test_stripes1.png')

Additionally, does the same failure occur when creating an all white or all black image?

Regards,

Austin

Hi Austin,

I've tried various settings and checked wheter the EVM resets or operates normally. I've tried settings with horizontal and vertical lines, with several line grey levels, and several distances between lines (skip). Here are the results:

Koen,

Thank you. This is very helpful data.

I will need to collaborate with my colleagues on this. Please expect a response some time in mid next week.

Regards,

Austin

Koen,

Do you have an oscilloscope at your disposal? I would like to know what activity you see on RESETz during these failure modes. RESETz can be probed on the resistor pad for R15 that is closest to U7.

Regards,

Austin

Hi Austin,

I've measured the pin you indicated. I did notice that the resistor on R15 is not placed and I was measuring directly on the pads. Also, most of the times I've probed the pad the DLP also went into reset when the probe made contact. I had to run the init_parallel_mode.py again while holding the probe on the pad to make it work again.

When I opened an image that would crash the DLP by using:
$ feh crash_image.png -Y -F

The signal on the oscillosope rapidly went from 2.5V to 0V. Then I closed feh by using ctrl+c and restarted the DLP with init_parallel_mode.py. When running that script the pin went from 0V to 2.5V again.

Koen

Hello Koen,

Thank you for running this test. It suggests that the DLP PMIC is triggering the reset.

Does the same behavior remain true for the images that you mention above that do not cause the reset, or does the 2.5V hold?

As a side note, I suspect that the contact is causing the DLP to go into reset because there is either a short being created on the board, or that the probe is pulling some current. Even if this is a small amount, the DLPA2005 has a max current output of 1mA on RESETz, so this may be the cause. You might consider blue wiring the resistor pad if this is challenging to hold in place.

Also, R15 should not be placed by default. It is a pull-up resistor that is not needed.

Regards,

Austin

After some discussion with our team, this issue sounds familiar to this ticket:

https://e2e.ti.com/support/dlp-products-group/dlp/f/dlp-products-forum/985312/dlpdlcr230npevm-evaluation-model-flashing-when-plugged-into-5v-power-supply

What power supply are you using for your EVM? What are its power ratings?

Regards,

Austin

Hi Austin,

The PSU I'm using is rated at 5.2A at 5V. When an image is opened that does not create a reset the RESETz pin is stable.

I've measured the Voltage on C33 during normal operation and I measure 4.9V. This does not seem to variate by more than 0.1V when when an error-causing image is opened.

Initially I was testing the the DMD because I was curious how you are able to get 4 pixels out of each mirror. That's why I created the pixelwise point and line images. Could you tell me a bit more about this method how this is done?

Regards,

Koen

Thanks for your response Austin. Last small follow-up question. Is it possible to disable the wobulation? I can't find this term in any of the datasheets of the EVM.