DLP2010EVM-LC: I flashed new firmware on to the EVM while it was running and now it acts as if the display board is not connected

Hello Kim,

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

Looks the flash might have been corrupted. Our expert will get back to you soon on this for further support. Appreciate your patience

Regards,

Akhil

Hello User,

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

What version of the FW are you using, is it the EVM FW or is it from the Pico FW Selector Tool?

I prefer to update the FW with the EVM connected to power but not displaying. I believe you are already doing that.

What type of output are you getting when the projector is turned on? Is it only failing to connect to the GUI or is it not projecting at all?

Regards,

John

Hello John,

The firmware I used to update initially was from the firmware section of this page

www.ti.com/.../DLP2010EVM-LC

The filename of the firmware is dlpc3470_fw_8.0.0.img

I'm not entirely sure what is loaded on the EVM currently but I last tried to load the backup I made of the original firmware back on to it.

The EVM is currently not projecting at all. There is no splash image or pattern on the DMD when SW1 is flipped. The LED lights for the light engine are not connected because custom illumination and optics are being used, but there is no voltage on the display board J2 pins. On the main control board when SW1 is flipped the PORJ_ON LED (D3) is on and stable, with RESETZ (D2) being off.

Hello User,

Would you be able to flashing the EVM with this firmware from the product page once more? Its possible that something went wrong in the process?

Also, can you say what activity you are seeing on HOST_IRQ (D1) when the board does not boot?

Regards,

Austin

Hello Austin,

I went ahead and re-flashed the EVM with the firmware on the product page after downloading a new copy. The EVM still behaves the same. HOST_IRQ (D1) is on and consistent whether or not SW1 is flipped on or off.

Hello User,

Can you give us more detail on this custom setup that you are running?

For illumination, is the EVM display board powering the LEDs or is there some external driver?

Are these the LEDs that came with the EVM or some of your own?

How is the DMD connected to the custom optics and the Display board?

Did the EVM work with the original engine and LEDs included?

Regards,

John

John,

For Illumination we are using an external laser that has its own power supply. The beam is de-collimated and appropriately angled to go through the optics and blaze the power away from the 0th order. Here is a picture of how the DMD is held in place.

https://imgur.com/a/EQ4cUk5

The board is held in place with some 90 degree brackets that are on an optical breadboard. The optics are attached to the same optical breadboard, but kept a distance from the DMD. The optics and DMD are not connected together, there is are physical gaps between the DMD, Illumination laser, and the projection optics to allow for fine tuning. The DMD itself is connected to the EVM with the cable that came with the system. I don't need the LED's to work I just need to be able to put a pattern on the DMD. I brought up there being no voltage on the LED's to show that maybe the display board is not connecting properly. However now with the last flash I am getting 3.5V on pin 1 of J2,

VRST is reading -14V

VOFS is reading 9.96V and

VBIAs is reading 17.9V

when SW1 is in the on position and then the ground probe is touching TPGND4. Before I was getting 0 for all of these values. When I connect the original light engine up to J2 the LED's do not come on. I am still getting I2C errors when trying to connect through the GUI. Is there a way to step through the software and see what is going on? I have an MSP430 with a launchpad and an MSP-FET somewhere around here.

Hello Miguel,

The controller does check if LEDs are connected and will not allow you to pattern if it thinks the LEDs are damaged. I think there may be an issue here. Either the LEDs are not connected properly or they may be broken. Even if you don't intend to use the LEDs there must be something in their place.

Can you check if the LEDs are connected properly or broken?

The MSP430 is not in charge of this process. There is a boot process for the controller itself and we will have to figure out what test it isn't passing.

Did this EVM work out of the box before?

Regards,

John.

John,

It turns out that the connector for the red LED was broken on the LED board that came with the original set of optics. I rigged up a set of LED's on a breadboard in series with one 1k resistor per LED. The LED's wont turn on, the DMD still doesn't have a pattern, and I still get I2C errors when trying to connect through the GUI. However I do now get voltages at each pin on the J2 connector with ground referenced to TPGND4.

The voltages are:

Pin 1     3.5V

Pin 2     2.1V

Pin 3     2.1V

Pin 4     2.1V

 When measuring  voltages from pin 1 to pins 2 to 4 the reading is 0V

The circuit from the J2 connector to the LED's is here

 The EVM did work out of the box and was even working without the LED's connected before the update. Did I connect the LED's up properly or is there a configuration that would be better for EVM? The LED's were tested with a 9V battery to make sure they worked but maybe this configuration wont work for the EVM. Let me know what you think.

Hello Miguel,

Are you able to replace the broken part of the connector for the Red LED? 

If so, please try to replace the part and see if this will solve the issue. If not, we might need to provide you a test firmware for further inspection.

Regards,

Lori 

Lori,

I was not able to replace the connector but I did manage to connect the combination Blue and red LED by soldering a few headers to the interface board then scraping the PCB the LED is on with a razor blade until the traces were exposed and soldering wires directly to the traces. Here is a picture of what I did.

The interface board was the connected to the display board on the EVM, then the EVM was powered on. The DMD still does not produce a pattern and the LED's do not turn on but I do get voltages on all of the J2 pins when ground is referenced to TPGND4. The values of the voltages are

Pin 1     3.52V

Pin 2     2.1V

Pin 3     1.31V

Pin 4     1.3V

The voltages from pin 1 to pins 2 - 4 is still zero.

Hello Miguel,

Our team will look into this and get back to you by middle of next week.

regards,

Vivek

Vivek,

Thank you and your team for looking into this. If you need any more details from me don't hesitate to ask. 

Best,

Miguel

Appreciate your patience.

regards,

Vivek

Vivek,

Has your team found anything? I understand that your team is very busy but I need to start putting together new timetables and budgets for this project.. Could your team tell me where they believe the issue to be? Does your team believe it is a hardware issue or a software issue?  Does your team believe the issue is in the display board or the control board? Does your team believe this EVM can be salvaged?

Hello Miguel,

It is very difficult to remote diagnose such issue especially when hardware has either broken  or modified. We will continue to investigate and consult internally what could be the issue. Currently, I do not have any specific feedback or suggestion.

regards,

Vivek

Vivek,

Thank you for the response. I understand how difficult it troubleshooting is, especially remotely. I was just checking to see if your team came up with anything since you said a response would be provided a week ago. If your team comes up with anything I will keep an eye on this thread. Thank you and your team for looking into this.

Best,

Miguel

Hello Miguel,

Thank you for your patience on this.

It is a bit unclear on what is being said here. Is there a voltage detected on pins 1 - 4?

Additionally, can you read the resistance between the anode and cathode of the LED to as close to the DLPA as possible? It would be ideal to ensure that a connection is made here. The schematic for the board may be useful. It can be found on the TIDA-080001 page.

Regards,

Austin

Austin,

When the voltage is measured with respect to the display board ground there is a voltage on all of the pins. When the voltage is measured across each anode cathode pair there is no voltage. Looking at the schematic it does not look like like the cathodes of the LED's are connected to the board ground so the measurement relative to the board ground may be superfluous.

I don't think I can reliably measure the resistance from anode to cathode of the LED because the resistance of a diode depends on the forward voltage applied. Due to the way my multimeter takes readings it will give me a different reading depending on the setting(resistance scale) selected. Now the power provided from the multimeter in the diode check or continuity setting is enough to turn on the LED's individually when measuring from pin 42 on the DLPA2005 to pins 37 to 39 on the same IC (anode to cathode for each LED) with the reading on the multimeter being OL. The green LED does have a visible 90k resistor that I was able to measure. The combination red/blue LED likely has a similar value resistor for each LED inside of the packaging that I cannot access without destroying the LED's.

Yesterday I did more testing of the system, pulled out the MSP-FET and stepped through the program. It seems like it never gets past the i2c_master routine. The HDMI interrupt does seem to work but I didn't get much else out of it. Before attempting to measure the LED voltages I did re-flash the EVM with the limited version of the latest firmware for more consistency. 

I don't know if any of this helps but if you see anything or have any ideas, please let me know.

Best,

Miguel

Miguel,

Thank you for the additional information.

To confirm, the only modifications that were made to the EVM were updating the flash image and the alterations to the LED channels? Do the flash processes complete and verify that the flash was successful?

Regards,

Austin

Austin,

The only other major thing that was changed on the EVM was the display board. The display board replacement appears to be identical in layout and components. The EVM was working with the replacement board before flashing the image. When putting the old board on the EVM the results are similar but the RESETZ LED blinks. The old display board when everything else was working would also do this but the DMD would blink a checkerboard pattern for a while before scrambling. Now the old board will not blink the DMD at all. The DMD itself was also replaced with a new one to eliminate that as a possible issue. The flex cable was also replaced with the light board but it was also working before the flash update.

The flash process does complete and verify that the flash is successful every time I re-flash it through the GUI with SW1 in the off position. The first time I flashed the EVM I was using the GUI to project 8-bit alternating horizontal and vertical patterns running continuously from the internal pattern tab of the light control page at the same time.

If you want anymore details feel free to ask.

Best,

Miguel

Hello Miguel,

Thanks for the confirmation, Our experts will get back to you by mid of this week.

Appreciate your patience.

Regards,

Akhil

Hello Miguel,

The leading reason is that the FW is corrupted during the flash process.

We also have the firmware selector tool, here. Can you try the FW from that page to tell us what happens to the EVM.

Without having the EVM and with the current behavior it is hard to debug the matter. When this issue usually happens a reflash is the proper way to fix the matter.

Regards,

John

Miguel, 

In order to re-flash the firmware, please toggle the SW on/off switch to off position (The GUI now should indicate "Connected, Powered off") and the re-program the firmware available on the EVM page (https://www.ti.com/tool/DLP2010EVM-LC) via GUI (Firmware page)

Let us know if you are able to revive the EVM after following the above steps. 

Regards,

Mayank

Manyank,

This was the current procedure being followed since the thread opened. The procedure did not work for me but if I had done this initially the EVM would likely still be working. I know this is a discontinued part but if you update the GUI include those steps on the firmware update page. This is a very easy mistake to make,

Best,

Miguel

John,

I did try the firmware selector tool a couple of times. I tried uploading uploading firmware with the alternate I2C address as well but that didn't work either. I know this is very difficult to diagnose and repair I'm right here with you that why I appreciate all of the help that has been given in this thread. This is just a strange problem.

Best,

Miguel Hernandez

Tto give an update on the board. A few days ago I tried to trace the signals to see if they would dead end anywhere. I got to a level shifter, shorted a couple of pins and blew the 3.3V rail. Now there is a short from the 3.3V power plane to ground. Without individually removing each component and testing them I don't think its possible to diagnose where the short is. At this point it's time to move on. I did manage to find an alternative for the EVM but we'll see how well that works.

I would like to thank all of you that helped in this thread especially John McKnight and Austin Snyder. You guys provided a lot of help and suggestions for over 2 months the amount of help with the limited information you had available is jaw dropping. Thank you all for the time and effort you put into this.

Best,

Miguel Hernandez