DLPC3470: DLPC3470/DLPA2005/DLP2010LC Won't initialize after sitting without power

Hello User,

In your design is there a switch that you have to flip to get the device to project or do you just apply power to the system and it boots on its own?

What FW version do you have and is it from the Pico FW selector tool?

We will also need to focus on the DLPA2000 and system VIN for system behavior as well.

I do want to note that with the holidays there may be a delay in response time till January.

Regards,

John

Hi John,

We have an additional CPU that control the PROJ_ON signal.  Normally on power-up, the signal is put into the state that allows the projector to boot and enter splash screen mode on its own.

We downloaded the firmware from your website March 25'th using DMD routing option 4.  Yes, it's from the Pico select tool.  The download's name is: FwSel_DLPC3470+DLPA2005_pm4_isc0x36_v8p01p.img.

Like I said in the original post, we have built around 50 of these, and most work regardless of how long they have been powered off.  We have a few that work correctly if powered up within 10 minutes of sitting powered-off but don't power-up if they sit longer than this magic threshold.  On these units, if we cycle power a second time not allowing them to sit powered-off longer than the 10 minute threshold, they work correctly.  This has all the ear-marks of a hardware issue, not firmware/software, so I'd respectfully ask that you concentrate in that area.

Thanks,

Scott

Hi Scott,

It's hard to directly point to whether there is a problem on your design board that is related to your issue. 

Do you follow our reference design and layout recommendation? 

Are you able to use oscilloscope to capture power up sequence and check if there is any fault flag when these failed boards couldn't power up properly?

You can refer to the Power sequence diagrams in the DLPA2005 Datasheet section 7.3.1, and share your findings. 

Regards,

Lori 

Hi Lori,

I did follow the design recommendations.

When the device doesn't power up, most of the generated supplies don't power-up.  -14, +18 and +10 are all zero.  Our 1.8 regulator is dragged down to 0.9V.  On the same board when it starts-up properly, the 1.8V supply is 1.8V as is should be.

On failed start-up, the 3470 doesn't respond to I2C communication.

The only input to the 2005 is System power, 5V in our case, there's nothing to look at.

Thanks,

Scott

Hi Scott,

It seems like the issue could be cased by the 1.8 V regulator dropped down to 0.9 V, then VOFS, VBIAS, and VRST go into fast shut down. I have some questions and suggestions for you to check:

1. When does the 1.8 V drop occur? It might be some components drive the voltage, and you can check the load on LS_OUT.

2. Test on the 1.8 V regulator when Proj_on is low.  Do you still notice the voltage drop?

3. For sanity test, use the same optical engine (i.e. same DMD and flex cable) on both working and non-working boards. Check if there is any differences.

4. Check and compare the assembly of working and non-working boards. You may need to check with your manufacturer and do an assembly test. 

I hope these may help you find the cause.

Regards

Lori 

Hi Lori,

The 1.8V regulator being loaded down is a manifestation of the 3470/2005 not starting correctly, not the other way around.  When the system starts up properly, which happens when power is turned off then back on within 10 minutes, the regulator outputs the correct voltage.

3) It is the same unit, DMD and flex cable that both won't start correctly and will start correctly depending on how long power is off.  If over 10 minutes, the unit doesn't start, if under 10 (or so), it starts and operates normally.  This is consistent between a variety of computers.

4) Please read my comments: it is exactly the same boards, cables, DMD, everything that both works and doesn't work.

Thanks,

Scott

Hi Lori,

The 1.8V regulator being loaded down is a symptom of the 3470/2005 not starting correctly, not the other way around.  When the system starts up properly, which happens when power is turned off then back on within 10 minutes, the regulator outputs the correct voltage.

3) It is the same unit, DMD and flex cable that both won't start correctly and will start correctly depending on how long power is off.  If over 10 minutes, the unit doesn't start, if under 10 (or so), it starts and operates normally.  This is consistent between a variety of computers.

4) Please read my comments: it is exactly the same boards, cables, DMD, everything that both work and doesn't work.

Thanks,

Scott

Hi Scott,

For clarify my question 1, do you see any other voltages/signals behaving differently when the 1.8 V dropped down? Did the drop happen before or after Proj_on pulled high? 

Meanwhile, can you monitor both LS_IN and LS_OUT?

For "3) It is the same unit, DMD and flex cable that both won't start correctly and will start correctly depending on how long power is off.  If over 10 minutes, the unit doesn't start, if under 10 (or so), it starts and operates normally.  This is consistent between a variety of computers.

4) Please read my comments: it is exactly the same boards, cables, DMD, everything that both work and doesn't work."

You mentioned that you have 30% of PCB assemblies suffer from this issue, and the rest of the boards work fine and regardless of how long they have been powered off. I would like to check if you did the test and comparison (question 3 and 4) between 30% "non-working" boards and the remaining 70% working boards that regardless of how long they have been powered off? 

Regards,

Lori