Startup problems with a new NIRscan EVM. Blue LED D1 will not light up

Hello Phil,

Welcome to the DLP Advanced Light Control (ALC) section of the TI-E2E community.

Please see my answers:

1. Startup status LEDs-  Green power LED D8 indicator ok  Green LED D3 pulsing (heartbeat ok) and Sitara power line indicator blue LED  ok. Diode D7 flashes red momentarily on power up then goes out. This is expected behavior. It is not unusual for D7 to briefly flash on power up.  The second green LED is D3. D10 (near the USB between the Ethernet port and two USB A ports) should light showing power is stable to the Sitara, and then the Sitara should load (see answer to 2).
2. After 20 seconds or so , Blue LED D2  light up every time(nothing in the user guide about D2 under normal operation).  This LED indicates that the Sitara has finished loaded and the OS is running.
3. Blue LED D1 fails to light up at all - even after 10 minutes or more ( The quick guide states we should wait until D1 lights up before plugging in the mini USB – however, under normal operation we have not seen D1 light up whether the USB is connected or not).  This must be a typo - I believe it should read D2.  We have logged this to update in the next revision.  Thank you.  During normal operation D1 should not light.
4. Plugging the USB cable into the mini USB Location 14, RNDIS/Ethernet gadget driver starts up ok- Pinging 192.168.0.10 gives ok response. This is OK.
5. The DLP NIRscan  browser page can be opened on 192.168.0.10 ( with just Simple and Custom Scan  options ( V1.0)) This is also as expected.
6. Using Simple Scan, the  device fails to  collect reference spectra- it just sits there stating 'Reference Scan in progress' and   waits  and some time later (~10 minutes) displays  'Reference Scan Failed'.  This suggests a corruption of files on the flash image.
7. In Custom Scan mode, if any of the pattern variables are changed  requiring a new set of Patterns - selecting 'Generate Patterns' results in the response 'Generating patterns'  and  then waits   and some time later, displays  'Reference Scan Failed' (No reference scan was actioned).  This also suggests a corruption of files on the flash image.

I suggest rebuilding the v2.0 flash image on a different SD card and try reloading yet again and let us know the result.  It is only during reloading the flash from an SD card that D1 should light.  During reprogramming D1 & D2 should flash "erratically" for a while and eventually (5 - 7 minutes) come on solid blue.  Make sure to wait for about 30 seconds after D1 & D2 have come on solid before powering down to remove the SD card.

Also just to be sure re-flash the DLPC350 v2.0 firmware.

Please let us know the result and we will go from there.

Fizix

Hi Fizix

Thanks for your quick response

We've been through the suggested programming cycle from scratch again, downloading and rebuilding the new NIRscan image and loading from a new unused SD card, (I've tried 3 now) Programming goes smoothly ending with both D1 and D2 on solid Reloading the DLPC350 firmware using LightCrafter4500 interface, goes smoothly.

Web browser opens up ok and the Info page shows V2

Unfortunately - it still refuses to complete a scan in either Simple or Custom scan mode. The symptoms are as before with earlier attempts , with V1 ( out of the box) and 3 attempts with V2 firmware.

I think this EVM , despite being brand new, has probably got a hardware fault.

Hi Fizix

Is there anything else you can send me that we can try to prove to ourselves that this unit can function?
The NIRScan application browser page is very limited in its diagnostic capability

For example- What can we do with LightCrafter4500 interface directly into the dlpc350?

I'm minded to send it back to our distributor for replacement.

What do you think?

Hello Phil,

Thank you for attempting the items I listed.  Before we pronounce it DOA, please try one other thing.  We have and several other customers have experienced marginal USB cables that can cause issues.

If at all possible, try another (or two if you have them available) A to mini-B USB cable, preferably of a different brand, with the kit to see if anything changes. 

Fizix

Hi Fizix
I had anticipated this possibility having seen that mentioned on the Forum already and have had a go with reloading flash images with alternate mini USB cables-. Again programing seemed to go ok but the unit will not scan.

Hi Fizix

We have persevered with this TI technology and  sourced a second unit  and  this second one also  does not  fully function out of the box either-. using V1.0  application firmware as delivered.

The symptoms are  slightly different this time in that in Quick Scan mode,  initially  we  had 2 or 3  successful completion of reference scans from  it, interspersed with numerous 'scan failed'   but  we have never got a completed sample scan - that just returns a scan failed..

If this is all due to flash corruption,  why would it appear to  partly work? 

What are the causes of these failures?  we started off with  our  'best' USB Cables and an isolated Laptop  

Is there a history of failures/fragility' with these EVMs?

Do you have  a recommended  USB cable spec/brand  that is a known minimum requirement? (in the same way you stipulated the Digikey 12V5A supply?)

Is there a way to work around the  'Ethernet over USB' cable  connection issues using the RJ45 Ethernet port?   to get this going quickly   

 

Dear Fizix

I have now worked through this and found that the root cause of most if not all these problems is that the NIRScan EVM is incompatible with Windows7/IE Explorer 11 (despite what it says on p5 of the user guide )


Google Chrome works ok (so I'll stick with that)


Phil