Analog Discovery not recognized (suddenly after several hundred hours of use)

[Mike-san]

Hi Folks,

I am a prolific user of the AD2 platform.  I currently own 14 units.

The devices are used to capture the IV curve of a transistor-like device over its lifetime.

Basic workload of the AD2 is to sit and wait for a trigger while the rest of the system puts hours of use on the transistor.

I've found the AD2 to be very unreliable over time... mainly due to the device ceasing to be recognized by the system after 50-100hrs of waiting for a trigger.   After the failure, power cycling, new cable, different host do not remedy the problem.

Something wears out over time.  And renders the AD2 unusable.

Do you guys have any ideas?

Thanks in Advance

Mike-san

 

Statistics:

Qty 1 - PASS: Up/running for debug only.  No long hours of use.

Qty 4 - PASS: Up/Running,  hours of use in the 200-500hr range (simply waiting for a trigger and logging a sweep periodically)

Qty 8 - FAIL: failure mode as described above.

Qty 1 -  FAIL: System reports PLL failure

Other details:

- PCBA removed from AD2 packaging and integrated into a larger system  (yes... risk of ESD)

- Temperature inside enclosure can be 30-35C (not sure if the AD2 is rated for this.. most of the components look to be 0-75C rated)

- External 5V was added in a few cases, but did not solve the problem.

- I've tried R/R of the FTDI device (not sure what would wear out in that chip).

- I've tried R/R of the EEPROM (could be data retention/temp problem there).

- Wish I had some debug tools for the AD2  :)

 

 

 

 

[attila]

Hi @Mike-san

I'm using some devices since many years and have no problem with them.
The only issue I have is with some non-original USB cables wearing out. I prefer to use shorter cables.

As problem with your 8+1 failing boards I could think on the ferrite beads.
These are under the USB connector supposed to filter noise but it could act as fuse.
In case you are using external supply for the circuit under test and don't have proper ground connection with the device, ground loop over USB cable could stress the GND ferrite bead.

 

[Mike-san]

Hmmm... very interesting.

I've discovered something else... as a hard core Digilent fan and customer... I also use a CMOD6 Spartan 6 FPGA.

If I reprogram my FPGA with the AD2 installed... AD2 dies a terrible death.

Wondering if I'm reprogramming the CMOD6 you guys used to develop the AD2?????

I will also check my ferrite beads..

Than ks

Mike

[Mike-san]

btw - do you have a part number for the ferrite bead?   Sounds like another Mouser order ....

Also ... do you have a .bit file I can use to reprogram the AD2?

 

Cheeers

Mike

[attila]

Hi @Mike-san

Could you provide more info about the issue? What tool are you using to program the CMOD ?
I have no problem programming a board (Arty Z7 similar USB as CMOD S7) with Vivado 2019.2 and using AD2 at the same time with WF app.

The AD2 FPGA is programmed automatically when connecting to the device with WF or custom app/script.

BK2125HS470-T but a 0R or simple short solder should be fine.