ESD damaged FPGA?

[Matthew]

I have managed to get myself in a real twist today. I set out trying to re-attempt writing a software based solution for reading the MAX11645 on the HL2-MRF101 PCB (see thread here). For prototyping, I was using a modified version (modified to try and read/write MAX11645) of the hermeslite python program. I wasn't seeing a difference in behaviour for a read at addr 0x36 (MAX11645 address), and an i2c address not on the PCB. I think the port 1025 discovery response doesn't report back the error (per extract from wiki below):

If a write or read is attempted when the I2C or AD9866 subsystems are busy, an error response will returned. An RADDR of 0x3F is indicates error. RDATA will match the original request.

So I then decided I needed to get my scope out and see if the ADC was responding at all to i2c requests. I soldered 2 flying leads onto the HL2-MRF101 PCB (with the board powered off(!)). I connected the 2 scope probes to these flying leads and powered up the HL2. I've used this same iron for the handful of hardware mods I've done to my HL2 boards over the past few years.

I was then unable to discover the HL2 with a direct wired connection. When I plugged it into my LAN it picked up an address by DHCP and responded to a discovery (in SparkSDR). The last 2 bytes of the MAC had changed to [FF],[FF]. I used SparkSDR to try and change these bytes of the MAC. I cycled power but still had FF, FF. I unplugged the HL2 from the HL2-MRF101 PCB and connected the N2ADR filter board. I tried changing frequency in SDR software to see if the MCP23008 on the same i2c2 bus was working. No familiar sound of relay clicking. At DB7 on scl2 I measure (with scope) 2.92 V and on sda2 I measure 3.3 V. On my other HL2, I confirm both voltages are 3.3 V and see the logic toggling when changing band (i2c2 writes to n2adr filter board).

I have programmed (via SparkSDR) 72p3 (was on 72p4) and cycled power, but no difference in behaviour.

I am out of ideas and my only explanation is that I must have somehow caused a ESD damage to pin 31/30 on the FPGA via the HL2-MRF101 PCB? I have become slack in recent years and I don't wear an ESD strap when working at my bench (and I don't have large pile of failed circuit boards either!).

My only other idea is to remove U15 in case that has failed in a manner that is holding the i2c2 lines? This seems unlikely, but is easier to replace than the FPGA.

Any other ideas? I'm keen to exhaust all options before changing the FPGA (this is a build9 board).

73 Matthew M5EVT.

[Matthew]

I removed U15 and I still cannot see any communication between n2adr board and HL2 on i2c2 lines. I've re-flowed the relevant pins on the FPGA and done a full inspection of the i2c2 lines across the PCB. I think I must have damaged the FPGA IO. I've ordered a new FPGA.

Given the number of boards out there and no one ever reporting such problems this definitely seems like an extremely rare occurrence and not to be discounted, I'm using it with a board still in development (the HL2-MRF101). Nonetheless, I'm surprised.

73 Matthew M5EVT.

[Steve Haynal]

Hi Matthew,

Did you also inspect R48 which pulls up SCL2? It may be that if R48 is bad or absent, the FPGA can't fully bring SCL2 to high. Do you see activity on SCL2 when sending any I2C command?

Makerfabs once sent me a failing board which turned out to be a bad FPGA pin for SCL1. I recommended replacing the FPGA on ~3 more boards they had with similar failure.

73,

Steve

kf7o

[Matthew]

Hi Steve,

Thanks for your thoughts. I have inspected R48 (and re-flowed) and re-flowed the pins on the FPGA and on the pull up resistors. I see no activity on SCL2 when sending an i2c command.

I re-read my original message and perhaps it wasn't 100% clear, this was working as expected (i2c2 functioning) for the past few months.

I will report back when I have replaced the FPGA.

73 Matthew M5EVT.

[James Ahlstrom]

Hello Matthew,

Sorry to hear you need to replace the FPGA. It is not easy. Hopefully you will do better than I did, as I wrecked my board when I tried. Good luck!

Jim

N2ADR

[Ed Murphy]

Hello Matthew. I hope you can read thid. Have you started removing your FPGA yet?. I have some ideas on a safe method. If you are interested please contact with PM. BTW, I cannot see any of this text that I am typing ! Such is the google groups page formatting. Cheers Ed GM3SBC.

[Matthew]

Jim,

FPGA now replaced. I wanted to be silent until I had replaced it, your issue was my primary concern! Thankfully, I now seem to be in a better position and i2c2 lines are now working and changing filters on your filter board. Still no comms with the bias EEPROM, so I will have to order a replacement before I can fully consider myself in the clear.

Ed,

Thanks for the offer of advice, I have a hot air rework station which makes swapping an IC a bit less "heart in mouth".

73 Matthew M5EVT.