Ulink2

[adaniel]

Posted on November 08, 2014 at 21:27     IDE: Keil 4.74 (microvision)     Board: proprietary board with STM32F102C8     Program: only activates UART 1 to communicate with a terminal     1 - I can download the program to the board using the Flash Loader Demonstrator,          the program works fine     2 - IDE debug mode: the Ulink2 is detected but: ''Ulink2 communication failure''     3 - This same Ulink2 works perfectly on another board with STM32F103VB and           exactly the same program.     4 - The JTAG pins are routed the same way on both boards. I checked them with           an auxiliary program disabling the JTAG and outputting a 1Mhz square wave           at these pins, one by one.  Any clue? Could it be that the Jlink2 doesn't support the STM32F102 anymore?

[Tesla DeLorean]

Posted on November 08, 2014 at 23:11

I don't have an F102, but I'd be surprised if that is the issue.

  Check that you connect with SWD mode, and the ''Connect & Reset'' options. Ideally you'd want more than one reference board. Keil claims it's workable

http://www.keil.com/dd/chip/4802.htm

Check grounds.

[adaniel]

Posted on November 09, 2014 at 00:10

 Thanks Clive for your reply (but it doesn't help). I have of course tried all possible

 SW settings, starting with the ones which work with STM32F103.  I know that the Ulink2 is supposed to support STM32F102, but does it? From my  experience on one board, the score is 1-0 against Ulink2.  Ground problem? it is usually what we say when we don't know what to say. But it  is still a possibility. I hope to get more boards these days to test them. I am going  to try ground bypasses and bigger and better decoupling capacitors.  Note that this is a 4-layer board with the two external layers being exclusively ground.

[Tesla DeLorean]

Posted on November 09, 2014 at 00:35

Actually my concern was the ground over the debug interface, and that in fact they shared a common ground. Your previous post suggests the signal pins are viable. If the ULINK is using the target reference, check that also.

I have very little insight into your board, or debug connection. Here I'd try a JLINK and ST-LINK if I suspected the ULINK.

[adaniel]

Posted on November 10, 2014 at 15:21

 No progress so far. I have checked two more boards and the score is now 3-0 against Ulink2 (and ST-link). Ground bypasses and bigger capacitors have no effect.

However there is still a point to investigate: the STM32F103 board which works, has an 8Mhz crystal, while the STM32F102 board has no crystal: maybe that Ulink2 and ST-link need and external crystal to start? I am going to check this tomorrow. I have also ordered pin-to-pin compatible STM32F103 chips which I hope to receive soon. Anyway thanks for your suggestions.

[Tesla DeLorean]

Posted on November 10, 2014 at 15:41

The parts have no dependency on external crystals, the chips don't boot from them, and it's impossible to guess what someone might attach for generic software to be dependent on them.

Make sure NRESET/NRST is pulled high, but not driven high/low by external circuitry. VSSA powers the PLL and POR. Doesn't sound like a problem capacitors can fix. Seeing as you have three boards, and JTAG pods failing, you might want to focus on connectivity in the design, boards/soldering.

[adaniel]

Posted on November 11, 2014 at 13:28

As expected, crystal makes no difference. I am waiting for the STM32F103 chips I have ordered.