cancel
Showing results for 
Search instead for 
Did you mean: 

ULINK-ME cannot detect target device on a new board

tomverschooten9
Associate II
Posted on August 12, 2008 at 11:34

ULINK-ME cannot detect target device on a new board

13 REPLIES 13
tomverschooten9
Associate II
Posted on May 17, 2011 at 12:41

Hi all,

I've designed a board based on the Keil MCBSTM32 eval board and I'm having trouble detecting the Cortex-M3 that is present on this board. I copied the JTAG interface from the eval board so I think it's kind of weird the ULINK-ME can't detect my target device. Both SWD and JTAG protocols won't work.

Has anyone encountered a problem like this before? If necessary I could put the schematics online.

Many thanks in advance,

Tom

16-32micros
Associate III
Posted on May 17, 2011 at 12:41

Hi Tom,

I advice you to check your SWD connections as it is simpler and requires less wires : you need only VDD, VSS, SWDIO (PA13), SWCLK(PA14), SWO (PB3) for the serial wire view and also RESET signal (NRST) and then verify that all signals are correctly routed and soldered to your connector JTAG/SWD connector. Hope this helps you to solve your board's connection.

Cheers,

STOne-32.

tomverschooten9
Associate II
Posted on May 17, 2011 at 12:41

I verified the connections between the SWD connector and the controller. I see some signals appear on the SWDIO and SWCLK pins, however I don't see anything on the RESET pin.

What should I learn from this information? Am I supposed to see anything on the other two signals?

Quote:

On 08-08-2008 at 15:47, Anonymous wrote:

Hi Tom,

I advice you to check your SWD connections as it is simpler and requires less wires : you need only VDD, VSS, SWDIO (PA13), SWCLK(PA14), SWO (PB3) for the serial wire view and also RESET signal (NRST) and then verify that all signals are correctly routed and soldered to your connector JTAG/SWD connector. Hope this helps you to solve your board's connection.

Cheers,

STOne-32.

[ This message was edited by: tom.verschooten on 08-08-2008 16:03 ]

jj
Associate II
Posted on May 17, 2011 at 12:41

Recall from similar posts that the Jtag/SWD clock line may have

problems caused by external pull-up/down resistors. (believe Olimex

board suffered from this - although ours works)

Assume you used short, direct routing micro to connector and that

no/minimal digital switching signals are adjacent.

Are you providing adequate power and bypass caps on your design?

Are regulators proper and verified?

Perhaps you ''shrunk'' your Jtag/SWD connector - and the ULink

''needs'' one of those unused connections???

Should above all fail - set-up for SWD and scope SWData & SWClock

on your working board. Then compare these traces to those obtained

(at the micro pins) on your new (failing) pcb. Good luck.

tomverschooten9
Associate II
Posted on May 17, 2011 at 12:41

I actually compared my schematic design to the reference design described in one of the STM application notes and this is a good match. I don't use external pull-ups except for the RESET line. Using an oscilloscope shows that the RESET line is stuck to the 3.3V supply. When I remove the pull-up it gets stuck at the ground. In any case it won't do anything as long as the JTAG connector is ''loaded'' by my failing board.

I've reassembled a second board and I'm still having the same problems. I did verify all lines work fine and do not suffer any kind of distortion. It seems the controller itself is refusing to use it's own pull-ups. Is there any way to check the controller itself is healthy?

I find it quite distressing that a copy of a working schematic (the KEIL dev board) isn't working. We're quite at a loss as to how this is possible.

Any suggestions are welcome.

[ This message was edited by: tom.verschooten on 11-08-2008 21:26 ]

16-32micros
Associate III
Posted on May 17, 2011 at 12:41

Hi Tom,

Could you please provide me your schematics so I can help you, as it is very strange and I haven't heard this kind of connections problems before with STM32 on assembling boards. Could you check also if you have a similar screen-shot as mine here attached where I have selected SWD instead of JTAG.

Cheers,

STOne-32.

tomverschooten9
Associate II
Posted on May 17, 2011 at 12:41

Working on it, ETA 15 minutes, together with some other information.

tomverschooten9
Associate II
Posted on May 17, 2011 at 12:41

Ok STOne,

here's the screenshot and schematic you asked for. I also set u my scope on the SWDAT and SWCLK lines while uVision tried to detect my target device. As you can see it fails and tries a second time (and fails again).

What concerns me most is measurement (4). It seems to me the controller is trying to send something back but fails horribly. What do you think?

I should also mention that I disconnected the pull-up for SWDAT to resemble the reference design in the application note. I also did the measurement again with the pull-up connected. This looks slightly better, but still has the same amplitude and thus is unreadable.

[ This message was edited by: tom.verschooten on 12-08-2008 11:43 ]

16-32micros
Associate III
Posted on May 17, 2011 at 12:41

Hi Tom,

I think that there is a big mistake on the schematics and you should remove the R20 that creates a short between VDD and GND and VDDA pin should be connected to VSUPPLY instead of Ground :o , all the analog parts of the MCU are connected to VDDA power supply.

Hope this helps you.

Cheers,

STOne-32.