HSE do not oscillate anymore

[riccardo caira]

Posted on March 22, 2018 at 10:57

Hi guy,

I was developing my lcd application under discoveryf4 (see photo) and everthing works fine.

I decided to develop my custom board, and use discovery to flash and debug my stm32f407vgt6 mounted on my custom board.

In order to make this, jumper on CN3, jumper on JP1, bridge on SB11 of discovey has been removed.

After connected my custom cable (gnd, swdio, swclk, nrst signals) SURPRIIIISEEEE!!!!! everthing works at the first attempt (successfully loaded blink_led_firmware).

So I decide  remove flying cables of lcd and, SURPRISE, nothing works anymore.

I can't connect via swd to my custom board. Neither using st-link utility.

Vdd is stable 3.3V.

Voltage on VCAP is 1.26Volts.

Xtal is 

https://www.tme.eu/it/details/8.00m-cfpx104/risonatori-al-quarzo-smd/iqd-frequency-products/lfxtal055663bulk/

 . 

Replaced with xtal removed from other discovery board. Replaced with other 10mhz xtal.

Tried 0 ohm, 220 ohm, 1kOhm on rcc_osc_out pin,

Capacitors on Xtal are 2x20pf COG 0805. Replaced with capacitor  removed from other discovery board.

Replaced mcu.

Soldered new pcb with new components.

Replaced mcu.

SWDIO and SWCLK signals on discovery are good.

The only bad design (I see) is xtal analog gnd sorrounded in digital gnd.

My xscope  alway find 0v on rcc_osc pins.

I've already read many posts about not working hse but they were not helpfully for me.

I'm having this problem by 4 days, please give me some ideas.

Thanks.

other photos

[waclawek.jan]

Posted on March 22, 2018 at 12:57

Check SWD/reset connections up to the mcu pin (i.e. check for bad solder joints); the same for all ground and supply pins (including VDDA/VSSA).

Try shorter SWD cable. Try to make a cable with signals interleaved with ground wires.

Do you use 'Connect under Reset'?

JW

[riccardo caira]

Posted on March 22, 2018 at 15:11

Made swd cable shorter, connection are ok 

(still not working).

Cleaned flux residue with acetone(still not working).

I've re-soldered all 100nf cap on vdd, 2.2uf cap un V_CAP pin,  changed 100nf cap c16 and c17, changed R5 by 51 to 10ohm, re-soldered r6 10kohm reset pin pullup resistor, re-soldered r7 10kohm pull-down resistor on BOOT_0 pin, re-soldered c44 100nf V_BAT pin.

Now it WORKS!!!!!

Now I can connect under system workbench and st-link utility (mode: normal | connect under reset) at 4mhz.

I don't understand why after the first attempt my pcb wasn't working anymore.

I'll re-solder my 2nd pcb and I'll let you known what will be results

Thanks JW....

[Tesla DeLorean]

Posted on March 22, 2018 at 15:39

If BOOT0 is not low then your code will likely not execute out of reset. Debugger redirects execution to code under test after resetting, so could explain different functionality when debugger disconnected.

[Andrew Neil]

Posted on March 22, 2018 at 22:10

my custom cable (

gnd

, swdio, swclk,

nrst

signals)

NRST has already been mentioned.

With your custom cable removed, is your board getting a good GND?

[T J]

Posted on March 23, 2018 at 02:13

its looks like a soldering issue,

remove the crystal board,

clean all contacts, reseat crystal board using flux.

in your final design,

put the crystal directly on the pcb - no daughter board arrangements,

the connections fail often because they are brittle.