32.768kHz clock doesn't start sometimes!

fabio239955_stm1_stmicro_com · ‎2011-02-11

Posted on February 11, 2011 at 18:08

#lse-rtc #lse-low-temperature

santi · ‎2011-10-25

Posted on October 25, 2011 at 17:19

Hi clive1,

Thank you for your suggestion. I have checked that the 32.768kHz crytal is 12.5pF, this seem to be a problem, right? But some how the crytal on the development board that I have is 12.5pF too. I desolder the crytal from the devlopment board and the put it into to my board but it does not make any good. Anyway, I am purchasing the proper crytal( 6pF), hopfully this should solved the problem. Thank you again.

emalund · ‎2011-10-25

Posted on October 25, 2011 at 17:49

These circuits are hugely sensitive to capacitive loading, and resistance, as well as the physical characteristics of the PCB, traces, crystal, and soldering.

......

layout, layout, layout

and the response is:

I am purchasing the proper crytal( 6pF), hopfully this should solved the problem.

BTW I have never heard of a 6pf crystal

Erik

santi · ‎2011-10-25

Posted on October 25, 2011 at 18:20

Dear Erik,

I am reffering to 32.768KHz CL=6pF.

emalund · ‎2011-10-25

Posted on October 25, 2011 at 18:48

OK, let's try this way

are the traces to the xtal and the burden caps less than 1cm ?

are the traces to the Vdd decoupling caps less than 1cm ?

is every Vdd pin decoupled?

does at least one of the Vdd pins have a tantalum in addition to the ceramic

santi · ‎2011-10-25

Posted on October 25, 2011 at 19:10

Dear Erik,

Yes, I have.

1 yes, the traces to the xtal and the burden caps are less than 1cm. ~8mm.

2 yes, the traces to the Vdd decoupling caps less than 1cm . ~5mm.

3 yes, the every Vdd pin is decoupled with 0.1uF ceramic.

4 yes, there is one 22uF tantalum at Vdd.

My PCB already fit these but still no luck.

thank you

nguercetti9 · ‎2012-02-20

Posted on February 20, 2012 at 09:30

I have a similar problem: at low temperature My LSE start whitn some seconds of delay, aboaut 12s at -20°C.

This delay is very shot at +20°C, 3seconds.

My cristal is ESR=40K, CL = 6p Co= 1.2p and CL1=CL2=6.8p, layout very similar to KAIL evaluation board.

Any Ideas?

fiim · ‎2013-04-19

Posted on April 19, 2013 at 21:12

For the stm32L15x problem solved: From the pinOut32kHz to the crystal must be resistor 1kOm.

zzdz2 · ‎2013-04-20

Posted on April 20, 2013 at 09:30

It may be interesting to measure LSE transconductance of the chips that fail to start RTC.

Maybe it's out of specs, who knows?