ESP32 Forum

Hello,

the built-in 32kHz oscillator on WROVER-KIT-VE board is not working
I am getting and later the 32kHz crystal is working without any problem on the old variant of board (WROVER-KIT-VB)

I've ordered one more WROVER-KIT-VE to eliminate possibility of a broken part and the other one has also the same problem.

I've played also with some settings in menuconfig as found in other posts, but without success

Best regards

Hi,

Can you check the following resistors?
Are you setting "RTC clock source" with "External 32kHz crystal" from the menuconfig?

Hello,

yes, the 0ohm resistors between crystal and module pins are present
yes. I've set external crystal in the menuconfig

Hi,

No clue on this issue. Maybe you can try re-soldering the crystal.

m.madela wrote: ↑

Tue Apr 27, 2021 11:41 am

Hello,

yes, the 0ohm resistors between crystal and module pins are present
yes. I've set external crystal in the menuconfig

You should also make a continuity check from the pins on the wrover module to the pads of the crystal. This will check if everything is soldered correctly and the 0ohm chip resistors are not cracked.

I am sure the paths are correct, there is connection between both end of crystal and module pins

Was there any solution found for this?

I also have a ESP-WROVER-KIT-V4.1 and I'm seeing the same behavior. No matter what clock settings I select, I see: I've tested continuity, and I can confirm that the 32.768kHz crystal does have continuity to the WROVER module's pins.

"me too"

May be relevant: https://github.com/espressif/esp-idf/issues/6621
Altough I've tried commits all the way back to August of last year, and I can't get it to work. We've got a batch of 5 custom boards, on 3 the crystal works, on 2 the crystal doesn't. It also doesn't work on the ESP32-WROVER-KIT-v4.1, but works on all of the DFRobot ESP32 FireBeetle boards that I've got laying around (3 of them).
No amount of tweaking the calibration cycles (up to 10k) or bootstrap pulses (up to 100) can get the crystal to start (atleast to an osciloscope I've got here). The "Additional current" option also doesn't seem to be doing anything, although I can't even find any documentation for that.
Seems like a hardware issue at first, but I'm confused as a few boards from the same batch work and others, including the WROVER kit don't - with the same firmware.

Well.. those are my 2 cents

Could you try to add a parallel 10 M resistor across 32.768k crystal on ESP-WROVER-KIT-VE ?

I've tried adding a parallel 10MOhm resistor across the 32.768k crystal on the ESP-WROVER-KIT. I've attached a photo of the resistor, and confirmed that it is properly soldered in place. With it in place, the measured resistance across the crystal is now ~6MOhm. This did not fix the problem: I still see "W (686) BTDM_INIT: 32.768kHz XTAL not detected, fall back to main XTAL as Bluetooth sleep clock
light sleep mode will not be able to apply when bluetooth is enabled".

I also have the same problem on a batch of custom boards that use the ESP32-WROOM-32D. Sometimes I don't see the "BTDM_INIT: 32.768kHz XTAL not detected" message, but the 32.768k clock is unstable and causes the board to usually crash when entering light sleep. Other times, I see the "BTDM_INIT: 32.768kHz XTAL not detected" message and the board boots somewhat normally. This happens regardless of whether I add a parallel 10MOhm resistor.