Using OCXO as an External Oscillator for STM32H533

[modecs]

Dear,

I am planning to use an OCXO (Taitien NI-10M-2503) as the external oscillator (HSE) for synchronizing the MCU to a highly accurate clock source. The OCXO provides an RF output with a sine wave swinging between -2V and 2V.

According to the STM32H533 datasheet, it appears that the input signal on the OSC_IN pin must not exceed Vdd or fall below Vss. In my case, Vdd is 3.3V, and Vss is ground.

I would like to seek your guidance regarding the following:

1. Based on the datasheet, my understanding is that directly connecting the OCXO output to the OSC_IN pin would not work. Could you confirm if this is correct?
2. Would it be necessary to use a circuit such as a bias-T to condition the signal before feeding it into the OSC_IN pin?
3. Are there any example circuits or application notes available for this type of setup?
4. If there are alternative or better approaches, I would greatly appreciate your recommendations.

Thank you very much for your time and assistance. I look forward to your advice.

Best regards,

[Billy OWEN]

Hi @modecs 

 

This post has been escalated to the ST Online Support Team for additional assistance.  We'll contact you directly.

 

Regards,

Billy

[Uwe Bonnes]

10dbm is 2 V pp, not 4 Vpp. Add a capatictor between OXCO out and STM32 Osc_In and 100 Ohm each from OSC_IN to GND and VCC and that should be do the job.

[modecs]

In my simulation setup, I have set the OSC IN resistance at 200 kΩ due to the R_feedback value being 200 kΩ. For the voltage divider resistor, I initially used 100 Ω but found it did not achieve a 3.3 Vpp ; hence, I adjusted it to 560 Ω. Regarding the capacitor for DC blocking, setting it to 1 nF seems appropriate to achieve a 3.3 Vpp signal level.

 

 

[DevZen]

1. Based on the datasheet, my understanding is that directly connecting the OCXO output to the OSC_IN pin would not work. Could you confirm if this is correct?
-> If the RF output of the OCXO ranges between -2V and 2V, it cannot be directly connected to the OSC_IN pin of the STM32H5. When BYPASS mode is used, OSC_IN is a digital input and it requires >0.7*VDD for high level and <0.3*VDD for low level signal. (Refer to Table 39, "High-speed external user clock characteristics (1)", in the datasheet.)

2. Would it be necessary to use a circuit such as a bias-T to condition the signal before feeding it into the OSC_IN pin?
-> Typically, a Bias-T is not used for the OSC_IN pin of STM32 MCU. The OSC_IN pin is intended to operate with standard crystal oscillators or external Clock sources that comply with the specified voltage levels and signal requirements. Before deciding whether a Bias-T is needed, check the output format of the Crystal Oscillator and the requirements of the circuit it will be connected to.

3. Are there any example circuits or application notes available for this type of setup?
4. If there are alternative or better approaches, I would greatly appreciate your recommendations.
-> You can refer to ST's AN2867 document for more information.
[AN2867 Guidelines for oscillator design on STM8AF/AL/S and STM32 MCUs/MPUs]
https://www.st.com/resource/en/application_note/an2867-guidelines-for-oscillator-design-on-stm8afals-and-stm32-mcusmpus-stmicroelectronics.pdf

[Uwe Bonnes]

Your OCXO is power by 5 Volt. So absolute highest possible swing is 5 VPP. They guarantee 10 dBm at 50 Ohm, that is 2 VPP. The OXCO output seems to have a capacitor at the output, so you see +/- signal. RM0481 in 11.4.1External clock source (HSE bypass)  talks about what to set for "low swing (analog) " signals.  The needed HSEON bit needed to be set for analog input probably biases OSC32_IN in the right way for AC coupled inputs.
With your 5 Volt OXCO directly connecting the OXCO output  without 50 Ohm termination may cause voltages to high at the input. I assume there is also another consumer for the OXCO, that probably terminates the line. So AC coupled input should be fine. If not, perhaps consider a 3.3V OXCO or terminate OXCO with 50 Ohm and AC couple to OSC_In.