Crystal drive level

[shadewind]

Posted on September 22, 2011 at 11:31

I have read the ST app note about crystal oscillators and I am concerned about the drive level. Before reading this app note, I chose the TXC 7M series crystals for its small size and reasonable cost. In the datasheet, its drive level is specified between 1 �W and 50 �W. Unfortunately, I do not have the equipment to measure the drive level in place on the board. Is there some advice you could give me here? Is the 7M series likely to work properly without a series resistor or do I need to change my design?

#stm32-crystal-drive-level-txc-7m #stm32f2-stm32f4-crystal-gm

[Rhodes.Ethan]

Posted on February 13, 2014 at 18:30

I am also curious about the oscillator transconductance spec for the STM32F2 and F4 series. I think I read somewhere that this spec was lowered from the STM32F1 series to the F2, probably for power consumption reasons? By looking at the data sheets, it appears the F2 has an oscillator transconductance that is a factor of 5 times less than the STM32F1 series. By using app note AN2867 as a guideline, it is fairly difficult to find crystals with a small form factor and higher frequency that can meet the 5/1 Gm to Gmcrit ratio suggested for the F2 and F4 specs. Am I reading the F2 and F4 data sheets correctly that I will need to find a crystal with less than 1 mA/V Gmcrit as calculated in AN2867? Thanks for any clarity that you can add.

[derek]

Posted on April 18, 2014 at 08:53

I  just ran the numbers for my design and I found the same result, the 25Mhz oscillator I'm using has a really low gain margin (1.62).   To this point I've never had issues with the design not starting up.   Its seems some of the specs (ESR, C0, and gm) are worst case.   Its likely that in most cases the gain margin is much higher than expected.

However, it really worries me that the gain margin might be so low in the right set of circumstances. 

I'm wondering what solution to the issue of low gain margin with your design since you first posted?