Surface mount high speed CMOS oscillators

[Adraen]

Hi there,

I am currently working on a project with an STM32 and a TI Ethernet Phy. The main point is for the Phy to work a 50MHz CMOS oscillator needs to be used.

I don't have many problems soldering surface mount components in QFP packages, but every single surface mount oscillator i've found don't expose their pins such as this one : https://uk.farnell.com/abracon/ase-...rch=Featured New Products&MER=i-9b10-00002068

The soldering of such oscillator is quite hard without an oven/heat gun and I've been looking for oscillator with leeds but couldn't find any.

So the question is does any one know if and were we can buy oscillator with leeds or an easy method to solder such oscillator ?

P.S: I'm aware of the through holes oscillator but if all the component have to be on the same side then a via is required and most ICs say not to put vias on the clock track. (also most of the time they are much larger than surface mount ones)

Cheers !

 

[keith1200rs]

Farnell's site is down at the moment so I cannot look at the device you are referring to, but the simple answer is likely to be to make the pads of the footprint slightly larger than the device so you can hand solder it. The solder will flow under the device and the exposed pad gives you something to heat up with a conventional soldering iron.

Keith

 

[FvM]

but the simple answer is likely to be to make the pads of the footprint slightly larger than the device so you can hand solder it.

The footprint suggested by Abracon will already work for hand solder, but if you don't like it, there's no problem to use slightly larger pads.

Seriously speaking, I never experienced problems in hand soldering of oscillators or crystals in ceramic package. I wonder if you tried at all or are only raising a theoretical problem?

 

[Adraen]

Hi cheers guy I've been using slightly larger pads for a while now, I was mainly asking this because last time I soldered a 50Mhz oscillator for an ethernet PHY I had a _LOT_ of noise on the clock line (track was about 1cm long) and was blaming it on a dodgy soldering. I will give it another shot for my new project and see if it was only bad luck last time.