MEMS oscillator

Technical discussions around xCORE processors (e.g. xcore-200 & xcore.ai).
Post Reply
Gothmag
XCore Addict
Posts: 129
Joined: Wed May 11, 2016 3:50 pm

MEMS oscillator

Post by Gothmag »

I'm looking to cut down a board I have and I have 0 experience with MEMS oscillators but from the looks of it I can essentially supply most of them power, ground, decouple it and I get my clock signal I can feed right to the CPU with a very small area taken up for the clock. Is there something I'm missing, will this work fine for a 200 series processor? I've typically used a pierce-gate clock circuit and I'm not seeing a reason to not switch over.


User avatar
Bianco
XCore Expert
Posts: 754
Joined: Thu Dec 10, 2009 6:56 pm
Contact:

Post by Bianco »

I have some prototypes here with ASTMLPE-25.000MHZ-EJ-E-T and ASFLMB-24.000MHZ-LY-T without problems.
MEMS oscillators do draw more current than regular crystal oscillators in general but they should also be a bit more affordable.
Gothmag
XCore Addict
Posts: 129
Joined: Wed May 11, 2016 3:50 pm

Post by Gothmag »

Excellent! The 24mhz abracon is on my short list so that's good to know.
User avatar
akp
XCore Expert
Posts: 578
Joined: Thu Nov 26, 2015 11:47 pm

Post by akp »

We use DSC1001CI5-024.0000 and it works as expected. They are more reliable than quartz, so that's a benefit even though they draw more power.
Gothmag
XCore Addict
Posts: 129
Joined: Wed May 11, 2016 3:50 pm

Post by Gothmag »

Just as a follow up. I decided to check stability and check specs between a pierce clock using ABLS7M2-24.000MHZ-D-2Y-T and DSC1001AL5-024.0000 after hearing they work well. Probably use a smaller footprint of the MEMS clock for final board but for proto the large package is easier. The difference between the images is time/div from 20uS down to about 10nS. They all showed that the MEMS was more stable overall but sort of expected compared to a 20ppm freq. tol./30ppm temp. tol. crystal circuit. For the price and small footprint I'm pretty happy with it. Probing wasn't ideal in either case but it was the same for both, running from same 3.3v power supply. Also for the 10nS/div the temperature changed about 10* C.
Image
Large timebase - 20uS
Image
Medium timebase - 40nS
Image
Small timebase - 10nS
Post Reply