Hi,
Everyone, I've just checked the reference design and i.MX53 user guaide, the contradiction is that the ref connecting to ECKIL but the user guide suggesting to CKIL.Which one is better?
Best Regards
Rain
Solved! Go to Solution.
Hello,rainustc,
There are 2 kinds of methods to connect 32.768k to i.mx53 , you can select one of the following :
ECKIL is 32.768k OSC which is from PMIC.
OSC or Crystal can be used here .
Regards,
Weidong
Hello,rainustc,
There are 2 kinds of methods to connect 32.768k to i.mx53 , you can select one of the following :
ECKIL is 32.768k OSC which is from PMIC.
OSC or Crystal can be used here .
Regards,
Weidong
Hi, so we still have some questions pertaining to the 32KHz clock. We had our Arrow rep contact Freescale and here was the response in regards to CKIL vs. ECKIL:
From the i.MX53 Team:
CKIL is input, ECKIL is output. Generally it is recommended to use input.
But ECKIL is weak output, that can be externally (over)driven.
1. If CKIL is the input and is recommended by the above Freescale engineer, why do the reference designs have the PMIC 32KHz output going into ECKIL?
2. Is CKIL synonymous with ECKIL (internally)?
3. Does the imx53 need to be configured or does it pick between ECKIL and CKIL automatically?
Below is Figure 18-5. from page 825 in the reference manual...
There is no mention of ECKIL in the clock tree diagrams, so is it safe to assume CKIL and ECKIL are synonymous?
Basically we've had some issues on a couple boards and we want to route the external 32KHz into CKIL instead of ECKIL. So far, everything works fine when we prototype this change, but would like more clarification if anyone has any. Mainly some info on SYNC block the CKIL goes into and if we have to do anything in the kernel to make the change from ECKIL to CKIL.
We are also contemplating hanging a crystal right on the arm instead of relying on the PMIC. Are there any software changes that need to be some to accommodate this or does it happen automatically? I've prototyped the change and it seems to work fine without any software changes.
Thanks
Hi,
What is ECKIL used for? It is only mentioned in the reference manual once in all 5000 pages. Manual says signal use: Pervasive.
Is it necessary?
Thanks
EDIT:
Looks like it is necessary…
CKIL is used in conjunction with ECKIL if you hang a crystal on the imx directly. ECKIL is used for feeding external clock like from the PMIC. In this case CKIL is left floating.
Here's a good nugget of information about the clocks for the imx53...http://cache.freescale.com/files/32bit/doc/data_sheet/IMX53IEC.pdf
A little snippet...
The i.MX53 device has four external input system clocks, a low frequency (CKIL), a high frequency (XTAL), and two general purpose CKIH1 and CKIH2 clocks.
The CKIL is used for low-frequency functions. It supplies the clock for wake-up circuit, power-down real time clock operation, and slow system and watch-dog counters. The clock input can be connected to either external oscillator or a crystal using internal oscillator amplifier.
EDIT2:
The reason I'm looking into this is because were having some boards exhibit slow boot. We've narrowed it down to the PMIC switching between internal and external crystal. We're using the MC34708 PMIC, which has a rather large errata and is not recommended for new designs.
I'm willing to bet that hanging a crystal on the imx is better than relying on the PMIC. Furthermore, if you're not relying on the PMIC RTC function then using its internal oscillator (with 60% error rate) to feed the imx should be good enough.