What's the LDOs（VDDARM_CAP 、VDDSOC_CAP、VDDHIGH_CAP） start-up sequence within i.mx6 ？
Is the POR_B signal first or LDOs output first？
That depends on the version of the i.MX6 and the version of the PMIC you are using, for example, if you are using the i.MX6Q, you should use the MMPF0100F0AEP version of the PMIC as you can see in Table 1 of the MMPF0100 PMIC datasheet: https://www.nxp.com/docs/en/data-sheet/MMPF0100.pdf
Now, if you check Table 10 of the same PMIC datasheet, you can find the start-up sequence configuration for all the regulators.
And from the I.MX6 point of view, you can typically find the requirements of the power up sequence in the device datasheet, for example, in section 4.2.1 of the i.MX6Q datasheet (https://www.nxp.com/docs/en/data-sheet/IMX6DQCEC.pdf), you can find the following sentence:
For power-up sequence, the restrictions are as follows:
But my board (2/4) can't startup,and I added 400ms delay before POR.
The LDOs has no output and the clock（32K） did not oscillate.
i.MX6DL and MMPF0100F0AZES
I’m not sure if I’m properly understanding, but when you mention “my board (2/4) can't startup” you are saying that 2 of 4 of your boards are presenting this same issue? Please confirm, this could be important and could tell us that there are no problems with the design, but with the part itself (like a quality problem) or manufacture issue (like soldering profile).
You also mentioned “The LDOs has no output”, please confirm which LDO you are referring to.
It is also important to know the HW configuration of the VDDOTP pin, connecting the VDDOTP pin to VCOREDIG through a 100 kΩ resistor. And using this configuration, LDOs VGEN1 and VGEN 3 are disabled as you can see in Table 9 of the datasheet: https://www.nxp.com/docs/en/data-sheet/MMPF0100Z.pdf
If a quality problem，please confirm the reason and What part of it might have caused it?
“The LDOs has no output”——VDDARM_CAP 、VDDSOC_CAP、VDDHIGH_CAP，32K and 24Mhz can't start too！
the VDDOTP pin followed evaluation board “MCIMX6DL-SDP”.
Quality problems are rare and typically affects more complete lots, this is why I asked to clarify your “my board (2/4) can't startup” comment, are you saying that 2 of 4 of your boards are presenting this same issue? Please confirm, this could be important and could tell us that there are no problems with the design, but with the part itself (like a quality problem) or manufacture issue (like soldering profile).
VCOREDIG hardware pin configuration is correct.
Only one 22 mF bulk capacitor should be connected to each of these on-chip LDO regulator outputs:
If the nominal capacitance value is larger than recommended, power-up ramp time is excessive and operation cannot be guaranteed. Note that the ramp up time is constant. Larger capacitors mean more inrush current. Select small capacitors with low ESR (equivalent series resistance).
Do not connect any loads to these LDO outputs: VDDARM_CAP, VDDARM23_CAP, or VDDPU_CAP. VDDSOC_CAP is restricted to MX6 loads.
Only one 10 mF bulk capacitor should be connected to VDD_HIGH_CA.
If the nominal capacitance value is larger than recommended, power-up ramp time is excessive and operation cannot be guaranteed. Select small capacitors with low ESR.
These LDOs should only be used to power the loads as described in the reference manual or data sheet. Do not connect any loads to these LDO outputs: NVCC_PLL_OUT or VDDUSB_CAP.
VDDHIGH_CAP is restricted to MX6 loads.
Are you sure 22mF?
But my reference design is 22uF！
I can give you a answer with the following question YES！
“my board (2/4) can't startup” comment, are you saying that 2 of 4 of your boards are presenting this same issue?
But what part of it might have caused it? i.max6 itself？
You are correct, 22uF is correct.
If only of 4 of your boards are presenting the failure, I don’t think the problem could be related to a quality issue, since this kind of issues are typically affecting the complete batch, so, only if the failure is present on 4 of your 4 boards, we could think in a quality issue.
Quality issue are the ones that are present since the fabric, but this does not seem to be the case.
One thing that I would recommend you to do is a swap to make sure that the problem is with the i.mx6 itself of with the board or with the manufacturing/soldering process. Get a new sample of the i.mx6 or one from another board that you know it is working properly and solder it to a failing board to test it.
Retrieving data ...