RT1052 1V2 rail

cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 

RT1052 1V2 rail

Jump to solution
2,964 Views
martinsmith
Contributor III

Hi,

I am bring up a custom PCB and the SNVS, DCDC and 3V3 rails all come up OK in the right order. As a result I was expecting the 1V2 rail to come up to power the M7 core, but it doesn't. I have the inductor and caps connected right but nothing. Is there anything I have to do to enable it?

Thanks 

Labels (1)
Tags (1)
0 Kudos
1 Solution
2,394 Views
martinsmith
Contributor III

UPDATE: I populated another PCB and all the rails come up OK now. I guess I can put this one down to manufacturing issues?

Thanks for the advise

View solution in original post

0 Kudos
14 Replies
2,395 Views
martinsmith
Contributor III

UPDATE: I populated another PCB and all the rails come up OK now. I guess I can put this one down to manufacturing issues?

Thanks for the advise

0 Kudos
2,394 Views
igorpadykov
NXP Employee
NXP Employee

pastedImage_2.jpg

0 Kudos
2,394 Views
igorpadykov
NXP Employee
NXP Employee

Hi Martin

one can check design, also paying attention to clocks, using

sect.3. Power supply, sect.4. Clocks

Hardware Development Guide for the MIMXRT1050 Processor
https://www.nxp.com/docs/en/user-guide/MIMXRT1050HDUG.pdf

Best regards
igor
-----------------------------------------------------------------------------------------------------------------------
Note: If this post answers your question, please click the Correct Answer button. Thank you!
-----------------------------------------------------------------------------------------------------------------------

0 Kudos
2,394 Views
martinsmith
Contributor III

Hi, thanks for your help.

I've been looking at it this morning. The SNVS rail (3V3) comes up first, along with the DCDC rail (3V). I can see the DCDC start to come up (PSWITCH is high around 1ms after DCDC). The 1V2 rail (VDD_SOC_IN, aka the output from the internal DCDC) rises to around 1V then gets very noisy. At this point the DCDC rail dips and the 1V2 rail collapses. 

I have had a good look through the Hardware Development Guide for the MIMXRT1050 Processor and can't see what i'm doing wrong. 

The DCDC rail just goes to balls L1, L2 and K4, and the 1V2 rail coming out of the inductor / Cap filter just goes to the VDD_SOC_IN balls and the DCDC_SENSE ball (J5). I have all the required caps fitted around the RT1052.

I am using rev B silicon if that makes any difference?

0 Kudos
2,394 Views
igorpadykov
NXP Employee
NXP Employee

so do you have working clocks in XTALI/XTALO, RTC_XTALI/RTC_XTALO,

may be useful to look at Figure 45-1. Power system overview i.MX RT1050 Processor Reference Manual
https://www.nxp.com/docs/en/reference-manual/IMXRT1050RM.pdf

sect.4.2.2 Integrated LDO voltage regulator parameters, sect.4.2.4 On-chip oscillators

i.MXRT1050 Datasheet
https://www.nxp.com/docs/en/data-sheet/IMXRT1050CEC.pdf

Best regards
igor

0 Kudos
2,394 Views
martinsmith
Contributor III

No the clocks are not working - but I think this is because they are powered from VDD_SOC_IN? I don't have an RTC clock fitted anyway. 

0 Kudos
2,394 Views
igorpadykov
NXP Employee
NXP Employee

according to sect.4.2.4.2 OSC32K - it is powered from VDD_SNVS_IN.

from sect.4.2.4.1 OSC24M - it is powered from NVCC_PLL.

from Figure 45-1 - NVCC_PLL produced from NVDD_HIGH_IN through LDO_1P1.

So seems they all should work.

0 Kudos
2,394 Views
martinsmith
Contributor III

OSC32K appears to be the RTC - which I am not using (N9 tied to GND and P9 floating)

NVCC_PLL is connected to ball P10 (from Table 81) - which I can check and confirm has 1.1V on it. But VDD_HIGH_IN is powered up after DCDC so shouldn't be interfering with VDD_SOC_IN should it?

I will try to get the 24MHz clock working in the meantime

0 Kudos
2,394 Views
martinsmith
Contributor III

Update: Using the Eval PCB (MIMXRT1050-EVK) I have been testing the clock and VDD_SOC_IN rail.

1. I can see that the VDD_SOC_IN comes up around 270ms before the 24 MHz xtal starts oscillating

2. If I remove J36 to stop the VDD_SOC_IN rail from coming up the 24 MHZ xtal never starts ups. I can see repeated attempts about every 270ms where it tries to start but fails. 

On my PCB the XTAL never even tries to start.

So the DCDC seems to be able to interfere with the XTAL operation, even though it doesn't power it directly.

0 Kudos
2,394 Views
igorpadykov
NXP Employee
NXP Employee

NVCC_PLL does not depend on VDD_SOC_IN rail, it is produced from VDD_HIGH_IN.

Do you have VDD_HIGH_IN, VDD_HIGH_CAP, NVCC_PLL powered ?

For dcdc one also can look at AN12146 Table 2 "Power", notes for DCDC_PSWITCH

https://www.nxp.com/docs/en/nxp/application-notes/AN12146.pdf 

0 Kudos
2,394 Views
martinsmith
Contributor III

Yes VDD_HIGH_IN, VDD_HIGH_CAP and NVCC_PLL are all powered. I have looked at the app note and have made any adjustments but still no 24 MHz xtal output and no VDD_SOC_IN

0 Kudos
2,394 Views
igorpadykov
NXP Employee
NXP Employee

one can check PMIC_ON_REQ, as described in sect.20.5.1 Turn on DCDC i.MXRT1050 RM:
To turn on the DCDC, PSWITCH and PMIC_ON_REQ must be both high

0 Kudos
2,394 Views
martinsmith
Contributor III

Yes I can see that they are both high (3.3 volts). In the screenshot below the blue trace is the VDD_HIGH, VDDA_ADC_3P3, and DCDC rails (they are all joined together, the SNVS comes up 250 ms before). The yellow trace is the PSWITCH and the purple trace is the VDD_SOC_IN (1V2 rail). PMIC_ON_REQ is connected to the enable line of the DCDC reg so comes up just before the DCDC rail.tek00000.png

0 Kudos
2,394 Views
igorpadykov
NXP Employee
NXP Employee

actually in the screenshot above the blue trace is sagging below 2V which violates power-up sequence and
voltage levels for VDD_HIGH, DCDC rails (min. about 2.8-3.0V depending on chip revision).
They should be smooth and its levels comply with Table 9. Operating ranges i.MX RT1050 Datasheet
https://www.nxp.com/docs/en/data-sheet/IMXRT1050CEC.pdf
Reason for voltages sagging may be layout errors (short circuits, broken traces), too big capacitors
(check sect.7.2. Placement of bulk and decoupling capacitors Hardware Development Guide
for the MIMXRT1050 Processor), insufficient power supply current which feeds processor.

0 Kudos