I'm trying to bring up a custom board using MIMXRT1024CAG4B. The chip draws >1A, causing the 3.3v DCDC supply to go into short-circuit protection mode and limit current to ~100ma and the RT1024 gets quite hot. I see other folks (@mjbcswitzerland, @jaymonkman) have had trouble like this with RT series parts when the DCDC converter does not start properly and/or shuts down (and no solutions were posted):
https://community.nxp.com/t5/i-MX-RT/DCDC-converter-troubles/m-p/1170855
https://community.nxp.com/t5/i-MX-RT/Large-current-surge-problem-during-power-sequencing-of-i-MX-RT/...
The errata for the RT1024 shows no power issues unlike some other RT series parts.
The design exactly follows the RT1024-EVK (power sections shown below). DCDC_Pswitch is a 30k/.22uF delay. All bypass caps are mounted per design guidelines and EVK. All power 3.3v supply inputs for DCDC and other power domains are provided by the same source. 3.3v is supplied by a Recom 1A DCDC R-7833 supply. The other loads on the 3.3v supply are tiny except inrush on bypass caps. USB Vbus is from a separate 5v DCDC (VDD_USB_CAP is 2.6v but should be 3.0v)
Any ideas on where to start would be appreciated!
Thanks,
Best Regards, Dave
Solved! Go to Solution.
Hi David
My problem was related to the i.MX RT 105x HW errata and was exasperated by a glitch when switching through power, which was not related to the i.MX RT. In any case with accurate control of the timing and control of surge it was solved.
In the meantime a number of designs with various i.MX RT parts (without the HW errata and including B versions) have not had any issues.
In your case I don't understand the DCDC output connection:
The voltage on C26 will ramp up to 1.1V when the internal DC/DC converter is triggered (by a rising edge on the DCDC_SWITCH input). This is then used to feed the VDD_SOC_IN power inputs.
I don't understand why it is labeled with V3V3. If it is really connected to the same power supply as its input it would be short circuited to its input and also the core voltage (VDD_SOC_IN) would be short circuited to the input voltage instead of the DCDC converter's 1V1 output, exceeding its absolute maximum voltage of 1.6V.
Can you clarify the VDD_SOC_IN voltage rail circuit?
Regards
Mark
Hi David
My problem was related to the i.MX RT 105x HW errata and was exasperated by a glitch when switching through power, which was not related to the i.MX RT. In any case with accurate control of the timing and control of surge it was solved.
In the meantime a number of designs with various i.MX RT parts (without the HW errata and including B versions) have not had any issues.
In your case I don't understand the DCDC output connection:
The voltage on C26 will ramp up to 1.1V when the internal DC/DC converter is triggered (by a rising edge on the DCDC_SWITCH input). This is then used to feed the VDD_SOC_IN power inputs.
I don't understand why it is labeled with V3V3. If it is really connected to the same power supply as its input it would be short circuited to its input and also the core voltage (VDD_SOC_IN) would be short circuited to the input voltage instead of the DCDC converter's 1V1 output, exceeding its absolute maximum voltage of 1.6V.
Can you clarify the VDD_SOC_IN voltage rail circuit?
Regards
Mark
You're absolutely right Mark, DCDC_OUT/VCC_SOC_IN was supposed to be a separate net.
Blunder....
Thanks!