Other Parts Discussed in Thread: TPS65217, TPS65218
Hi Team,
My customer is seeing some failures with the DDR on their board. They detected the DDR voltage to be between 2.9 to 3.1V. It is not always that high, but it seems to happen randomly on some units. Here are more details:
1. PMIC used - TPS65217D
2. Processor used - AM3351
3. DDR used - Hynix H5TC4G63CFR
4. The PMIC has a default output voltage of 1.35V on DC-DC1. This is routed to the DDR. The rail is not modified by SW.
We have already been in discussion with the PMIC team on the issue (see this thread). For quick reference, I've outlined the steps already taken below:
1. Reviewed schematic of the PMIC section and confirmed everything checks out good.
2. Since the failure is random, capturing data of this occurrence is extremely difficult. The team has been unsuccessful in understanding the exact conditions of when the problem occurs. Having this data would be super helpful.
3. PMIC team tested one -D rev EVM in the lab to try and recreate issue, but results were inconclusive. Test procedure is outlined below:
- Use the default assembly of the TPS65217D EVM
- Applying 5V DC into AC pin
- Allowing initial power-on sequence to occur normally
- Monitor LDO1 (Strobe 15) and VDCDC1(Strobe 1)
- Press nRESET button (S1) and allow the PMIC to reset after 1 second
- Test repeated 100 times
- RESULTS:Max value recorded on VDCDC1= 1.44V
- No problem identified
In an effort to explore all options, I’d like to review the customer's timing diagram with the Sitara team and ensure everything on this end looks good. We are wondering if, say, the sequencing of the TPS65217D does not match the AM3351 device exactly, could the timing of the rails potentially be the problem and the AM3351 leakage may be damaging the DDR memory?
Thanks,
Mitchell