Stresstest fails at 3 frequencypoints but passes with 1 frequencypoint, LPDDR4

[NikJur]

Hello everyone,

I had asked a similar question when i used the mScale DDR Test Tool and that error was resolved when switched to ConfigTool V15.

I am currently working on a board bring-up with the following combination of SOC, RAM and Software for the RAM-Stresstest:

- i.MX 8M Plus (1600 MHz) + Micron LPDDR4 4GB (MT53D1024M32D4DS)

- iMX Config Tool v24.12

I have used the NXP RPA Excel sheet for generating the according ".ds"-file for the Stresstest program. Also using the standard parameters in the "DDR Configuration Sheet" results in the same error. 

The question is as following:

Why does the Stresstest run smoothly and without errors, even under temperature fluctuations, when chosing 2000 MHz as the single frequency for the test, but fails when i chose the same parameterset, but now with three frequency points enabled. In that case the default, additional 200 MHz and 50 MHz, are tested as well. 

Any help is hugely appreciated

Greetings

Niko

[JorgeCas]

Hello,

It is suggested to run the stress test at the target operating frequency for an extended period of time to verify the DDR performance on the target board. Running the stress test higher/lower than the target operating frequency is not meaningful.

The signal waveform may change a lot as the frequency increases/decreases. Therefore, passing the stress test at a frequency higher/lower than the target operating frequency does not necessarily mean that there is more timing margin.

To pass the stress test you may need to change:

ODTImpedance

Desired ODT impedance in Ohm. Valid values for DDR4=240,120,80,60,40. Valid values for LPDDR4=240,120,80,60,40

TxImpedance

Write Driver Impedance for DQ/DQS in ohm (Valid values for all DDR type= 240, 120, 80, 60, 48, 40, 34)

ATxImpedance

Write Driver Impedance for Address/Command (AC) bus in ohm (Valid values for all DDR type = 120, 60, 40, 30, 24, 20)

Best regards.

[NikJur]

I have already extensively tested the ODT, PDDS, Vref etc. For our highest target frequency i can get the stresstest running succesfully, without errors, when its the only tested frequency -> i assume our layout is good, because the most critical operating point works. Also in the first test run when 200 and 50 MHz are enabled. Only from the second run errors arrise in the highest frequency operating point as well -> i assume some the firmware or the Config Tool dont work as expected:

So you´d test the 2000 MHz/200 MHz and 50 MHz seperately and then just enable the 3 frequencies in the RPA sheet? 

[JorgeCas · Accepted solution]

Hello,

If selecting a frequency range to test, ensure that the start frequency test is within ±50MHz of the frequency that the DDR initialization script is tuned for and make sure end frequency is no less than start frequency and within 100Mhz of the start frequency.

Best regards.