How To Measure DDR Memories? (DDR5 / DDR4 / DDR3)

Explains how to connect an oscilloscope to DDR bus, what signals to measure and what to look for. Thank you very much Randy White from Keysight.
Links:
- How to test DDR5 Transmitter Compliance: keysig.ht/6tpd01
- How to Characterize and Test DDR5 Receiver Compliance: keysig.ht/OQ2MiJ
- FEDEVEL courses: fedevel.com/courses
Chapters:
00:00 What this video is about
00:29 The setup
04:25 Bit error ratio tester
12:15 Probing DDR5 / DDR4 / DDR3 memory signals
19:43 What software to run during DDR memory testing
23:07 Connecting and setting up oscilloscope to measure DDR memories
37:15 Interposer effects, equalization and de-embedding
46:36 Recognizing read and write cycles
52:56 Equalization in oscilloscope
58:04 Measuring and verifying DDR5 signals
1:12:40 Starting the automated test
------------------------------------------------------
Would you like to support me? It's simple:
- Sign up for online courses hosted on our platform: fedevel.com/
- You can also support me through Patreon: / robertferanec
It is much appreciated. Thank you,
- Robert

[@christopherjackson2157]

Randy white is someone worth following. Theres a great interview with him on the eecosystem podcast about ddr5

[@boonlau4171]

@15:00 I am wondering how the module going to inset into memory slot when there are Kapton tapes stick on the gold fingers.

[@ahyungrocks5509]

Towards the end, a test was run with all the options selected. Looks like there are only one measurement for each timing parameter. Will that sufficient?

[@DaSmik101]

I'd have a question for Randy :) Why did Keysight go the route of having the entire length of the flex tip in microstrip and as a passive element instead of having the amp on this tip? These flex tips will pick up a looooot of noise for your board.

[@hedleyfurio]

Very interesting - is the summary “ don’t try and design a PCB with DDR5 unless you can afford a 16GHZ scope and all the probes and fixtures ? “

[@hedleyfurio]

Wow I just googled that scope - STARTING PRICE $ 530,000 - I guess I will be sticking with a rigol 200Mhz , and just hope they keep manufacturing slow chips 😊

[@muhdiversity7409]

Would have been nice to introduce the guy, who he is and who he works for? I guess Keysight ? Private contractor ?

[@RobertFeranec]

Randy White from Keysight. More information can be found in description.

[@muhdiversity7409]

thanks @@RobertFeranec

[@girideshsekar7465]

Hi, I'm new to hardware design, can anyone please suggest me any resource to learn about DDR, also I'm pretty sure there will be alot of experienced hardware designers out there, is yhere any book or website that you would suggest for a beginner, mid level experienced me.

[@RichardKCollins]

You have some really beautiful noise in those signals, but the cost and time and care to work with it ... especially when all the examples I find are people like you trying to remove it, not study it in its own right. I am looking for innovative ways to use memory chips for gravitational detectors. Time of flight electron interferometry with lots of data, if only the analog signals were not so hard to get to. The sun moon tidal signal is large in gravitational terms. Then atmospheric density and temperature variations. Then a few ways to measure speed of gravity. Seismic events, real time imaging, earth and solar 3D imaging (interiors). Besides memory chips which are just marginal, there are entanglement detectors, some of the more sensitive cameras, and many "quantum", "pico" and "femto" experiments. Pretty much all the fields now are pushing deeper into the heart of electronics and there everyone should be starting to pick up gravitational variations. The atomic clocks change with the potential, but the clocks are small and cheap enough to make gravimeters, gravity gradiometers and time of flight detectors for mass movements with precise times and directions and locations. I classify the memory chips (particularly floating gates) and camera sensors (particularly LWIR, submillimeter, THZ, XUV and soft x-ray) under "electron interferometry". Then there are atomic microscope and atom interferometers and other sensors that groups have used but only gotten to "earth tides". When more groups start picking up "time of flight gravity signals" from cars and trucks on the road, detailed mapping of atmospheric density and temperature in the atmosphere and oceans, ocean waves, and many seismic events, then some of those sources - once calibrated - can be used as standard sources to calibrate newer sensors. Most of gravitational engineering can be done using electronics and electrodynamics. It is not mysterious, it is just small signals and lots of data to start. Richard Collins, The Internet Foundation