【Semiconductor】18. What Was PSRAM Intended to Achieve?
— The Premise of Reusing DRAM
topics: [“Semiconductor”, “PSRAM”, “DRAM”, “Memory”, “Architecture”]
🧭 Introduction — Why Did PSRAM Become an Option?
Around the year 2000,
a memory called PSRAM (Pseudo-SRAM) was introduced to the market.
Its external interface looked like SRAM,
while DRAM cells were used internally.
In this article, we organize:
- What PSRAM was intended to solve
- The structural assumptions on which it was built
Discussions of failures and limitations are deferred to the next article.
📱 Background of PSRAM — Constraints of Mobile Systems
In the late 1990s,
systems for mobile devices faced the following requirements simultaneously:
- A memory with simple external control
- Extremely low standby power consumption
- Cost kept at the level of DRAM
However, the available options at the time were polarized.
| Type | Characteristics |
|---|---|
| SRAM | High speed, simple control, high cost |
| DRAM | High density, low cost, complex control |
PSRAM was conceived to fill
the gap between usability and cost.
🎯 The Core Concept of PSRAM
The central idea of PSRAM was extremely simple.
Keep the inside as DRAM,
but make it look like SRAM from the outside
Rather than abandoning DRAM and creating a new type of cell,
the premise was to maximize reuse of existing DRAM technology.
🏗 Key Points of the Internal Structure
① Using the DRAM Cell Array as Is
- The storage element is a standard DRAM cell
- High density and low cost are exactly those of DRAM
There is no special physical design
unique to PSRAM at the cell level.
② Internal Refresh Control
The defining feature of PSRAM is that it:
- Automatically executes refresh internally
- Eliminates the need for external refresh control
As a result,
it can be accessed externally with the same ease as SRAM.
③ An Access-Priority Control Policy
The internal control policy was:
- Access takes priority when requests are present
- Batch refresh is performed during idle periods
This choice was reasonable for mobile applications because it:
- Avoids access latency
- Reduces average power consumption
🧩 Assumed Usage Conditions
PSRAM was designed as a
memory dedicated to mobile applications.
The assumed usage conditions were:
- Long standby periods
- Intermittent access
- High-temperature environments (up to ~90 °C)
- Strict low-power requirements
These differed clearly
from the conventional usage patterns of DRAM.
⚖️ The Decisive Difference Between DRAM and PSRAM
The critical point here is the following:
PSRAM uses DRAM cells as-is,
but it is not used in the same way as DRAM
- DRAM
- Is always refreshed at fixed intervals
- PSRAM
- Refresh intervals depend on usage patterns
This difference would later take on major significance.
✅ A “Rational Design” for Its Time
The structure of PSRAM was:
- Not technically radical
- Compatible with existing DRAM processes
- Directly aligned with market requirements
In this sense,
it was an extremely rational design for its time.
At this stage, words such as:
- Failure
- Limitation
- Withdrawal
had not yet become reality.
🧾 Summary (Structure)
PSRAM was a memory with a
clear and consistent objective:
- Reuse DRAM technology
-
Provide the usability of SRAM
- Not a special cell
- Not based on special physics
- Only the usage model was different
The problem was that
this usage model quietly violated
the physical assumptions of DRAM.
🔗 Primary Sources (References)
-
Legacy Technology Archive
https://samizo-aitl.github.io/Edusemi-Plus/archive/legacy/ -
PSRAM (2001) Cases
https://samizo-aitl.github.io/Edusemi-Plus/archive/legacy/psram_2001/ -
PSRAM Architecture
https://samizo-aitl.github.io/Edusemi-Plus/archive/legacy/psram_2001/psram_architecture/
⏭ Next
In the next article,
we will cover the failures that actually surfaced in PSRAM.
Why Pause and Disturb
became problematic in combination.
No evaluation, no hindsight—
only the facts.