Fast – Precise – Open for many 8-/16-bit systems like C64, C128, Amigas, Atari 800XL, 1040ST and others, Apple IIe, Spektrum and many more
A fast, open-source DIY DRAM tester for vintage RAM chips used in C64, Amiga, Atari, and other retro computers. Tests memory thoroughly in seconds and supports a wide range of chip types with optional OLED feedback.
As featured on:
- Hackaday (Dec 2025): "Cheap and Aggressive DRAM Chip Tester"
- Elektor Magazine (Dez 2025): "Ram-Tester Is an Open-Source DIY Solution for Retro Computer RAM"
.jpeg)
Most Arduino-based DRAM testers need up to 2 mins or even more for a single 41256 RAM and check only basic functions.
This project completes a full memory, address, data-line and retention time test in 15 s or less. (or even 2.5sec faster without the Display).
It is probably the fastest Arduino solution which also covers static-column DRAMs, cell retention time and 20-pin ZIP packages.
| Feature | Benefit |
|---|---|
| Most tests are <= 10s | Rapid diagnosis on the workbench or at retro repair events |
| Retention-time measurement | Detects weak chips by checking it meets the min retention times |
| Static-column support | Reliable testing of 44258, 514402 static column functions |
| Nibble Mode support | Testing 41257 RAM with nibble Mode access |
| 20-pin ZIP socket | Direct test of 20-pin ZIP DRAMs without an adapter |
| Optional OLED display or LED blink codes | Full text feedback or minimal hardware setup |
| Open hardware and firmware | KiCad, Gerber files and Arduino source under an open licence |
| Self Test Mode | Check the HW for any defects like short circuit, broken soldering, etc |
| Capacity | DIP | 20-pin ZIP | Static column | Nibble Mode | Retention Time | Test Time |
|---|---|---|---|---|---|---|
| 4 K x 1 | 4027 1) | - | - | - | 2ms | 1.3sec |
| 16 K x 1 | 4816 | - | - | - | 2ms | 1.6sec |
| 16 K x 1 | 4116 1) | - | - | - | 2ms | 1.6sec |
| 16 K × 4 | 4416 | – | – | - | 4ms | 4.2sec |
| 32 K x 1 | 3732 2) | - | - | - | 4ms | 2.8sec |
| 32 K x 1 | 4532 3) | - | - | - | 4ms | 1.9sec |
| 64 K × 1 | 4164 | - | – | - | 4ms | 2.8sec |
| 64 K × 4 | 4464 | – | – | - | 4ms | 6.4sec |
| 256 K × 1 | 41256 | - | – | 41257 | 4ms | 7.4sec |
| 256 K × 4 | 44256 | both | 44258 | - | 8ms | 4.1sec |
| 1 M x 1 | 411000 | !NO! | - | - | 8ms | 25.9sec |
| 1 M × 4 | 514400 | both | 514402 | - | 16ms | 12.8sec |
1)requires the 4116 adapter board.
2)currently only -H Type verified with real RAM HW
3)only verified by simulation with 4164 due to current unavailability of real RAM Chip
Note: Above test times include the OLED Display. Without Display, the test durations are approx 1 sec shorter.
- Insert the device (16, 18 or 20 pins, DIP or ZIP).
- Dial the needed DIP Switch to ON - depending on the number of Pins of your RAM
- Connect USB power supply.
- Read the result
- OLED version: plain-text report on the display
- LED-only version: green = pass, red = fail
There is a short YouTube video demonstrating the tester in action.
- GND Shorts, if you RAM has a Short to GND on any Pin
- Power Supply Shorts - well there is a resetable fuse to protect the board
- Addressline or decoder faults
- Stuck Cells or Crosstalk by using various patterns
- Random patterns combined with retention time checks
- CAS-before-RAS Refresh Timer Function
- All of the above uses Fast Page Mode, Static Column or Nibble Mode control depending on RAM Chip type
Here is the analytics of this algorithm vs. March-B
| Aspect | Actual Code | MARCH-B |
|---|---|---|
| Pattern Coverage | ✅ 0x00, 0xFF, 0xAA, 0x55 | ✅ |
| 0, 1Transition Tests | ✅ Thru Pattern-Sequence | ✅ Thru R0W1, R1W0 |
| Address Sequence | ✅ Asc-/Descending | |
| Coupling Detection | ✅ By Retention-Delay | ✅ Systematically |
| Real Retention | ✅ 2-16ms Tests | ❌ Only µs-Range |
Compared with many other open source projects:
- They usually don't test for address line faults. Tests will pass even if you bend one address pin up.
- They usually don't check multiple rows at a time, since they use slow Arduino read and write commands.
- Many use simple all 0 and all 1 tests, no patterns or random data tests.
- No retention tests possible since writing and reading one row alone with Arduino I/O takes more than one second - usual retention times are within a few milliseconds.
- No checks for broken chips (i.e. shorts to GND).
- No protection against short circuits on supply lines. Some use off-the-shelf DC-DC converters which supply high currents in short circuit situations.
- No tests of refresh or static column functionality.
- Very often limited to 1-bit or 4-bit circuits and no ZIP sockets.
- They are usually MUCH slower.
Sales thread on Amibay: https://www.amibay.com
Product Listing on Tindie: https://www.tindie.com
Product Listing on EBAY: https://www.ebay.ch
DIY order at PCBWay for Thru-Hole Version: https://www.pcbway.com
Use the provided Gerber Files from this Repo
- Wiki – assembly and operating guide
- Software – source code
- Schematic – KiCad project and Gerber files
- Changelog
Pull requests, issues and forks are welcome.
Questions: GitHub Discussions or contact tops4u on AmiBay.
Open-source hardware under GPL v3 – use at your own risk.