3
\$\begingroup\$

I have a mask ROM in a synthesizer module whose contents no longer match the expected CRCs, with it likely being just a single bit error. The ROM is circa 20 years old.

Assuming this is not a manufacturing fault, what are the mechanisms by which a mask ROM might degrade, and do mask ROMs have an expected lifetime?

\$\endgroup\$
7
  • \$\begingroup\$ Which synthesizer and ROM version? Can you compare to a working ROM? \$\endgroup\$
    – Justme
    Commented Aug 25, 2021 at 20:37
  • \$\begingroup\$ @Justme this doesn't seem to be a repair question \$\endgroup\$
    – Criticizing Israel not allowed
    Commented Aug 25, 2021 at 20:52
  • \$\begingroup\$ I'd speculate thermal/background-radiation-induced atom migration and dissolution. \$\endgroup\$
    – rdtsc
    Commented Aug 25, 2021 at 20:59
  • \$\begingroup\$ @user253751 I understand the question was triggered by an assumption that the ROM actually is faulty, but what if the ROM is not faulty but some data or address bit between the CPU or just power supply is faulty? Knowing if the ROM contents read with a ROM reader matches to some known reference will help to see if the ROM is not the faulty part to begin with, but something else? \$\endgroup\$
    – Justme
    Commented Aug 25, 2021 at 21:07
  • \$\begingroup\$ @Justme the ROM is actually a plug-in 72 PIN SIMM format ROM for an E-mu Proteus 2000. I built a standalone card reader that on other "failed" ROMs has detected single bit errors when compared to other working ROMs. \$\endgroup\$
    – Alnitak
    Commented Aug 25, 2021 at 21:49

2 Answers 2

Reset to default
1
\$\begingroup\$

"Read-only memory" from WikiPedia:

Masked ROM and fuse/antifuse PROM do not suffer from this effect (data retention related), as their data retention depends on physical rather than electrical permanence of the integrated circuit, although fuse re-growth was once a problem in some systems

I agree with that. The problem might come from the physical (external) or electrical stress first, but not from the self lifetime of the ROM. Cleaning the board and re-soldering dry joints may help.

\$\endgroup\$
2
  • 1
    \$\begingroup\$ joint issues would cause an entire data line or address line to act weirdly, not single digits of bits. \$\endgroup\$
    – Alnitak
    Commented Aug 25, 2021 at 21:53
  • \$\begingroup\$ @Alnitak, Curios, I was not clear: Does the diagnostics says a single bit failure? About the cleanup and soldering; Bad contacts can cause marginal signal, if not fail at once. \$\endgroup\$
    – jay
    Commented Aug 25, 2021 at 22:58
1
\$\begingroup\$

The failure mechanisms for masked ROMs are no different from the failure mechanisms of the the logic circuits that surround them. There is nothing special about the transistors or circuits of a masked ROM. Manufacturing defects can be latent, and not cause a fault for many years after the product is put into service.

Yes, masked ROMs have an expected lifetime. The value is different for every masked ROM that is made and depends very heavily on the quality of the manufacturing process, the design rules used in physical layout, and the expected operating environment.

\$\endgroup\$
2
  • \$\begingroup\$ IIUC correctly, you're saying that in effect a single failed transistor (which is not at all surprising after 20 years) could cause the bit associated with that transistor to change state? \$\endgroup\$
    – Alnitak
    Commented Aug 25, 2021 at 21:52
  • 1
    \$\begingroup\$ Absolutely. Each bit in a mask ROM is represented by a single transistor. Any fault associated with that transistor, or the connections to that transistor, could cause the effective value of the stored data to change. \$\endgroup\$
    – Elliot Alderson
    Commented Aug 25, 2021 at 23:37

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.