• Please review our updated Terms and Rules here

Life expectancy of PLDs?

8

8008guy

Veteran Member
Joined
Jun 25, 2014
Messages
584
Location
Salt Lake City, Utah
I have a few boards that have PLDs (Programmable Logic Devices). Looking at the spec sheets I see that most show a life expectancy of 20 years before they lose their programming. My clearpoint memory board from the 80's has a dozen or so.

Are you guys seeing failures of PLD logic?
 
8008guy said:
I have a few boards that have PLDs (Programmable Logic Devices). Looking at the spec sheets I see that most show a life expectancy of 20 years before they lose their programming. My clearpoint memory board from the 80's has a dozen or so.

Are you guys seeing failures of PLD logic?
Click to expand...

It would depend a lot on the type of PLD. Are they one-time fuse programmable, or UV EPROM, or EE EPROM based.

I suspect one-time fuse-based devices will last as long as similar logic devices might last (ie, 74Sxxx, etc).

Whereas UV or EE reprogrammable devices have a lifetime of the programmable element which may decay over time.

So one would have to know the exact manufacturer and device type to give a more specific answer.

I have many PDP-11 boards that are 30-35 years old built with one-time fuse programmable PALs that work just fine.

Don
 
Last edited:
8008guy said:
PLS153AN plds

There are also two ncr chips. Ncr0380971 and ncr0380968, which I believe are gate arrays.
Click to expand...

The PLS153 devices use Ni-Cr one-time programmable fusible links. Once they are blown they stay blown.

As to the gate arrays, they are built with specific metal masks that implement the desired functionality.
They are not programmed in the classic sense of a programmable device.

My experience is I have not seen an such PLD type failures on any of my 30+ year old boards.

Highest failure rate I see occurs in the 7474 classic ttl dual flip flop. Maybe just because there are lots of them.

Don
 
Hi All;
Don, Can You clarify, are You saying that the classic 7474's are PLD's or just that You have seen a high Failure rate, the same as M_Thompsom has mentioned in His postings on the PDP 8's ??

THANK YOU Marty
 
I think Don has seen quite a few 7474 failures as have I. The 7474 is a straightforward small scale bipolar integrated circuit (non-programmable). The early programmable ICs used metal fuse technology which weren't themselves prone to failure though the circuitry would have the same reliability as any other of the period. Later on as technology moved to MOS and current handling reduced fuse technology moved towards poly silicon fuses which early on had a nasty habit of regrowing especially if they were incompletely blown during the programming.
Off-topic slightly there was a comment on another thread about DEC designers making use of an undocumented feature of some manufacturers' 7474s in their designs. I don't think this would aggravate any likelihood of failure. I haven't seen any documentation as to which brands are more likely to show failures but since the individual designs and processing are quite different it's almost certain it will be down to design or layout of a particular brand.
Actually I continue to be quite surprised by the long-term reliability of the early plastic-packaged TTL. ISTR that expected lifetimes then for commercial ICs were more in the range of 15 years than in the 45-year range we're seeing just now, especially with equipment which has been stored in pretty unsatisfactory conditions. I do seem to see quite a few failures on ICs interfacing to the bus though so I do sometimes wonder if the (pdp8 omnibus) bus integrity is all it might be.
 
EPROMS start to loose their electric charge on the insulated gate over time. This is a known failure mechanism. Before the EPROM fails (hard) it starts to fail (soft) where occasionally it misreads a location and the program crashes - or the machine resets. It is also sensitive to light, electromagnetic interference and other environmentally-related effects. I have seen such failures like that in equipment at work.

Programmable Array Logic that utilises Nichrome fuse technology also suffers a failure mechanism known as "fuse regrowth". The fuse can actually regrow - although this may be related to the fact that the fuse may never have actually have been 'blown' properly by the programmer originally. There is a US Patent describing this failure mechanism and how it has been overcome in later designs. Again, before the PAL fails hard - it may exhibit soft failures as the fuse 'makes' and the one atom 'fuse' is 'burnt away' by the current passing through the 'fuse' just as it makes. You get a glitch in the logic as this occurs. We have identified a number of 25-30 year old boards at work that have failed because of this mechanism. Swapping the PAL for a freshly 'burnt' one fixed the board.

There is also an internal IC failure mode known as electromigration where metal atoms migrate in the presence of an electric current (http://en.wikipedia.org/wiki/Electromigration). This affects not only programmable devices but also conventional 'non programmable' devices (such as 7474 etc.). Usually the higher the current- the worse the effect.

Dave
 
Hi All;
Dave, Thank You for the information..
"" EPROMS start to loose their electric charge on the insulated gate over time. This is a known failure mechanism. Before the EPROM fails (hard) it starts to fail (soft) where occasionally it misreads a location and the program crashes - or the machine resets. It is also sensitive to light, electromagnetic interference and other environmentally-related effects. I have seen such failures like that in equipment at work. ""
Dave, is there a Fix for This ??

THANK YOU Marty
 
Marty,

Yep - erase the EPROM and reprogram it...

Unfortunately, we are at the mercy of physics. You can't contain electrons where you want forever.

It may be possible to just reprogram the EPROM with the same contents - but I prefer to either (a) use a new EPROM (if available) or (b) a complete UV erase and reprogram cycle.

This is why it is vitally important to preserve the EPROM contents whilst you still can and prepare for the day when the EPROM(s) start to fail from 'old age'. Full marks to the websites that are already preserving 'bits' for posterity.

Dave
 
You must log in or register to reply here.
Back
Top