TTL-Pin corrosion as fault indicator?

[EEngo]

Hi!

I recently got a Tektronix VM700A video analyzer, which was mentioned as a subject for a little happy repair session

One of the faults was a missing 976.5625 kHz interrupt clock, which is provided by a 74HC4020. After tracking down the fault by measuring I was surprised that this chip was the very only one in the unit which had corroded pins (see pictures).

I am wondering, what was first? the corroded pins or the faulty chip which got too warm and then got corroded pins?

Did someone see else see something similar?
Any suggestions?

After replacing the HC4020 the unit was working like a treat 

ingo.

[Kleinstein]

The chip in question might have had some corrosion from the beginning and thus poor solder joints. At least the solder quality looks relatively poor on that chip.

[duak]

The HC4020 leads look almost like the copper leadframe is showing thru the plating.  Does the color come off or is it the underlying metal?  If the plating is tin, perhaps it's "tin pest" where tin simply disintigrates.  I believe the date code is from 1989 and I don't recall any manufacturer having any particular problem with pin corrosion.  Perhaps heat caused this or at least accelerated it.  Does the chip have a VDD-VSS short?

Back in the 70's TI used silver (or at least something that tarnished like silver) on their ordinary DIPs.  A blackish deposit would show up after a few years and if IC sockets were used there'd be a good chance of intermittant connections.  One system spent a couple of years in Edmonton, Alberta where there are a number of oil refineries emitting sulfur dioxide.  The TI pins were totally black, as if I had used a marking pen on them.  On a lark, I tried some Pepsi - the phosphoric acid took the tarnish off but then I had a problem with the sugar.

I was looking at some chips that I were in some conductive foam that had disintigrated.  For the most part the pins were not corroded, but there was some black glop that's hell to remove.

Cheers,

[EEngo]

the faulty chip shows no current flow left between VSS and VDD, even at 13V and I can remove the corrosion by scraping it off, so it seems to be the case that the chip failed, got hot and the corrosion was induced somehow.

damn, I should have tested that at first 

thanks for your help!

[texaspyro]

I got in a couple scrap of ESI Video Bridge LRC meters ($25,000 when new).  All the pins on all the 40 pin Mostek Z80 chipset chips were totally rotten away.  Replaced them and all is well.

[EEngo]

so some chips were rotten away and some others were in perfect condition in the same unit? Or were all chips somehow corroded?

[texaspyro]

so some chips were rotten away and some others were in perfect condition in the same unit? Or were all chips somehow corroded?

Only the chips made by Mostek (Z80 and peripherals) were rotted.  I went to pull the CPU out of its socket and ...  surprise...

[schmitt trigger]

VM700A! I do have fond memories working with that instrument.
At the time it was released, it was heaven sent. It truly was.

To your question:
The corrosion on the pin itself may not be of consequence, but it may mean that the solder joints are compromised, which could have lead to the failure.

[EEngo]

oh yes, I kinda like those old instruments with full schematic service manuals. Apart from a few EPROMS (which are of course read and saved already), these units are fully repairable (and of course hackable too) 

I also like the opportunity to learn so much from these old designs 

In my case, the corrosion or solder points were not the actual problem, but I was surprised, that I could have seen this particular fault instead of tracking it down with the service manual. (Although that was definitly more fun!  )

[doktor pyta]

My quite long servicing experience is that in most cases (70%) the corroded IC indicates failure of the component ( I mean typical rust, not an oxidation of silver plating).
If I see something like this I replace the IC by default as it will fail sooner or later.

[wraper]

so some chips were rotten away and some others were in perfect condition in the same unit? Or were all chips somehow corroded?

Only the chips made by Mostek (Z80 and peripherals) were rotted.  I went to pull the CPU out of its socket and ...  surprise...

Means difference in materials used for pins/plating.

[texaspyro]

so some chips were rotten away and some others were in perfect condition in the same unit? Or were all chips somehow corroded?

Only the chips made by Mostek (Z80 and peripherals) were rotted.  I went to pull the CPU out of its socket and ...  surprise...

Means difference in materials used for pins/plating.

They both appeared to be tin which was the standard at the time.

[wraper]

so some chips were rotten away and some others were in perfect condition in the same unit? Or were all chips somehow corroded?

Only the chips made by Mostek (Z80 and peripherals) were rotted.  I went to pull the CPU out of its socket and ...  surprise...

Means difference in materials used for pins/plating.

They both appeared to be tin which was the standard at the time.

They both bay be covered with tin, however tin layer may be very different. It may be chemically plated or immersed in tin/lead alloy which will give way more thicker and more corrosion resistant layer layer. Pins themselves may be made from steel or from brass or bronze as well.

[Cyberdragon]

I really hate it when the corrosion creeps down into those stupid tin traces and makes them all bubbly. If the chip is bad you have to jumper the pads since they would disintegrate on removal. Never had a chip in this condition go bad yet, but I won't be happy when it does.

EDIT: I've reflowed joints in this condition just fine though, although the pads will lift up a bit on the pins.