MLCCs encapsulated in glass?

[RoGeorge]

Found them in the scrap box.  No idea where from are they.  The size is similar with that of a glass diode, and what is inside the glass looks like a MLCC capacitor.  The marking is 104Z 50V 8220.  I guess they are 100nF/50V MLCC in glass case.

Never seen such glass encapsulated capacitors.  Searched online and couldn't find any.

Anybody know what are these for, and why a glass case for a 0.1uF/50V, anyway?

[Gyro]

I've got some of those too, stripped from prototype boards - surprisingly, about 30+ years ago, certainly before we were using smd on our boards. There were/are epoxy encapsulated ones in the same form factor.

I've no idea if the glass encapsulation is intended to improve reliability. I suppose they would be more tolerant of physical stress than epoxy ones (auto lead bending, auto insertion machines?) and moisture-proof. Mine are 100nF decouplers too.

[Kleinstein]

I have seen these odd looking caps used for decoupling in old computer boards (I remember a ZX81 RAM expansion board). So these are not precision caps, more like small size low ESR / ESL caps.  These could be class 3 ceramic caps, using some special surface effects.

[2N3055]

An old datasheet.

Kleinstein is right. They were good for automatic insertion, small for the time, quite robust, MIL grade.
Used mostly as decoupling caps..
I also remember if from Sinclair ZX81, they were used in modems, mainframe...

[RoGeorge]

Thank you all, pretty clear now.

Was hoping for something high precision, but they are just nice looking vintage, probably from the week 20 of 1982.

Seems this glass encapsulated MLCCs can still be bought as NOS.  Searching for Unitrode I found some on ebay, https://www.ebay.co.uk/itm/100x-Unitrode-axial-100nf-50v-ceramic-capacitor-0-1uf-C104-202-104-0-1uf-/391907472239 and from the last picture on ebay, the label says tolerance is +80% -20%.

[btb]

I also came across some of these, on the badly damaged keyboard of a (rare?) serial terminal I'm trying to restore. (Telcon Ambassador, built in 1980 - thread)
I was mighty confused until I decided because of their placement they couldn't really be anything other than decoupling caps. Glad to find this info to confirm it.

Here's a pic showing a handful of them in place, along with a glimpse of the kind of damage I'm dealing with

[Tomorokoshi]

Good luck on that restoration.

Note to self: Don't drink battery acid over the computer keyboard.

[David Hess]

You can find those glass DO-7 packaged MLC capacitors in equipment built throughout the 1970s and 1980s.  They were gradually replaced with epoxy packaged axial MLC capacitors which are not nearly as tough.

There is nothing special about them, although their reliability should be higher than normal because of their hermetic packaging.  I suspect they were taking advantage of old packaging equipment for glass DO-7 diodes.

[Neomys Sapiens]

Interesting. They are clearly discernible as MLCCs, yet they share the form factor with the GLASS CAPACITORS which came first to my mind.

[David Hess]

Below are some examples that I pulled a few years ago from the main boards of some trashed Tektronix 22xx oscilloscopes.  You can see that they were marked for voltage, capacitance, maker (WC?), date, and came in two lengths.  These have 1984 date codes.  A full range of values up to at least 0.1 microfarads were available.

The big package in the middle is a DO-7 1N270 made by BKC.  The smaller package at the bottom is a DO-35 1N4148.

[btb]

Can't believe I never noticed this before, they're all over the TI-99/4A

[T3sl4co1l]

I haven't seen that board in ages...

I think I still have one floating around in the basement, too.

I have however looked at a somewhat contemporary module more recently, it's documented here:
https://seventransistorlabs.com/DC-DC_RE/index.html
which appears to be somewhat better designed than that one (given its dual comparator and 723!)... maybe.

Tim