GaAs Infrared emitter failure

[N_Cook]

TO18 size of can but inerference fit? mounted in a brass mount with
lens, no type number seen.
Actively passing 60mA drops 1.3V and DVM-D test shows 1V one way, so
diode function , but no IR out as tested with IR zapper tester of IR
photodiode in series with optical LED , indicates on or flashing of any
working zappers ever tried with it.
So are they known to fail optically, perhaps overdriven, but not go o/c
or low ohms , still functioning as a diode though. ?
Kit specs say 910nm and no tab seen , but TO-18 can type , although
that could be machined off to clear the brass barrel , any machining not
seen as axial view only, due to the surrounding brass barrel. So
something like
SFH400 to 2, FPE500 to 530, OP223,224,231,232,233, or TIL23 to 25 but
none of those are specified as exactly 910nm , ranging 880 to 950nm.
Initially I will try heating the assembly to 120 deg C and blast of
freezer spray to TO18 package , to try breaking presumed assembly-heated
interference fit ,
to split apart. Other than finally drilling out or seriously heating if
laquer fit, any ideas to break apart without destroying any marking.
Hand scratched on the brass is number 3738 presumably relating to a test
of the lens combination or perhaps coding of date in 1978 , as 1979 kit.

[Jon Elson]

On Thu, 14 Mar 2019 11:04:55 +0000, N_Cook wrote:

> TO18 size of can but inerference fit? mounted in a brass mount with
> lens, no type number seen.
> Actively passing 60mA drops 1.3V and DVM-D test shows 1V one way, so

Is this a laser or just an LED? 60 mA sounds to me like it might be a
laser, and fairly powerful.

1979? Hmmm, probably NOT a laser at that date, and hitting such an old
LED with such current would be pretty hard on it, unless the pulses were
short and the duty cycle low.

Possibly heating the whole assembly might make the emitter slide out.
I doubt these would be interference fit, as that could easily damage the
emitter. Likely, a TO-18 package had a fairly wide dimensional tolerance.
So, I'm guessing they used some kind of glue to hold it in place.

Jon

[N_Cook]

No type number found, but at least the likely failure mechanism found,
manufacturing flaw waiting to happe, although all well butchered now.
Not a TO18 device fitted to a brass barrel. Heat and freezer spray made
no difference. So decided to dremmel 0.5mm grinding disc a slot in the
brass to relieve the pressure. Firstly it was gold plated copper , not
brass and no TO18 device, just a ring of something containing brown
epoxy and the pins , so looking like TO18 can.
The GaAs die .4mm or so square sitting on plated Cu "top hat"
heatsink/plinth, inner diameter of 10mm and outer diameter of 15mm.
Interference fit to that ledge, dislodged in trying to force open the
slot in the copper, machined Al barrel 13mm long. Machined to a cone
inside with pin-hole at end, taking an unsupported length of bare fibre
optic approx 8mm long, presumably was cemented to the die.
Then a lens fitted over the other end of the fibre over the outer end of
the internal cone.
So with resonant mechanical vibration from carriage at some point ,
thermal or g-shock stress , or just post fibre manufacture internal
fibre stress-relief over time, the fibre could break from the die and
end up 0.2mm off-axis , so IR in the cone recess and next to none in the
fibre,so to the outside world no optical output.
As no make/number perhaps an early run, just post prototype, before
someone twigged the built in self-destruct design flaw.

[Look165]

Maybe can be a UV LED !

[N_Cook]

On 15/03/2019 12:43, Look165 wrote:
> Maybe can be a UV LED !
>
>

Specifications in the user manual say IR 910nm and another
contemporaneous (1979) source says the same.