Hi Mark,
Excellent work.
And probably more understandable for the crowd then what i manage to
explain.
My guess is number 5, and i think i can make that plausible.
I have been finding and replacing faulty microswitches in an industrial
environment, and the cause of the failure was usually water ingress in
the switches. What happens is that the water, which only has to be one
drop, boils between the contacts in a flash. The metal vaporizes and
forms a conductor on the insulation. The steam breaks the housing, with
could then hang the contacts up the mounting. Very much like the
microswith David describes. It sounds like you would need a lot of power
for that, but don't forget, these are microswitches, it's all very light
and feeble. An open contact could be 0.1 of a mil.
I think what happened is that the controller run the prop fine. The fine
limit microswitch(es) opened, exploded and shorted to ground. The
controller reversed to increase pitch and the breaker popped.
You don't mention Ivor's narrow escape with a broken motor, which looks
a lot like mine, with the difference that he had speed and altitude. The
investigator told me that the motor was loose in the spinner but he
blames that to the impact.
I still don't understand your willingsness to accept a 5 A diode, which
has to endure regularly 15 A. It could work in a toy car, but is not
acceptable in an airplane. You a talking about pulse load, In my book a
pulse is microseconds, a reversing motor will take considerately more
time. Does somebody have the typenumbers of the old and newer
microswitches? From memory the AC to DC max current is about 5 to one,
which would ask for 75 AC switches. I bet they are 6 A :-)
Regards,
Jos
28.6.2011 2:09, Mark Burton kirjoitti:
> --> Europa-List message posted by: "Mark Burton"<markb@ordern.com>
>
> OK, let's put our thinking caps on and ponder switches and diodes.
>
> We're only concerned here with the operation of the fine pitch limit switch
> and
its associated diode (incidentally, the SR3000 has two fine limit switches
and diodes in series so failures 1-4 described below have to occur to both
switches/diodes,
likely?)
>
> When the pitch is being reduced, the switch is carrying the current and the
> diode
is reverse biased. The current level is not high and well within the breaking
capability of the switch so it should open as expected and the pitch freezes
on the fine limit (+ some overrun).
>
> Assuming that the switch has operated as it should, the pitch will now be
> fully
fine. When the controller next commands the pitch to go coarse, the switch
is already open so the diode will have to carry the current. That current can
be large as it's basically the pitch motor stall current (initially around
25-30A,
falling quickly as the motor spins up). Now the diode is only rated for 5A
continuous current but it has a good pulse capability (check the data sheet)
so it probably won't fail (we will talk about what happens if the diode fails
next) - assuming the diode did not fail and the pitch motor is actually moving,
the switch will close shortly afterwards. Exactly how much current will flow
through the switch is difficult to predict because we don't know how much the
motor has spun up so it could be quite a lot - mind you, there's only around
1V across the switch (the diode drop) so I should think the switch could hack
it OK (perhaps with a reduced life).
>
> Now, let's consider some failure modes:
>
> 1 - the diode fails open circuit - in this case, when the switch opens as the
pitch goes fully fine the pitch will freeze at that point (+ some overrun) and
when the controller tries to coarsen the pitch nothing will happen because the
switch is open and the diode is open too. Therefore, the end result will be
the prop will be stuck in fully fine. But, this would not cause the C/B to pop.
>
> 2 - the diode fails short circuit - in this case when the switch opens as the
prop reaches fully fine, the pitch motor keeps going and will drive the pitch
all the way to reverse position (and the reverse switch would operate). If the
controller subsequently tried to coarsen the pitch, it would succeed unless
the diode decided the current was too much and died in which case the pitch
would
freeze. Now as it has been reported that the pitch of the prop was not in
the reverse position this probably didn't happen. Again, the C/B would not blow.
>
> 3 - the switch fails open circuit - the pitch will not be able to reduce, only
coarsen. C/B still intact.
>
> 4 - the switch fails short circuit - similar to 2 above, pitch could go to
> reverse.
C/B still a happy bunny.
>
> 5 - there is a short circuit from one of the switches/diodes/wires to ground
- now this is interesting: it could make the pitch go to the reverse position
and then when the controller tried to move the pitch the other way it would blow
the C/B. But as it's already been stated that the pitch was near the flyable
region that makes this scenario unlikely but not impossible.
>
> So, from what we know about the accident, the only one of these failures that
could have occurred is a short from one of the pitch motor drive wires to ground
(either the wire itself or a switch/diode/motor failure). Is that likely,
I don't know, but I reckon you could poor a bucket of water into a microswitch
and it's not going to conduct substantial amounts of current to ground (sea
water
would conduct better).
>
> So, Jos, that's why I don't have much to say about the quality of the diodes
and switches because apart from them physically flying apart and causing a short
circuit to ground, I can't see how they could be responsible for the reported
chain of events.
>
> Regards,
>
> Mark
>
>
> Read this topic online here:
>
> http://forums.matronics.com/viewtopic.php?p=344300#344300
>
>
|