|
|
Holy crap. I am so impressed Dale. What you're doing is looking at the
next part of our failure analysis.....What are the systemic causes? What
about the way we build planes leads to this? This is where we start
saving some lives. What is it about this connector that contributed? Ian gave us some
insights. He describes how much effort the auto manufactures put into
this stuff. It's historically been a real problem for them. I've heard
that Volvo went so far as to remove every intermediate connector. One of
their biggest causes for failure I'm told.
There are all sorts of design characteristics that reduce odds of good
connection. The clearance between the pins and the housing. The slop in
housing lets pin move off center, which forces pin out of connector
instead of into mating piece. I just witnessed complete engine shutdown 3
weeks ago due to this. Also wiped out the guys radio. See if this sounds
familiar. He bought used experimental, went thru it, checked stuff. In
the process he disconnected firewall connector and then re-plugged it in.
Two of the 10 pins only partially engaged. Plug was fully together, just
pins not seated. He was taxi testing when bam, lost all power. I was able
to cause power to go on and off by just pushing and pulling on one wire.
Looked at the plug and you could see two pins were pushed out. They were
only lightly touching the female connection.
So ONE systemic failure is that we are using low volume NON Toyota plugs.
The plug design is not optimized. So one solution is to write down MIRROR
and FLASHLIGHT on our action list. Then look at the backside of all our
plugins to verify the pins have not backed out of position. If we do
unplug one, poke around those pins and see if they wiggle. That indicates
you have a kmart plug connector. Better put a tag on that one that says
"Look for extruded pins after plugging in". Or something more clear, or
more permanent.
I've known of this type failure for years. When I built my plane I would
plug it in, then look at back side of male and female plug. Verify pins
fully engaged. Then on to next one. I did find pins not fully seated. I
also pulled on any screw connection to verify I had valid connection. I
found quite a few that looked good, but was open circuit.
Nice job Dale!
-al wick
On Mon, 6 Jun 2005 11:40:25 -0400 Dale Rogers <dale.r@cox.net> writes:
John,
I suspect what Al W. was getting at, is that every connector which is put together in the same way as the troublesome one is now *suspect*. If you _are_ using one on a flight critical system, you'll want to inspect it and verify its integrity under some reasonable level of motion.
As a semi-aside, this sort of thing is why I really don't care for soldered cable connections. I work in the computer industry as a field engineer, and soldered connections are rare except inside a chassis where the wires are well-
secured against motion. Soldering a stranded wire connection focuses all the force of motion onto the point at which the solder no longer binds the strands together. There is a reason why the connector industry can get more money for crimp connectors than for the solder type. They're worth the extra cost.
My $.002,
Dale R.
> From: "John Slade" <sladerj@bellsouth.net>
> Date: 2005/06/06 Mon AM 10:49:51 EDT
> To: "Rotary motors in aircraft" <flyrotary@lancaironline.net>
> Subject: [FlyRotary] Re: EC2 problems - solved / rotary risks
> > Thanks, Al.
> After a slightly bad start, this is starting to get very informative and
> interesting. Please keep it coming.
> > In the case of the wire/connector, I'm not really worried about it because
> it'd not an engine out thing, just a loss of communication with the engine
> monitor.
> > I like the "best asset is that I know I'm an idiot" theory.
> It works well for me because Rusty keeps reinforcing it. :)
> John
> > -----Original Message-----
> From: Rotary motors in aircraft [mailto:flyrotary@lancaironline.net]On
> Behalf Of al p wick
> Sent: Monday, June 06, 2005 9:58 AM
> To: Rotary motors in aircraft
> Subject: [FlyRotary] Re: EC2 problems - solved / rotary risks
> > > When we are pursuing a problem like Johns, we are eager to find the cause.
> It's a great relief when we do. We say "Eureka!". We did it! We did it! This
> sense of relief is a root cause for failure. We are so eager to get the
> problem off our back, that we don't take the next step....
> Yes, I know you firmly believe this connector was the problem. But if you
> can force yourself to pretend it WASN'T, then you can do this:
> Is the cause logical? Like is that really the wire that causes that
> effect? If I remove that wire, does it have the same effect? What if I have
> two things causing the same thing? By pretending that really wasn't the
> cause, then you will do some more testing, looking around. Looking for
> similar connector issues, stuff like that.
> > Now I have to admit, this really does sound like he found the cause. But
> I've seen this scenario so often. So you use the disciplines I suggest to
> reduce your risk. Logical cause? Can I make it recur?
> Repeat after me: "Al Wick is an idiot". "Al will jump to conclusions". If
> you believe that, then you start finding ways to prove your theory with
> facts instead of just accepting your first conclusion. My best asset is that
> I know I'm an idiot.
> > Yeah, yeah, I know, you guys already knew I was.
> > We had a perfect example of this on Cozy list couple weeks ago. Subaru
> engine slipped 2 teeth on timing belt. Would no longer start. Keith talked
> to expert and the guy said:"You know, the engine normally is never rotated
> backwards. But you've been pushing your new prop backwards recently
> (installing new prop). I think you relaxed the belt tensioner when going
> backwards and caused it to skip tooth." So Keith said" Yes, all of that's
> true. That has to be it."
> > But then one of the guys looked into it, guess what? The direction the
> belt slipped is the opposite of that theory. That could not have caused it.
> The lesson? Prove all aspects of the theory are logical. Prove that all
> the various facts support the theory. Find a way to convert your theory to
> facts!
> > Oh, by the way, if you look at my analysis of my engine risks....you will
> notice that timing belt is the highest risk item on this engine. So we have
> exposed another root cause for his problem. He didn't focus on the leading
> cause for all engine failures. When we reviewed some facts he had, we found
> conclusive evidence he had loose belt from day one! It was installed wrong.
> > Regarding CAS risk. It's not just crank angle sensor that is the risk
> item. Going to redundancy with the CAS will dramatically reduce risk of all
> ECM causes. Like this connector risk. I'm not always proponent of
> redundancy, but with my limited info on this item, I SUSPECT it's
> significant, positive step.
> > -al wick
> Artificial intelligence in cockpit, Cozy IV powered by stock Subaru 2.5
> N9032U 200+ hours on engine/airframe from Portland, Oregon
> Prop construct, Subaru install, Risk assessment, Glass panel design info:
> http://www.maddyhome.com/canardpages/pages/alwick/index.html
> > > > > On Sun, 5 Jun 2005 21:44:00 -0400 "Tracy Crook" <lors01@msn.com> writes:
> Ahh.. Music to my ears John : )
> > And this brings up the subject of risk (rotary & otherwise) that Al W.
> (and every other builder I know) is concerned with. I agree with Al W. that
> getting to the major causes of failures is a (hell, THE) key issue. That is
> why I have not spent much time on the crank angle sensor single point
> failure question. I have never seen or heard of a confirmed Mazda 13B CAS
> failure. Can it happen? Of course. I am in the process of developing a
> dual CAS for the Renesis CAS but it is not a 'front burner' project.
> > I'm reading between the lines of Al's posts but it seems that he is
> emphasizing the importance of leaving the engine as un-touched as possible.
> I once wrote an article for Light Plane World (EAA's ultralight magazine
> back in the late 80's) and advocated the same thing after noting that many
> Rotax failures occurred soon after the owner opened up the engine for
> maintenance. Decarboning the piston ring grooves was important but many
> builders were causing more problems than they fixed when they went inside so
> I recommended some products and procedures that would do the job without
> opening the engine.
> > That was the basic gist anyway but I eventually decided this was not a
> reasonable approach for builders who planned on installing an alternative
> engine in 200 mph category airplanes. There were simply far too many areas
> where things could go wrong in this process. The root cause of the problems
> had to be identified. One of the names I gave to the cause is a term I
> recently used on this list - Shopcraft (or lack of). This referred to the
> ability to identify the quality or suitability of virtually everything that
> goes into the plane. Yes, I know this is a generality of the highest order
> but if we are to get to the root cause of failures in the field of
> alternative aircraft engines, this level of abstraction is required.
> > It has been suggested that a collection of 'best practices' might be a
> solution. This may help but it is not a solution. There is an unlimited
> number of potential problem areas so a list of them could never be compiled.
> So, how do you learn to recognize what is or is not a 'good thing'? I'm
> getting so frustrated just trying to describe the problem that there may not
> be a solution, at least not one that can be spelled out in something like an
> email message. Damn, now I can't even criticize Al W. for not spelling it
> out.
> > The best I can do for now is to emphasize two things. Pay attention to
> every detail and admit to yourself when you don't have the ability to
> execute something well. Another version of these rules was given to me
> long ago:
> > 1. Rules are for those who are not smart enough to make up their own.
> (Author unknown)
> 2. A man's got to know his own limitations. (Dirty Harry)
> 3. Always follow BOTH rules 1 & 2.
> > Small details like the problem of soldering thermocouple wire to a
> connector that Al Gietzen mentioned can be critically important. He was
> able to recognize the problem (he made a lousy solder joint) and devise a
> solution (acid flux) even though it violated one of the cardinal rules of
> electrical wiring. He recognized that too and took the steps necessary to
> achieve satisfactory results (knowing when to make up his own rules).
> > Out of time, I'll stop blathering now.
> > Tracy
> > Subject: [FlyRotary] EC2 problems - solved
> > > Tracy and others.
> Following more than 12 months of battling with EC2 issues I'm pretty
> sure it's Eureka day!
> After rewiring and testing for almost 4 weeks I plugged the EC2 in
> last night, and got exactly the same symptoms as before. NOP flashing
> indicating no communication. I took the EC2 to Buly's plane and tried it in
> his installation. Same NOP, so I was thinking I'd fried it again. Before
> sending it back yet again I decided to install it my plane one more time and
> see if there was a spark.
> > To my amazement it worked. No NOP, and I could bring up the EC2 data.
> The only thing that changed overnight was that I moved the cable to unplug
> it. I climbed in the back and found that I could make the NOP flash, or stop
> flashing, by moving the cable. I haven't taken the connector apart yet, but
> I'm expecting to find a broken wire inside the insulation, probably near a
> solder joint at the pin. Whenever I bent the connector outward for testing
> it made contact. When I bent it back to plug it in, contact was lost.
> > Bingo!
> John
> > Just guessing, but maybe the new EC2 can't communicate with a
> pre-autotune EM2 like Buly's. ???
> > >> Homepage: http://www.flyrotary.com/
>> Archive: http://lancaironline.net/lists/flyrotary/List.html
-al wick
Artificial intelligence in cockpit, Cozy IV powered by stock Subaru 2.5
N9032U 200+ hours on engine/airframe from Portland, Oregon
Prop construct, Subaru install, Risk assessment, Glass panel design info:
http://www.maddyhome.com/canardpages/pages/alwick/index.html
|
|