X-Virus-Scanned: clean according to Sophos on Logan.com Return-Path: Sender: To: lml@lancaironline.net Date: Tue, 18 Dec 2007 17:10:11 -0500 Message-ID: X-Original-Return-Path: Received: from elasmtp-dupuy.atl.sa.earthlink.net ([209.86.89.62] verified) by logan.com (CommuniGate Pro SMTP 5.2c1) with ESMTP id 2602937 for lml@lancaironline.net; Mon, 17 Dec 2007 21:58:04 -0500 Received-SPF: none receiver=logan.com; client-ip=209.86.89.62; envelope-from=colyncase@earthlink.net DomainKey-Signature: a=rsa-sha1; q=dns; c=nofws; s=dk20050327; d=earthlink.net; b=QJWnNSn0OjbVjMo+uCHsN7ECe47t+Zps47ZAlpWtZ79cT1RIp5DA4aGVFflULCNX; h=Received:Message-ID:From:To:References:Subject:Date:MIME-Version:Content-Type:X-Priority:X-MSMail-Priority:X-Mailer:X-MimeOLE:X-ELNK-Trace:X-Originating-IP; Received: from [216.57.118.89] (helo=ccaselt2) by elasmtp-dupuy.atl.sa.earthlink.net with asmtp (Exim 4.34) id 1J4Se5-00054o-Do for lml@lancaironline.net; Mon, 17 Dec 2007 21:57:25 -0500 X-Original-Message-ID: <010801c84121$b85762c0$0402a8c0@nvidia.com> From: "Colyn Case on earthlink" X-Original-To: "Lancair Mailing List" References: Subject: Re: [LML] Re: Skoppe lancair 4 pt X-Original-Date: Mon, 17 Dec 2007 18:57:24 -0800 MIME-Version: 1.0 Content-Type: multipart/alternative; boundary="----=_NextPart_000_0105_01C840DE.A9372520" X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2900.3138 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.3198 X-ELNK-Trace: 63d5d3452847f8b1d6dd28457998182d7e972de0d01da940648cc14613dfb6de366ccf18a0e6157c350badd9bab72f9c350badd9bab72f9c350badd9bab72f9c X-Originating-IP: 216.57.118.89 This is a multi-part message in MIME format. ------=_NextPart_000_0105_01C840DE.A9372520 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Colyn, He has the Dynon EFIS as backup. Do you not consider a second = independent EFIS with it's own ADAHRS as sufficient backup ? I am considering the same idea. Quite Glass No not at all. data points: - mems based systems have only recently become "reliable". mems is = based on a tuning fork which is part of the chip. when you accelerate = the platform the tuning fork changes frequency. The frequency change = is mapped to an acceleration which is integrated into your current = attitude/speed. early on, failure to provide adequate mechanical = isolation from vibrations caused inappropriate response in the mems = system. I can believe Chelton has beat on this problem pretty hard. = What's it take for a Dynon to guarantee that their system works in your = airplane? If you get past the mems part, you still have a computer, a = lot of connections, another computer, and a display, all of which have = to work flawlessly or you have no information. As I said in a recent = post, my venerable garmin 530 display in my cessna failed recently after = 5 years of service. It was sudden and complete, and Garmin still hasn't = figured out what went wrong. incidentally my infamous rc allen electric = ai in my cessna outlasted the garmin - failure analysis of computer systems is a combinatorial problem. If = you sell enough of them (e.g. 10's of millions a year) you get some back = that failed because of really bizarre situations like multiple = simultaneous but unlikely events, coexistence with unexpected other = equipment, proximity to large voltage transients etc. none of these = ever happen in airplanes, right ;-) - generally large volume chip/computer companies do a massive amount of = testing which likely exceeds the entire run time of the entire = production run of a company like dynon - the bugs that slip through the massive testing are returned by (the = even more massive) customer base and the bug gets fixed. some companies = have a big farm of testing machines. when they get a failure from the = field they add that configuration to the farm and use it to test the = next chip. - failure modes of these kinds of systems are generally sudden and = total. - lightning happens - even if the efis itself is infallible, it is still dependent on your = power system. Now consider your old mechanical AI. It has a rotating mass and some needles. It's a very fallible thing but it's failure mode tends to be gradual. In my airplane I have an electric ai with an internal battery backup and = a vacuum ai which depends on the engine and an electric hsi and a vacuum dg and an electric turn and bank. so to be out of luck I have to lose one of the following combinations: - both ais, both dgs - both electric systems, the ai backup battery, and the vacuum pump or = engine - both electric instruments and the vac pump or engine computers fail, and one of the failure modes is common between your = dynon and your main displays, namely something gets zapped in your power = system and perhaps every box that is attached to it. ------=_NextPart_000_0105_01C840DE.A9372520 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable

Colyn,

  He has = the Dynon=20 EFIS as backup. Do you not consider a second independent EFIS with = it=92s own=20 ADAHRS as sufficient backup ?

  I am = considering=20 the same idea. Quite Glass

 

No not at=20 all.

data=20 points:

- mems based = systems=20 have only recently become "reliable".  mems is based on a tuning = fork which=20 is part of the chip.  when you accelerate the platform the tuning = fork=20 changes frequency.   The frequency change is mapped to an = acceleration=20 which is integrated into your current attitude/speed.   early = on,=20 failure to provide adequate mechanical isolation from vibrations caused=20 inappropriate response in the mems system.   I can believe = Chelton has=20 beat on this problem pretty hard.  What's it take for a Dynon to = guarantee=20 that their system works in your airplane?  If you get past the mems = part,=20 you still have a computer, a lot of connections, another computer, and a = display, all of which have to work flawlessly or you have no = information. =20 As I said in a recent post, my venerable garmin 530 display in my cessna = failed=20 recently after 5 years of service.  It was sudden and complete, and = Garmin=20 still hasn't figured out what went wrong.  incidentally my infamous = rc=20 allen electric ai in my cessna outlasted the garmin

- failure = analysis of=20 computer systems is a combinatorial problem.  If you sell enough of = them=20 (e.g. 10's of millions a year) you get some back that failed because of = really=20 bizarre situations like multiple simultaneous but unlikely events, = coexistence=20 with unexpected other equipment, proximity to large voltage transients=20 etc.   none of these ever happen in airplanes, right=20 ;-)

- generally = large=20 volume chip/computer companies do a massive amount of testing which = likely=20 exceeds the entire run time of the entire production run of a company = like=20 dynon

- the bugs = that slip=20 through the massive testing  are returned by (the even more = massive)=20 customer base and the bug gets fixed.  some companies have a big = farm of=20 testing machines.  when they get a failure from the field they add = that=20 configuration to the farm and use it to test the next = chip.

- failure = modes of=20 these kinds of systems are generally sudden and total.

- lightning=20 happens

- even if the = efis=20 itself is infallible, it is still dependent on your power=20 system.

 

 

Now consider = your old=20 mechanical AI.

It has a = rotating mass=20 and some needles.

It's a very = fallible=20 thing but it's failure mode tends to be gradual.

In my = airplane I have=20 an electric ai with an internal battery backup and a vacuum ai which = depends on=20 the engine

and an = electric hsi and=20 a vacuum dg

and an = electric turn=20 and bank.

 

so to be out = of luck I=20 have to lose one of the following combinations:

- both ais, = both=20 dgs

- both = electric=20 systems, the ai backup battery, and the vacuum pump or = engine

- both = electric=20 instruments and the vac pump or engine

 

 

computers = fail, and one=20 of the failure modes is common between your dynon and your main = displays, namely=20 something gets zapped in your power system and perhaps every box that is = attached to it.

 

 

 

 

------=_NextPart_000_0105_01C840DE.A9372520--