December 28, 2005

Chilling

As John mentioned elsewhere, a Cessna 172 recently went down a little south of here, killing all four people onboard in rugged terrain that many of us Bay Area pilots fly over regularly. The NTSB brief makes chilling reading ("The non-instrument rated private pilot and three passengers were fatally injured. The airplane was destroyed. Night instrument meteorological conditions prevailed [...]"); I remember hearing about it on the local news the next day and wondering how anyone would want to have been flying that evening (a dark and stormy night, so to speak), let alone someone intending a VFR cross-country with passengers...

December 16, 2005

Flying Glass (Part 1)

Garmin G1000 Panel shotA quick early morning call from John as I'm getting coffee at Javarama, and an hour or two later I'm face to face with the Garmin G1000 in California Airways' glass cockpit C172, N14008, a few weeks earlier than I'd planned. I'm not sure what to expect -- I did my homework a few weeks ago, but with the short notice I can barely remember the various primary function display (PFD) features, let alone some of the crucial system and module failure modes. I tell John we should aim low -- the Cal Airways sign-off requirements for this plane are two hours of flying the G1000 with an instructor plus a written checkout for straight VFR use, and a full instrument proficiency checkout on top of that for IFR use -- so while I know we haven't got time to do the IFR thing today, maybe we can get most of the VFR flying done. I can fill out the four page checkout sheet (with all the usual questions about failure modes, etc., some of it common to the sheet I filled in for the SP's a few months ago) over the next week at my leisure. We'll do the IFR checkout early January, probably.

We go through the checkout sheet in the club, and I feel embarassed that I can't answer even some of the "obvious" questions about the unit and the plane -- luckily, though, these were mostly the sort of questions whose answers you typically wouldn't memorise anyway. But at least the questions about things like how to kill the autopilot, or display reversion, or soft field / short field (etc.) procedures have been burned into my brain one way or another enough that I could answer them (mostly). John shows mercy and lets me take the sheet home for filling out later -- he'll have to sign that off separately.

It's a marginal VFR day at best, so I file IFR to VFR on top, something I realise I've never actually done for myself before (which surprises me). DUATS says it's basically clear on top of a broken layer that doesn't extend much above 2,000' locally, so we'll head off towards Mt Diablo and make things up as we go. I do the pre-flight, which is essentially identical to the other SP's, but when we get in and prepare to start the thing, the information overload begins creeping up on me, and I slowly become more aware that this is a system I'm playing with, not just a bunch of disparate instruments. In particular, things like backup battery power, fall-back instrumentation, etc., have to be armed and tested in a certain order before and after engine startup. This is the classic time for an official checklist, so on John's prompting, out it comes. It's a lot more complex than the SP's, let alone something like 4AC, where startup is little more than priming a bit, yelling "clear", and turning the key.... I feel curiously like, erm, a real pilot; I start feeling maybe I should wear a uniform....

I get it all under control eventually, and we taxi off to the runup area with John asking me questions about how I'd do this or that, and giving me a few hints, etc., and somewhere along the line we pick up our clearance, the standard "cleared to San Jose VOR (SJC), runway heading until 400', then left turn to 160, radar vectors to San Jose, report reaching VFR conditions" IFR-to-VFR on top clearance. Nothing too difficult. And programming the G1000 for this -- and later escapades -- turns out to be nothing too difficult, either -- it's basically the same as programming the GPS 530, with all of the pluses and minuses that implies (I share John's frustration at the "enter" key thing). We spend the time waiting for release going over details like how to get the HSI to superimpose bearing to a station, how to set up the cramped little situation window (whatever it's called) next to the HSI, etc., and the different multi-function display (MFD) modes such as traffic, nexrad, XM, etc. This stuff's magic. And the MFD engine data displays are a model of clarity, too, laid out logically and very effectively (for me, anyway).

After the usual long wait for release we're finally cleared for takeoff on 28L. The big test here for me is how well I'll cope with the vertical tape airspeed and altitude displays -- everything else seems at least somewhat familiar, especially the HSI and huge AI (finally, an AI in a C172 that's big enough to use easily and accurately, especially for pitch information!). So we start down the runway, and ... disappointingly for some of you, I'm sure, for the rest of the flight the airspeed and altitude displays (especially the altimeter) turn out to be little or no problem. Yes, I make a few mistakes reading them a few times here and there, and yes, I miss the quick-glance mode of readout available from the analog versions, but really it's just not that hard to adapt. Half way through the flight I catch myself thinking "I'm doing 108 KIAS instead of 110 -- let's just push the throttle forward 30 RPM..." (or whatever) and think, hmmm, this three decimal place readout thing might just be a little spurious... (but it's fun).

So we cancel IFR at about 2,000 feet and spend the rest of the time doing maneuvers, then two times around the ILS RWY 27R into Oakland, both times under the cone of stupidity. Nothing much to report about any of this, except the information overload caused me to miss a few radio calls, to screw up a simple VOR interception, and to fly really pretty badly during some pure VFR segments. Urgh. And it was a good workout for getting on top of the KAP 140 autopilot as well -- most of the time we left it coupled -- and John was full of the usual hints and help. I've tended to use it more in vertical speed mode, dialing in a positive or negative VS to get to the altitude I want, monitoring it myself and getting it to level off by hitting the altitude set button at the right time; John's getting me more comfortable with using the dialed altitude method. Unfortunately, I keep making mistakes with this -- it seems to be way too easy to reset the VS while dialing in the altitude -- which keeps reminding me why I started using the VS method in the first place. I'll get it right every time... eventually.

The first time around the ILS at Oakland we can't get the autopilot to couple to the glideslope, so the next time around we concentrate on getting it right by intercepting it from well below; this time it works. It didn't help that the NorCal controller vectored us straight through the localiser (I saw it about to happen and called him, but he responded -- apologetically -- too late); this is a first for me -- NorCal usually vectors you nicely onto the localiser... several thousand feet too high. One embarassment for me: the first time around the ILS, somewhere outside FITKI (the notional GS intercept point), John casually asks me where the glideslope indicator is. I realise in horror that it's not where I thought it'd be -- on the HSI. He lets me sweat a little then points out the tiny moving green blob next to the altimeter that's obviously (ha!) the GS indicator. I guess it makes sense to have it there, but it's way too small to be easily visible while you're monitoring the HSI on final, and I'm unimpressed by that particular design decision. I'd prefer to have the choice of bars against the HSI as well.

After the second time around the ILS we depart VFR through the murk back to Hayward; I embarass myself further during this leg by some really bad flying -- again, it's easy to blame information overload, but this time it was also caused by having trouble finding the airport through the mist (it's still MVFR) and by my obsessing about the engine instruments. I'll learn....

So now I'm basically checked out and ready to fly VFR with the G1000 (I need to get John to sign off on the paperwork early next week to make it totally legal); I'll be making another flight with John sometime over the next few weeks to get the rest of the IFR checkoff done. This is a lot of fun, I have to admit.

* * *

All in all, a good lesson. Not so much because of the G1000 itself, but because the rough bits of the flight once again rammed home to me just how much you need to guard against the temptation to put aside the actual basics -- keeping the airspeed up, ensuring you're listening to the radio -- to get some minor detail or other figured out or completely correct just because the bloody thing's there in front of you on the PFD or whatever (and thanks to John for his patience with this).

My overall impressions of the G1000 so far? This is a very nice way to fly, even in a 172. Especially in a 172, if you're serious about flying single-pilot IFR. Basic VFR flying with it is -- with caveats about getting used to vertical tape displays and a few other fiddly bits that seem incomprehensibly-designed or laid out -- as easy as with "normal" 172 instruments, and for the limited amount of IFR flying I did under the hood, the displays and the way everything's nicely integrated becomes very addictive after a while. Yes, as always, your mileage may vary depending on how much you like integrated computer displays (I'm a nerd, so they seem utterly natural to me...), but, if nothing else, the various redundancies and fallbacks -- including the steam gauges of last resort under the main displays -- strike me as just an order of magnitude more reliable (in every way) than something like the conventional instruments in, say, 4AC. If you're looking for a detailed critique of flying with a G1000, or second thoughts on the wisdom of digital flying in a GA context, you're reading the wrong guy, unfortunately. I liked it immensely, I'm enthusiastic -- the only downsides have to be the extra costs and the training.

But it's obviously not for everyone -- and I couldn't care less about the G1000 when I'm doing aerobatics or just pottering about the Bay on the Bay Tour -- and I couldn't afford to buy an old M-series 172 with steam gauges, so if I weren't renting the damn thing for not a lot more than a conventionally-instrumented 172SP, I'd probably have a very different view of the cost vs. safety / convenience tradeoffs here....

December 10, 2005

Reading Glass

Well, reading the G1000 manuals wasn't as bad as I'd thought -- what's a few hundred badly-written tech pages to someone who's motivated? -- but, as usual, the proof's in the pudding. I haven't flown for weeks (weather, and a new project at work, mostly), and I sometimes wonder whether I'll even remember how to take off, let alone what this readout or that instrument is telling me in IMC.... I've been in touch with John a couple of times about getting checked out (or even just flying along with him as safety pilot), but there's always some excuse on my part. So it goes. It'll probably be January; the good news (on the grapevine) is that Cal Airways is getting more glass cockpit 172's. Cool!

Anyway, as an old hardware and software engineering nerd, I'm impressed and intrigued by the G1000 system architecture and implementation. For example, it warms my heart that it uses ethernet to join the components together -- the sort of thing I predicted decades ago when ethernet was still done with temperamental inflexible thick co-ax cable and vampire taps (don't ask...), and when 1Mbps was pretty optimistic. I just love this sort of thing -- and the second-guessing and mental reverse engineering I do when reading about hardware and systems architecture like this keeps me occupied when I should be doing real work.

Something I don't love is the likely difficulties I'll have with the vertical tape displays for airspeed and altitude. Almost everything else looks straightforward about the G1000 -- the GPS seems to be a glorified GPS 530, and the rest of it is "obvious", at least while sitting here on the ground -- but, as both John and David Megginson have pointed out (see e.g. David's article here), the transition to this style of presentation can be rough.

As my over-used catchphrase has it, we shall see.