« Posts tagged instruments

Rockin the clock.

5 hours.

Last time I went flying, there was a very odd glitch.  On the first takeoff of the day, if I do a quick pattern, on base leg before final, I’ll lose my AHRS and compass.   This was intermittent, and I couldn’t reproduce it on the ground.   This week it got worse: If I landed and taxied back to the hangar and shut down the engine, the instruments would come back.  With the engine running, the AHRS and compass are inop.  This is WEIRD.   Strange things happen when voltage dips below 10v, but my backup battery and system voltage both were above 12v.   A quick check of the RCA cables revealed one loose.

Bingo.  RCA cables as aircraft duty connectors are a the worst idea in a collection of bad ideas, but there’s not a lot I can do about it.    Added item to maintenance checklist.   No weirdness from the radio either.

But the futzing around cost me an hour of valuable good weather.  I had planned to get 4 hours of flight time today; I got 2.   That’s only 14 hours left on the Phase 1 countdown clock.   But I need all this crap to stop happening so I can do some real tests.


Even more legal, and some argh

10 hours.

Friday I got the static/transponder test done and that’s now in the logbook.   That took a bit, because I had to chase down some leaks.   And, somehow, I managed to snap off the hose barb that connects to the EFIS’s AOA pressure sensor.   That really, really sucks.   Now, once again, I have to dismount the EFIS and take it to MGL to have the AOA sensor replaced.  Can’t fix the barb.  The good news is that the AOA isn’t really necessary for first flight.   Also, my combination AOA/Pitot doesn’t lend itself well to the usual testing method of slipping a hose over the end of the pitot and hooking it up to the test machine.   But we worked around it by connecting the test line to the fitting where it connects to the pitot tube, so I’m going to assume it works.   We’ll find out.

Static tests eventually worked out OK, and we were able to calibrate using the EFIS’s setup menus.    All good!  Signed off, sticker in the airframe logbook.

Today was about picking up the pieces and extracting the EFIS backup battery for replacement.   I think it got run down too far too often and now it won’t take a charge and tops out at 6v.   To do that, I had to relocate the IO Extender module to someplace that would allow me to get to the screw securing the backup battery in its bracket.   When I had to move the IOX before to get it out of the way of the canopy, I’d put it on the end of a piece of angle behind the EFIS.  This solution was kind of half-assed, and I never actually felt good about it, so this is actually an improvement: I put it on top of the throttle quadrant bracket, up out of the way of everything.   I’ll have to dismount the transponder to make changes to the wiring, but that’s OK, it shouldn’t be necessary to change the wiring much.  It’s much more elegant, and gives me plenty of room behind the EFIS to do what I need to do, plus it’s more secure.

I did have to extend the power and ground wires for the IOX, but I was able to use the trim position wires as-is.  It also enabled a much neater wiring bundle instead of the slightly messy arrangement I had previously.

Magnetos timed (again) and RPM sensor fixed.

6 hours.

One of the day’s tasks was to replace the magneto gasket we tore with a new one.   This necessitated removing the right magneto, the one that was cranky about timing.   This is also the mag that has the hinky contact spring on one of the harness wires.  I figured what the hell, I’ll make sure both are timed, so I’ll pull both, put new gaskets on both, and set the timing again.   Big mistake.   Timing mags with two people sucks, as you might have gathered from an earlier post.   Doing it on your own sucks too, and then some.   Yesterday it was hot and windy at OXR, so the hangar was warm and the doors were rattling and making freakish noises every couple of minutes.

There are various sources of knowledge as to how to time mags, but the general idea is that you turn the engine over until the crankshaft is at 25 degrees before top dead center.   Then you zero the mag position by sticking a pin in the timing hole in the back of the magneto where it will sink into another timing hole on the gear inside, thus locking the thing into the position where the points are open.  Then you stick the mags back on your timed engine.

A number of factors are at work here.   There’s the impulse couplers, there’s the slop in the mags even with the timing pin inserted, there’s gear lash in the engine itself, and there’s the ambiguity of where 25 degrees BTDC actually is.   Near as I could figure out, there’s a little lump in the rotation of the mag gearshaft where the points open.   This is no great mystery.   A cam inside opens and closes the points to make a spark.  That fires the spark plugs.  This is no different from your common Briggs and Stratton-powered suburban lawnmower.

But the magnetos are buried amongst wiring, hoses, and engine mount tubes, and getting them back on can be extremely frustrating.  The mag gasket is a little wider than the magneto housing, so it’s really easy to dislodge it.   The engine gears of the mags are also wet with engine oil, so they’re slippery.   It’s also of primary, critical importance that nothing falls down the open mag hole into the accessory case.   If that happens, game over.  If that happens, you better hope whatever fell, like a nut or a washer, falls all the way through into the oil sump where it gets stuck to a magnetic plug — if you have one.  The obstacles in the way make it easier, but not by much, to insert the timing pin into the housing with the mags more or less in place, but supporting them, spinning the gear, and inserting the pin is one of the more frustrating solo tasks there are.  The trick is to get both mags installed so they’re within just a few degrees of each other, because there’s only about 70 degrees of twist available to bring them into line.    If they’re too far off, they’ll never sync up.

The next part is relatively easy, assuming the first part goes OK.   This is where the buzz box comes in. This thing has two lights on it, and you connect the leads to the P-lead connections on the mags.  When the points are closed, the lights are on.  When the points are open, the lights are off.   Or vice versa.   Can’t remember.   The point is, the status of those lights is supposed to change when the engine turns through the specified timing mark on the flywheel.   Both mags are supposed to change at the same time.    There’s actually enough wiggle room with the timing pins in the magnetos to put the timing off by a whole gear tooth.  If it’s off in opposite directions on each mag, there might not be enough range to bring them both back into line at the right point.     But eventually, I got it.   The trick is pulling the prop backwards a bit, then advancing to see the lights go.   The prop has to be pulled backwards enough so that going forwards takes out the gear lash well before the points open, otherwise the adjustments are crap.   But you don’t want to go so far back that you catch the impulse couplers again.    Maybe this isn’t the proper way to do it, but this worked for me.  Also of important note: pull the timing pin out of the magneto before cranking the engine around.   If you don’t, best case is you’ll bend your timing pin like a wet noodle.  Worst case is you damage the innards of the mag.

So I got them as close as I possibly could.    Then I put the shielded ground wires back on.   For safety, the shielded wires go back on the P-lead and ground before the distributor caps go on, because until they’re on and grounded, the mag is hot, and if a spark is triggered, it could swing the prop and really ruin your day, especially if you or your stuff is in the prop arc.   Imagine that.  The prop kicks and sends a blade into the tank of your air compressor or a nearby tool cart.   Prop strike!   Have fun tearing down that engine, sunshine.

By the time this process was over, I was already well past my scheduled time to head back to Los Angeles.   But if you’ll remember, this exercise was only part of the day’s activities.   The other things I needed to attend to were the RPM sensor and the oil temperature probe.   Earlier in the week, Matt at MGL responded to my email and said I needed to remove the ballast jumper on the RDAC and that I could delete the resistor between 12v power and sensor signal.   I pulled the ballast jumper, then proceeded to push the ship out into the wind for a quick test.    Ron and Melody helped me push out and get pointed into the Santa Ana wind.   With the wheels chocked, I fired it up, once again.


It lit right up, after Owen reminded me that I needed to run my boost pump for a second to get some fuel pressure on startup.   It ran strong and smooth, and I had an RPM reading this time, but like a monkey, I forgot to test each mag individually.  But here’s another ground run video:

This time, however, was another milestone.   We pulled the chocks and I actually taxied down to the end of the hangar row, then back.   At that point, my EGT’s were starting to get a little high, so I thought I’d quit while I was ahead.  We pushed it back into the barn and I went home.

Working on engine-start squawks.

7 hours.

Last week’s engine start, while exciting, wasn’t without issues.   The one I thought was going to be most difficult to fix turned out not to be.   The governor operation was reversed, and the fix turned out to be pretty simple.   I just flipped the governor bracket over and mounted it so the cable operates the arm from below.     All I had to do was put a small piece of angle on the side of the arm to secure the #2 fuel injector line.   We’ll have to see about conflict with the cowl.  I put the top cowl on and it looks like there might be about 1/8″ of clearance between the new governor cable arc and the underside of the cowl.  That’s not a huge deal though, and I’d rather put a blister on there than go through the hassle of dismounting the governor and sending it back to American Propeller to see if maybe, just maybe, they can reverse the operation.   Either way, that would cost me in time, in shipping, in ability to test the engine in other ways and debug the two sensor squawks from last time.   The hardest part was the firewall-penetration eyeball.  It’s almost impossible to get to and I still need to torque it down, because it’s not laying flat like it’s supposed to.  I had to take it out so I could get the extra half inch of cable slack necessary for the flipped bracket.

On the others, I had no luck.   No RPM reading, no oil temperature reading, although I have no idea how long it takes the oil to warm up to a minimum of 72 degrees.    The rpm thing is a little scarier because i don’t know if it’s the wiring, the sensor, or the magneto the sensor is attached to.   I’m going to have to break down the wiring of those two pieces and go through it piece by piece.   I really hope the senders aren’t bad, because those things are not cheap.  Replacing both is about a taildragger lesson’s worth.

Speaking of which, I have replaced the ON/ON/(ON) magneto switches with ON/ON DPST  switches.   One side of the switch handles the start circuit and the other handles  the mags.   When the start circuit is enabled, the magnetos are ungrounded.   Since I have two impulse-coupled mags, I can get away with this.   At this point, the starter button on the stick is the only way to kick the engine over now, which is OK, but I plan to install a pushbutton to forcibly engage the starter should the relay the stick button actuates takes a dive.

The magneto timing still needs work.  Even after the switch replacement, the right magneto still makes the engine backfire like mad when it’s on its own.   I’ll need to retime it anyway when I install the new gaskets, but it’s still a royal pain and I’m not keen on doing too many more runs with it in that condition.

Oh, and joy of joys, there’s a small crack in my canopy, in the aft right corner, coming up from the last screw hole on the canopy frame.   I didn’t catch it until now, because it has been  covered with masking tape, which I removed yesterday.   I removed all the protective plastic from the canopy, except for a swatch between the rear canopy and the roll bar brace.   Now the fun is going to be getting that last piece of plastic out from there.  I’ll probably have to dismount the rear canopy, which I would have had to do anyway to get some kind of seal in there.

Engine Start, Part 2.

  1. 5 hours.

Armed with last weekend’s knowledge and suspicion, David and I headed up to OXR to debug and correct the mag and start wiring, then actually start the engine.    We had a copy of The AeroElectric Connection and more importantly, the airport gate card, which allowed us to come and go as we pleased.

Even though we knew what was wrong, it was puzzling, because we were trying to follow Bob’s diagram and make an open circuit for one thing while making a closed circuit for another.   Maddening.

Eventually, we did figure it out and we were rewarded with the sound of an airplane coming to life.

Hooray! Huge milestone, and only three squawks:

  1. No RPM reading.   This is most likely a bad ground or a severed wire coming from the pickup on the magneto.
  2. No Oil Temperature reading.   Maybe this shares a ground with the tach sensor.
  3. (and this is a big one)  Governor operation reversed.   The prop does cycle and the governor doesn’t leak, but it operates the wrong way.   I need to set it up so the prop increases pitch when I pull the blue lever back.

No leaks, no fires, no pieces flying off.   I’d call it a good day!

A Fix and Some Finishing

7 hours.

Although the last three shouldn’t count because I spent them fixing something I should never have had to in the first place.

20130728-192738.jpgThat little corner of aluminum is the bane of my life.    Since I put the wings on, I’ve caught it on shoes, shirts, and finally, the belt on my jeans.   When this happens, it bends.   This time, I bent the crap out of it, so badly, that I had to remove the fasteners on the skin, peel it back, and hammer it flat again.  The light makes it look worse than it is, but it’s still pretty bad.   Fixed now, though.


20130728-192751.jpgThe day wasn’t all bad.   I got the fuel pump overflow plumbed.   I used some of the tubing I had from the MAP sensor install to create a flexible link between the output of the fuel pump to the hard line shown here.   The engine wiggles.  The aluminum tubing doesn’t.   I need a flexible line between the two.

20130728-192805.jpgThe fuel sensors concerned me for a minute.   When I connected the wings, I had some little extra wires that I thought I’d run for spares.   I guess this is why you label things.   After some pondering, I realized these were the fuel level sender wires.   Duh.  The good news is that I didn’t have to do any splicing and apparently I cut them to the right length.   A couple of connectors later and I had fuel level, which was, of course, zero.

No luck on the OAT sender though.   Either my EFIS or the probe is bunk.   OAT reads a steady 32 degrees F.  Have to contact MGL for a new one.

More odds and ends

4 hours.

A few things here and there. I installed the MAP tubing and put a couple of heat shields on the pipes to protect the throttle and mixture cables. I also installed the canopy seal, which is going to need some assistance from some RTV or proseal. I do think firewall forward is just about done, though. The cabin heat SCAT tube rubs on the engine mount a little, but some UHMW tape should fix that. The two things I did that were of major importance were the autopilot test and getting that ridiculous piece of lead off the flange of the left elevator counterweight rib.

A while back, I’d balanced out my elevators, or so I thought. You’re supposed to put the elevator tips on, then drill holes in the lead weight until the elevator balances. Well, guess what? You’re not supposed to have the elevators connected when you do this. I discovered this, freaked out, then riveted a flat piece of lead (cut from an extra counterweight) to the outboard rib.

When I put that away, thinking I was just about the smartest cat in the whole barn, I started imagining the kind of beating a control surface takes in flight. So what happens to a little piece of lead riveted to this structure with a couple of Cherry countersunk blind rivets? The piece of lead comes off and somehow jams the elevator in the dive position and I go screaming downward like a holed Stuka, straight into a busload of orphans on the 405. This has bothered me for months, but I could never find a good opportunity to fix it until Saturday. I drilled out the rivets and put the lead back in a drawer. I also read about a neat trick you can do when balancing your elevators: Pour some lead shot into the tip through the tooling hole in the rib until th elevator balances out, then stick it in place with epoxy resin. Even if it’s not perfect, bias it a little heavy, because paint will change the balance.

I also mounted the MGL GPS antenna on the top side of the glare shield. It works; I get an intermittent GPS position while still inside the guest house.

Oh and one other thing: I dragged the wing cradle over to the shop and tested out the bank servo of the autopilot. Since I actually followed a wiring plan and wired the fuselage-side and the wing side according to it, I was able to test out both servos simultaneously. I do need to make a new ground connection though. The ground from the servo bracket sucks and I was only able to get a good ground by cleco-clamping the ground terminal to a wing rib.

Fiddly bits.

8 hours.

So yesterday, I joined EAA 723 at Camarillo airport. I dunno why, but I’m drawn to Camarillo. Maybe it’s because my cousin finished and flew that little hot-rodded Vari-EZE out of there many years ago, maybe it’s because it was the first place I took passengers when I got my private ticket, who knows? But I like Ventura county, and I like KCMA. It also doesn’t hurt that the EAA hangar is right near the Commemorative Air Force hangar, where there are a multitude of interesting flying machines to gawp at. Everyone seems friendly, and the guest speaker for the meeting was an FAA official, who gave us the rundown on ramp checks and a few FAR’s that are very much misinterpreted by the likes of you and I.

The goal of this is to find a nest for my bird where I can final-assemble, certify, and test fly 313TD. Currently, there isn’t any room in either of the hangars, but one gentleman is 20 hours into phase 1 on an RV7A and another is getting ready to go fly, so maybe a spot will open up soon.

Another benefit to membership is that the chapter has a flatbed trailer suitable for moving a project to the hangar, which I will hopefully need very soon. Yet another is several more sets of eyes looking at workmanship and assembly techniques.

As for actual work on the plane, I got the left intercylinder baffle and the replacement fuel pump installed, which is nice. All the wiring is re-secured where it was, and the fittings on the sensor are properly installed with thread seal.

Today, the morning was spent with my neighbor’s vast crew-cab contractor truck, schlepping stuff to and from two different Ikea stores, one in Carson, one in Costa Mesa. We’re doing this kitchen remodel, and our designer was a complete monkey: he ordered a ton of stuff we can’t/didn’t use and failed to get a bunch of things we need, so Shelley and I had to kill our morning putting everything right in preparation for the final cabinet install tomorrow.

I did get some work in on the plane, but it’s cold in Los Angeles right now, down in the 50’s during the day. You East-coasters, Northerners and Midwesterners chuckle all you want, but the little space heater in my shop wasn’t able to take the edge off, so it kind of sucked to be out there today. Cold hands bashing on metal structures when wrenches slip reminds me way too much of college street surgery, changing out a part lying on my back under a car in an icy parking lot. One of the many reasons I moved out here and stayed.

So no, it didn’t go well. First event of the day was an oh-shit moment, when I snapped the head off one of the bolts holding the throttle bracket on. That was a study in anger manangement, and I had the presence of mind to self-soothe through the initial impulses of throwing the torque wrench through the sliding glass door. Tantrums do not fix airplanes.

Fortunately, there’s an O’Reilly Auto store not a city block from my house, so I picked up an EZ-Out (sorry, screw extractor- generic) and through some miracle, managed to get the offending bolt out of the sump. I guess I’m now 1 for 5 using those things. At least I didn’t snap the drill off inside the broken bolt like I did with the wing attach adventure from a couple of years ago.

I installed the breather tube and safety-wired the mixture bracket bolts after torquing them with a different torque wrench (A Snap-On dial-type I bought used) and those are fine. I’m hoping I’m nearly done messing around with Adel clamps. Space is getting tight.

More firewall forward progress… And regress.

5 hours.

The latest load from ACS allowed me to finish a couple of things. I got the oil pressure line/oil line adel clamp securing done, I finished the purge valve bracket assembly, I got the fittings installed on the oil cooler, and I actually got the prop governor installed, and the front center baffle reinstalled. I painted the throttle and mixture brackets and put them back on, and I re-secured the current sensor to the alternator wire. This is where it all went south.

I also finished attaching the fuel hose to the throttle body, and securing it to the intake pipes, but this is where it all went horribly wrong. I had purchased a steel 1/4″NPT-AN6 elbow, but I was using one of the blue aluminum ones to get my fit and mount done. The Floscan is on the firewall, and the elbow points toward the intake pipes on the left hand side of the engine. I had the fuel line connected to this. Somehow, in backing out the AN6-1/4″ NPT elbow on the output side of the Floscan fuel flow sensor, I managed to cross-thread the fitting as I was taking it off. Taking it OFF, not putting it on. So what it did was essentially pull on one side of the threads but not the other, kind of just bending the whole thing. Long story short, I completely destroyed the threads on the Floscan and it will have to be replaced. That was a very expensive ($210 from MGL Avionics) lesson in not connecting loads to dry-fitted NPT fittings loose enough to jam up. I did finally manage to extract the threads left in the Floscan housing and re-tap them, but I don’t trust the connection anymore, especially not with high-pressure fuel running through it. Now would be the time to switch to the Red Cube, except that now I’m reading about failures of these units, plus my firewall mount is drilled specifically for a Floscan. The time for experimentation is pretty much over.

I also got a tube of red RTV for sealing the baffles and putting some blobs wherever things might rub that I didn’t get with Adel clamps.

I still have to track down a hose from PHT (throttle body to fuel spider), but that’ll have to wait until tomorrow. My intercylinder baffles are on their way, and once those get installed I can permanently reattach all the baffles. I can also rivet the firewall to belly skin, because I’m giving up on making an exhaust fairing for now.

FWF Sensor Wiring DONE!

10 hours.

This is yesterday and today. Yesterday started with an epic shop cleaning. I have those every so often, and it’s helpful, extremely good to have a workbench devoid of clutter, if only for a little while. Stuff got put away, stuff got tossed out, floors got swept and vacuumed.

The fiddling with the FWF stuff has been progressing as well as I can make it, and the wiring has been a bit of a challenge, especially since the K-Type wires on 2 of the sensors were too short to make it all the way to the RDAC engine monitor module. Then there’s the business of where to actually run things so they don’t rub on other bits or get roasted alive by the exhaust pipes. I got the current sensor done, although I suspect I’ll have to flip it over, I got all the temperature sensor wires run, wrapped, and grounded. About the only thing that didn’t happen is the MAP sensor, which needs a 1/8″ to 1/8″ barbed elbow fitting. Searches have turned up no such animal so far. But technically, that’s not wiring, so legally speaking my FWF sensor wiring is done.

Here’s the beginning of the attachment process. The overbraid K-type wires are a ratbastard to work with. They’re a little like shielded cable, but they’ve got an extra woven sheath below the braid that absolutely sucks to try to get off. Maybe there’s a trick to it, but this was by far the worst suckage of this task.

A slightly different angle. On the RDAC (the grey box), the first 4 pairs are exhaust gas temperature thermocouple inputs. Can’t change them, which kind of sucks, because then I wouldn’t have had to extend a couple of TC wires to make them reach the proper input. Oh yeah, splicing TC wire sucks too. The ty-wrapped snarl at the bottom is the FloScan fuel flow sensor wiring. I installed the FloScan months ago, but never got around to properly wiring it until today.

Still a rat’s nest, but there is progress there. This is right after grounding all the things that needed grounding, being the FloScan, the tach sensor, and the RDAC itself. The EGT and CHT probe wires haven’t been cut to length yet, nor have they been secured by Adel clamps to the firewall.

Ahhh, that’s better. Still some wrapping to be done, but all the wiring is secured and attached.

This shows the oil pressure sensor hooked up and run into the bundle. This will benefit from a couple of Adel clamps attaching the main bundle to the engine mount, but we’re in a pretty good place.

Even more fun: Once I got everything wired, I fired up the EFIS and decided to test out whether or not I’d actually inserted the right wires into the right terminals on the RDAC. This was done by means of a heat gun and a blowtorch. The heat gun raised the temperature of the actual cylinder itself so I could check the reading on the EFIS. The blowtorch was for the EGT probes. Suffice it to say that the method of testing was a great success: I found I had miswired the 1 and 3 cylinders for EGT, and the 2 and 4 cylinders for CHT. It is now all fixed.