Monday, July 31, 2017

Fuel Pump Failure

End of June I wanted to depart on a breakfast flight to Benson with my wife, when ~3 minutes after take-off the cockpit quickly filled with a strong stench of gasoline. It wouldn't pass for a minute or two and I knew what that meant.
I turned around and climbed to 10,000 feet. Once over the field I turned off the electric fuel pump and watched the fuel pressure drop quickly to 0.2 PSI when the engine started running rough. I turned the electric pump back on and landed the plane, canceling our breakfast plans. Apparently the diaphragm on the engine-driven fuel pump had ruptured and leaked gas out the weep drain and overboard. At just 207 hours, that's rather frustrating.

out comes the old one....

It didn't take too much time replacing the pump, besides the inaccessibility of course. With crow's feet it's pretty impossible to put any controlled torque on the bolts and that includes the fuel distributor block.

... and in with the new one.
I did an extensive engine ground run at full throttle with only the engine-driven pump running to see if it was working alright. Then I headed for a test flight, steep take-off staying over the field and climbing to 10,000 ft.. Stayed there for 20 minutes before moving away, always staying in gliding range.
Without the electric pump running, the fuel pressure was oscillating between 2.6 and 3.4 PSI. I never saw that with my previous pump. It was a very repetitive oscillation pattern, almost rhythm-like.
Even on a cross country flight a week later I could still see this oscillation pattern but it never dropped below 2.6 PSI.
After coming back from Oshkosh (flying commercial) I had another flight where the mechanical pump produced rock solid 3.9 PSI. No oscillation and no jumps. Apparently, a new fuel pump does need some break-in period.

Let's see how long this one will last. Fingers crossed!

I'll disassemble the old pump to verify the failure mode of a ruptured diaphragm and I will update this post when I have the pictures.

Monday, January 9, 2017

Operating Limitations Overly Limiting

One of the things that most builders forget quickly after the exhilarating experience of turning a heap of metal into an airplane - also known as "receiving an Airworthiness Certificate - is that the whole world of experimental flying our kit planes is built upon something called Operating Limitations.
Without the OpLims, there is no experimental flying.

You will receive your Operating Limitations on the day you get your Special Airworthiness Certificate, better known as Pink Slip, and it explains in detail what you can do with the airplane in the initial trial phase, aka Phase I and what you can do once Phase I has been completed - which is referred to as Phase II.

You airplane will likely spent most of its flying life in Phase II, so those limitations are the most important ones for any owner of an experimental airplane.
When I received mine, I studied them and I found one odd set of limitations regarding Phase II and doing anything but flying in VFR conditions during the day.
I have to admit that because of the excitement of accomplishing a successful airworthiness inspection as well as the fact that I was going to fly the plane as a Sport Pilot this tiny weirdness did not bother me too much.
What it was you ask?

Well, here are the two limitations quoted from the OpLims:

10) This aircraft is to be operated under VFR day only

11) After completion of phase I flight testing, unless appropriately equipped for night and/or instrument flight in accordance with 14 CFR part 91.205, this aircraft is to be operated VFR day only.

The way the limitations were applied to Phase I and II respectively was by adding the numbers of the limitations that applied to a paragraph titled Phase I and II respectively.

Mine looked like this:

Phase I Limitations: 1, 2, 3, 4, ..... , 10, 12, ......

Phase II Limitations: 1, 2, 3, 4, ..... , 10, 11, 12, ......


Of course, you wouldn't want to fly your plane during the trial phase in adverse meteorological conditions like IMC or at night. So having 10) in phase I is perfectly fine. As 11) wouldn't apply, it is not listed under Phase I.
However, Phase II has apparently 2 limitations that deal with flying at night.
10) prohibits it altogether, and 11) would allow it if the plane was properly equipped (mine is).
10) is clearly more restrictive and I would think the FAA would use the more restrictive one as we are dealing with LIMITATIONS which are restrictive by nature.

Why do I even care? Didn't I say I was flying as a Sport Pilot anyway and Sport Pilots are prohibited from anything other than flying VFR day.

Yes, that was the case. However, I have started to work on my Private Pilot License and I wanted to use my RV-12 for all the required lessons I have to take with a CFI. That would include 3 hours of flying VFR Night and my OpLims clearly prohibit that.

I searched VAF and found that at least one other builder in Arizona was hit with the same mistaken limitation and that his DAR had told him that the FAA would not issue new OpLims but that this is clearly a mistake and only 11) would apply.
Hm, sounds like one would have to trust the federal government. That's quite a stretch, isn't it. And on top of that your CFI would have to be satisfied with this "explanation" too. Pretty unlikely as their job depends on it.

I contacted my DAR who had made my RV-12 an airplane about 3 years ago and explained the situation. He agreed that this set of limitations made no sense and contacted his person at the FAA who also agreed(!). What were the chances? Honestly!
My DAR was given authorization to issue amended  OpLims in accordance with the latest updated regulations and I should have them in my hands before the weekend. The new limitations are fully spelled out, much more narrative than the old ones and more importantly the issue of no night flying in Phase II has been eliminated.

What is the morale of this little story? Check your OpLims when you get them and get them straightened out if you find anything that strikes you as odd. It's likely easier when they just got issued and the copies haven't made it to the FAA yet.

Monday, November 21, 2016

Broken Wire, Or Is It?

What an interesting annual inspection I am having this year! After last years slip into the new year (I usually started inspection at Thanksgiving), I decided to make some changes this year.
One of those changes was to start earlier and instead of planing on a long downtime and doing it all at once, I was going to split the annual up into smaller packages that could be completed in a weekend and allow the plane to usable on the following weekend.
So I did the wheel bearing service with replacing the brake linings about a month ago, utilizing my newly acquired HF Racing Jack.

Worked like a charm! No more assistance needed to get the plane on sawhorses for this task.

I also completed the engine inspection 2 weeks ago and hopefully repaired an intermittent right CHT connection by recrimping the wire. Unfortunately, the problem only shows in flight. The sensor always reads fine on the ground but goes haywire right after take-off.

Other than that, no findings firewall forward.

Now the plane is down for a longer period as I opened the tailcone bulkhead and working on the flight controls. First item was to fix a problem that occured the first time after a SkyView software update (I think it was the update to V7.0) when it finally supported the standby network wires we had put in during the build. Mine was showing as defective and later versions even allowed the problem to get pinpointed to the standby network of the roll servo (behind the bulkhead).
Having access to it now, allowed me to finally see what was causing the problem. After a lot of measuring I was surprised to find that it is a problem with the wiring harness. Somewhere between the Y-crimp at the pitch servo which extends the orange wire (standby network 2B) from the roll servo through the tunnel to the aft of the bulkhead where it connects to the roll servo, this wire is broken.
If you have built an RV-12 or seen the tunnel of a completed one, you know that replacing a wire in this over-stuffed tunnel wire bundle is a very difficult task to say the least.
As I already had worked around this issue before when adding a wire harness for the ADS-B receiver (which I put behind the bulkhead) which runs from behind the panel under the longerons on the left side all the way aft of the bulkhead. I did make this wire conduit larger than necessary for such a case as this one where I see the need to add more to it.
I will utilize this wire path to connect the 2B network wire to the roll servo. Ideally these network wires should run in pairs and get twisted to reduce the noise they might pick up on the way. I am still contemplating if I just accomodate that by running a twisted pair back to it and just not use the one wire of the original harness that does work ok.

It turned out not be a broken wire at all.

After I ran the replacement wire for the failed connection to the 2B terminal of the servo and crimped it onto a showrt stub of the orange wire that connected to the servo, I still did not have a successful continuity from the panel to the servo. This could only mean one thing!
I disconnected the spade connectors for that wire on the servo side and measured again. No connection! A closer look into the connector revealed that I must have pushed the stripped wire passed the crimp part and had crimped the connector onto the insulation of the wire.  Duh!
Measuring between panel and the stripped part of the wire confirmed that it had a good connection.
I fixed this crimping issue and hooked up the servo and the Skyview display confirmed that everything was working now:

I also confirmed that the SB that most worries me (forogt the number , but it is the one about the little bridge taking the flaperon forces where the pushrods connect to the torque tubes) is still not an issue at 178 hours.

Left side

Right side

The rest of the inspection went without any hitch and a maintenance flight confirmed that the roll servo is no longer creating warning messages about the backup network within Skyview..

We also did not have any signs of overheating the voltage regulator:

Although I wish it was running a bit cooler. Maybe I do prepare to work on a scoop for the next oil change time when I have access to the radiator side of the lower cowl.

The ignition modules were doing fine too:

She's back in service since the Day after Thanksgiving.

Quick update from about a month later:
- the recrimp of the right CHT probe connector did its job for now. The CHT indication is solid but I expect that the repeated heating and cooling of the crimp as well as the engine vibrations might require repeated attention to these connectors in the future.
- It's really nice to fly without a constantly flashing warning indicator on the Skyview!

Monday, May 23, 2016

The End of The Blue Plague

I was out at the hangar on Saturday morning to fly out for breakfast but the wind had not developed as forecast. It had looked like I might have a small window to get to fly before 10am but it turned out that the wind had developed way before that. At 6:30am it was already at 11kts, gusting to 16, and 40 degrees cross to the runway with the crosswind runways closed for maintenance. The wind would turn further cross during the morning and also increase to 20+ knots, so I decided to just take a quick hop around the pattern as I has not flown for a few weeks due to a vacation.
The good thing about this blown out morning is that I finally got to remove the last trace of the Blue Plague from the bottom of the left wing.

Not a very good photo but I hope you can see the lack of any blue plastic sheeting on the bottom skin.
I guess that completed the build :-).

Next I will really have to take care of the unpainted plastic and fiberglass surfaces as the sun is starting to take its toll on the surfaces.

Thursday, March 17, 2016

Cracked Gas Can Caps

This is not really an issue with the airplane itself which is what this blog is focused on, but as the fueling system I use is one that many others are using as well, I thought it might be of use to some anyway.
After 3 years of use 2 of my 3 caps for the 7.5 gallon Flo-Fast jugs showed cracks like the one in the following picture.

I contacted Flo-Fast and they said that they had a batch around those 3 years ago that had some size deviations that led to cracks like these. They were sending replacements free of charge. They haven't arrived yet but I had already ordered replacement caps from a vendor (I am pretty impatient when it comes to exchanging emails. If a company doesn't answer within 24 hours I assume they don't care about emails) before we got some successful email exchange going. The caps I received are clearly made from a different material (I would guess it's fiberglass and not polyethylene or styrene like the old ones), so maybe it was not so much a size issue but a problem with the material used.
My jugs got exposed to a lot of temperature difference while closed and that might have added to these stress cracks.
Anyway, now you know that it might be a good idea to take a closer look at these caps when you notice a pronounced gas smell around them.

Update: The 3 replacement caps from Flo-Fast arrived over the weekend and they look exactly like the ones previously ordered from a car supply shop. The material looks like it is fiber-reinforced and seems to be a lot sturdier than the original caps that came with the jugs. We'll see how they hold up.

Monday, January 11, 2016

Yet Another Annual

The plane is down for inspection since December. I am by no means ready to formally document everything yet but there was one issue I should inform readers about before they run into the same problem.
The nice blue air filters I had installed during the last annual (and which appeared just fine in March when I installed the Ducati regulator) both showed cracked bodies just behind the attachment flange. They were both cracked almost to 50% of the circumference and clearly did not provide much filtration at that point. When I removed them, I found that the rubber appeared to have worked its way into the serrations of the hose clamp that was holding it on, on both. One was so deeply connected that it ripped the rubber of when I removed the hose clamp. It almost looked like the rubber or plastic was not able to take the heat under the cowling as it looked partially melted into the hose clamp.
I was about to order replacement filters from the same place and am happy now that they ran out of stock. That gave me a moment to reconsider that they might just not be the best solution as an aftermarket product anyway.
I found a smaller sized K&N but decided to give Green Filter a chance. They have one model that just fits the Bing flange and I like that it is a fully paper/rubber model with no metal parts.

It has a pretty nice clearance and a very good fit to the rubber flange of the Bings. Actually, I had taper the inside of the filter flange a bit to be able to get them to slide on. The type number is 7069 and it appears to be only one that would fit the Bing inlet. Let's hope this one holds up to the vibrations better than the blue filters.

It was also time to replace the spark plugs. I did not do it at the 50 hour mark as they looked prestine. I just cleaned them and put them back in. They still did not look like they needed replacement but I did notice a deeper drop during the 4000 RPM ignition test in the last months, so I went ahead and threw them out.


Of course I had to try something new for the replacement, so I installed the Iridium version of the same DCPR8E plugs which have an EIX suffix. They are a bit pricey but then again they go into an airplane so we are used to overpaying, right?
The result by the way was amazing. The ignition test dropped the RPM only by 50 instead of 70-90 with the 8E plugs. We will see how they hold up and if this deteriorates over time but for now I think they might be worth it.

I might have to improve the air scoop for the voltage regulator as the temp indicator hit the lower mark so far and I don't believe she flew a lot during Arizona's hot summers....

One of my concerns last year was the plastic buffer in the tail cone that should keep the lower cable for the elevator from sawing through a bulk head. It was worn down very close to the metal and I had greased it with silicone grease to reduce friction.

I pushed the cable out of the way to get a better shot of the actual friction area and it looks like it did not deepen any further since last year. If it was touching the metal the primer would have rubbed off and there is no evidence that this is happening yet. I just re-greased it and will check again next year. The tension of the cables held up pretty well too and I did not tighten them.

One of the improvements this year was to finally add the APRS radio beacon. I am a licensed Amateur Radio operator, so it was just natural to put that license to additional use with the RV-12. I took quite some time to choose the right antenna for this project. I had an antenna from DeltaPop Aviation but was afraid that it was too heavy and catching too much wind load for the super thin skin on the bottom of the tail cone. I looked into mobile Amateur Radio antennas and considered different antenna bases and so on. After long deliberation with myself I decided to go back to the initial thought of using the whip antenna from DeltaPop Aviation. It is exactly tuned to the APRS frequency, has a matching network that eliminates static build-up and it is proven to survive the 120 knots TAS of an RV-12. The mobile antennas won't get used much above 80 mph and their radiator could separate from the base and fall off the plane if the set screw ever gets loose. All I needed to do was to properly reinforce the skin on the bottom of the tail cone to avoid vibration cracks over time.


The little fin is the PCAS antenna that I had installed last year. I finally hooked this one up to the Zaon MRX unit this year too. The base of the APRS antenna is 4 feet away from the base of the VHF Com antenna and I only had to slightly increase the automatic squelch of the SL-40 to not open up when the beacon sends a position (from 56 to 75).


So you can see that I took the reinforcement very serious. The large sheet is 40/1000" thick and a foot long, I put the transponder reinforcement plate on top that came with the RV-12 antenna and which we never used and I also installed some left over angle on the perimeter of the base sheet to stiffen it up while keeping it very light weight.

On the interior I did not get very fancy for now but it might turn out to be a good spot anyway. The Byonics radio was attached with cable ties to the tank fill spout (some foam in between to dampen vibrations). And the GPS antenna for the radio was riveted to the center bracket of the rear window to give it a good view of the satellites.

The only required switch on the dashboard is a breaker of 3 Amps that is labelled "Beacon" which I felt was quite accurate.

It worked right away when I switched Main and Avionics on. If you would like some advice on where to wire this into the existing 12V bus without causing a fire, just email me.
As you can see the Zaon MRX was finally installed too and it is connected through the 1 Amp breaker to the Avionics bus. So far I have not seen a case where it would have detected traffic when ADS-B did not see it but I haven't been flying in rough mountainous terrain yet, so the jury is out if this PCAS install was a waste of time.
The APRS beacon works great and you can now follow the travels of N128TL by following this link.

Besides adding stuff I was also able to determine the cause of my problems with the EGT sensors. The sensors work just fine but the wire of the right sensor (on the sensor side) had vibrated off. The stiff wire of the sensor harness is not just hard to reliably crimp a connector too but is not itself not very happy about vibrations. I did a brute force approach with a tube of red solicone to address that after soldering (yes, yes, I know, we will see if it breaks again) an extension to the broken harness wire.

I just smothered the EGT temp wires in a layer of red silicone in the hopes of softly reducing the vibrations enough to avoid future breaks.

This should have been the last item on my list but then Dynon announced the release of the GPS-2020 that would be the last piece in my Avionics system to be fully 2020 compliant. I couldn't pass that chance up to replace my current GPS while the plane was all opened up now, could I?
Dynon was very quick in delivering the GPS and I figured from the RV electrical schematics that this was a drop-in replacement. As I did want to reroute the 4 wires through the firewall, I just cut off the wires of the old GPS and soldered the new wire harness to the existing wires.

Getting the old GPS out was a hassle. The metal lock screws are almost impossible to hold well enough to turn the bolts. I ended up dremeling the case of the GPS back to I could get a wrench on them. The GPS still works but sure doesn't look so good now. Oh well.
Please note that you do need to upgrade your SkyView software to V14.0 or beyond to be able to use the GPS-2020. And you will have to go into the setup and manually make all the changes to make it understand that the previous GPS has been replaced.

I also checked that worrisome Service Bulletin item about the flaperon torque base in the tunnel. Sorry I can't remember the exact phrasing or the number right now. It's the one where an aluminum bracket starts developing a crack from a rivet hole after around 200 hours of operation. I took photos and they show no issues yet at 113 hours.

And the final installation was a piece of plastic that I designed and printed out on a side project I completed in December last year. It's a Kossel 3D printer and it allows me create neat stuff that would be hard to find otherwise. In this case it is a headset hook that allows me to finally have a good place for my Bose headset while I am away from the RV.

This is a prototype and I am trying 3M double sided tape on the back (the red type) but I a mnot sure if it will stay there when the plane heats up in the summer on the tarmac. So I will design another one that takes rivets for installation - just in case. Anyway, for now I am very happy that the headset is nicely stored under the avionics bay and out of the sun.

Monday, March 23, 2015

Replaced the Ducati Regulator

My Ducati Regulator/Rectifier had started to show signs of upcoming failure. It had not reliably charged the battery on the ground while running at low RPMs after initial engine start. It seemed to have worked ok once in the air and for subsequent engine starts during a longer flying day but other have reported this behavior before the regulator eventually failed.
So I decided to not wait for that failure and to replace it. I also wanted to address the obvious reliability issue with this regulator as well as improve the cooling issue with the air pick-up from the cylinder head plenum which does apparently not provide enough air at low RPMs on the ground.

After a lot of research and consideration I decided to go the most expensive route. I had initially settled on an aftermarket Harley regulator from Compu-Fire (55120) which I still believe to be a very good solution. However, the installation of said regulator would have required wiring work (to adapt to the existing Rotax plug) as well as mechanical work to provide an adapter plate for mounting and to rebuild the metal cap to control the cooling air flow.
I eventually found a better solution that would not require such work, or so I thought. The regulator is from a company in Germany (and before you ask, at this time they do not ship outside of the European Union, but I will be talking with them to solve this issue) called Silent-Hektik. They provide a wide variety of regulator/rectifiers for different application and also one for the Rotax 912/914 series (English translation) of engines. They also supply regulators for Jabiru 2200/3300 engines and even for common auto/motorbike conversions (at least common in Europe).
(Edit: Back when, I ordered the F4112 which seems to be replaced by the F4118 now)

This past weekend, I decided to put it in. Unfortunately, I forgot to bring my cell phone on the first day, so I could only take photos on the second day when everything had already been installed, trying to catch as much of the installation as possible.
Removing the Ducati regulator was fairly simple. I had just removed the upper cowling and that was sufficient, albeit without the lower it might have been somewhat easier.
Getting the Silent-Hektik regulator in revealed two stumble blocks. While the case has the same outer dimensions as advertised, the top has one cooling fin more than the Ducati. This required either to rebuild the top cap or to bend it open and wing it. I opted for the latter. I also installed a flange on the cap, moving the hole over to the center. The flange allows for a 1" SCEET hose to attach with a worm clamp. The flange can be found here.

Take a close look at the corners of the cap in the following picture. I resealed the holes in the corners with silicone.

The other "issue" is that while the mounting holes on the Ducati are a lot bigger than you could ever need them to be, the Silent-Hektik has holes that are probably meant to be used with metric M5 or M6 bolts and an AN4 bolt will not fit. Of course, this is what Van's chose to mount the regulator to the firewall shelf. I drilled the holes open with a drill bit size that allowed for some slack of the bolt and it lines up perfectly with the existing nutplates on the shelf.

I wonder why Van's chose AN4 bolts as holding the regulator to the shelf is clearly not requiring a lot of force, but it already presented a problem with the Ducati as there was hardly any clearance to get a socket over the head of the bolt. It is even worse with the Silent-Hektik regulator as they really used every possible area to contribute to cooling the regulator and that means you will need a 7/16" crow foot to get this baby torqued down. With AN3 bolts the socket would have likely cleared just fine.

These were the only issues that I ran into that would make this installation not quite a perfect drop-in replacement. The plug fit like a glove though and snug right in and that is what I most cared about.

By the way, when I removed the Ducati regulator I could not see why it is frequently failing in this application. I had recently attached a temperature sticker on the side of it (around 50 hours) and it did not show a significantly high peak temperature.

It showed that it had reached 160F definitely and did not exceed 170F. Now, this is not a temperature you want silicon to operate at for a continuous time AND Ithe sticker was only on it during the cooler months in Arizona. So, yes, I can see that running it at 30-40F higher ambient temperatures could push it into the death zone.

The plug and the connector looked good though, no heat damage from high currents through bad connections or anything.

To avoid this overheating from happening again, I removed the existing plastic hose and plugged up the plenum pickup with a piece of scrap fiberglass glued on with silicone.

Not pretty but will definitely do its job. I also installed another temperature sticker on the Silent-Hektik.

For the new air pickup I drilled a hole in the top of the radiator tunnel and glued on another 1" flange using Pro-Seal.

The white insert in the flange is a piece of PVC tube that I glued in with silicone (so I could remove it later if I decide to install a scoop to increase pressure on the cooling pickup). From the tunnel it looks like this:

I hope this will help to divert air to the pickup line and increase flow during low RPM on the ground. I am currently at 87 hours and my next oil change is coming up. At that time I will check the temperature sticker and decide if need to install a scoop. If so, it will help to have the lower cowl removed and work on it from the radiator side.

As the Pro-Seal needed a day of curing before putting any stress on it, I had to return home at that point and let it sit.

The next morning I came back to the hangar only to find that the story about keeping Pro-Seal in the fridge to keep it from expiring are not true. My can had an expiration date of late 2013 on it and I kept it to do simple sealing jobs that would not require fuel resistance but where you wanted the paint-ability of the Pro-Seal. I kept it in the fridge ever since I used it for sealing the tank and the firewall in 2013.
Well, it does not cure anymore, at least not within 18 hours as the it used to. I know that it needs longer to fully cure but I would have expected it to be at least strong enough to hold the force of a bent SCEET tube. It did not and the flange was pulled off and tipped forward almost immediately after putting it on.
I had 3 options at that point. Remove the gunk and glue it on with 5-minute epoxy. I did not want to do that as the 5-minute epoxy gets brittle and could crack and break under vibration. I could wait and see if the pro-Seal eventually cures after a week. Not too excited about that idea either.
I could riveted the flange on and care about the Pro-Seal. That needed two people though as one would pull the riveter from the top of the tunnel while a helping hand would hold the washer on the shop end of the rivet that would relieve the stress on the thin fiberglass of the tunnel over a larger area and avoid cracking under vibration.

That very moment my friend and hangar neighbor Tim landed his Drifter and got ready to push her back. 5 minutes later the flange looked just fine with 4 rivets holding it down.

The last steps were to find a good path for routing the hose and to put the tie wraps back in place keeping everything tight together. The SCEET hose is 3' long and fit perfectly without any trimming.

It was still early enough to take her for a spin and I wanted to see if the regulator was doing its job. I pulled her out, jumped in and started the engine. While the Ducati needed at least 2500RPM at this point to drain the battery at only -1 Amps or just be at +/- 0, the Silent-Hektik presented me with a much better picture:

It charged with even more than 2 Amps before I could my phone out to take this picture. And all that at the lower range of the green arc. Perfect!

In flight I saw the voltage rising higher than with the Ducati. The Silent-Hektik charged the battery to 14.3V max. and kept it at 14.1V most of the time where the Ducati would be showing 13.7-13.8V. I heard that going for the higher voltage should enhance the battery life and maintain its performance by keeping the plates clean.
We shall see...

After my next oil change I will follow up and let you know how that air pickup is doing.