Today I spent some time measuring and adjusting the flight controls. My goal was to get everything rigged within the Vans specs, and to correct an issue where the flap and aileron training edges were misaligned.
I downloaded the Vans Specs and copied them into a spreadsheet, then measured the various deflections using a digital level. Note, the elevators are not well aligned, so I measured the neutral point as the middle of the left and right elevators.
Everything measured in-spec initially, but I wanted to fix the trailing edge problem. This came about because I used the wing-tip template to set the aileron neutral point. A few years ago Vans changed the plans, and they now recommend ignoring the template, and simply putting the flap in the up position, then setting the trailing edges of the ailerons to be aligned with the flaps. I had to disconnect the for-aft pushrods, unwind one of the rod end bearings one full turn (on both sides), and the alignment came out perfect.
This weekend I was busy with many tasks, most of them small and not newsworthy – just things I want to get done ahead of an airworthiness inspection. However, one important thing I had not done, was test the ELT. To be honest, I was kind of scared of testing it, and had installed it a long time ago thinking I would one day figure out how to test the thing. In hindsight, I should have learned all about it back then and performed the test. I ended up with a couple of problems that took an annoyingly long time to solve.
Testing an ELT is tricky, because setting off a locator beacon can trigger search and rescue operations, so you want to make sure you understand the rules and regulations, and the operations of the unit. Unlike older ELTs, this one will transmit aircraft info including GPS coordinates to a satellite constellation. I started by reading the installation manual (I had read it before, but needed a refresh). The first thing was to register the unit with NOAA, which is compulsory before triggering the signal, so I went ahead and did that. Derek, if you are reading this – fyi, you are one a small group of emergency contacts. There are a set of tests that require equipment I don’t have, so I skipped that and moved to the ELT self-test.
The ELT’s interface. The antenna port is at the top. The switch allows someone to manually activate the device, or to perform the self-test. Next to the switch is the LED which indicates the state the unit is in. The 15 pin D-sub port connects the unit to ship power, nav signal, the remote switch, and the buzzer.
The self does an internal system check, then broadcasts on 121.5 for two cycles, tests that it’s receiving Nav info, and sounds the buzzer. If anything is wrong during the test, the LED flashes and number of flashes indicates the error code. The annoying limitation on testing the device is that it’s only allowed during the first 5 minutes of each hour. That means if you have a problem, you really don’t have time to troubleshoot it before that time expires and you have to wait another hour to test if you have resolved the issue.
On my first attempt to run a self-test, I waited for the hour to strike, pushed the self-test button, and absolutely nothing happened. Hmm. I decided to take the unit out, remove the battery, and give it a more thorough inspection. Everything looked as it should, and I reinstalled the battery etc, still wondering why it wasn’t working. As soon as I plugged the battery into the receptacle on the circuit board, the LED lit up and it beeped. Then it kept beeping! I quickly consulted the manual and saw a warning that connecting the battery can sometimes activate the ELT. The fix is to turn it to “On” (activate it), then back to “Arm”. This quickly resolved the issue, but I still accidentally transmitted for a few seconds.
Backside of the circuit board with some QA markings
With that figured out, I reinstalled everything, waiting for the next hour, and performed the self test again. This time the unit beeped, and then gave a couple of different error codes. Hmm. Consulting the manual
I ended up pulling apart the D-sub connector and inspecting the wiring going into the ELT and buzzer. I discovered the buzzer power wire was in the wrong position, but everything else seemed to check out. I confirmed the correct wiring at the remote switch on the panel, and couldn’t find any other issues. I put it back together, waited for the right time and retested. Still an error.
After some online research, I discovered that while the ELT supports several different formats of data from the Nav system, when it’s programmed, it’s set for one format – it doesn’t dynamically switch between protocols and baud rates. With that info, I went into the avionics configuration and found there were dozens of different available formats that the G3X can output, and the one selected for the transponder was probably not the right one. I adjusted to the “Aviation” format, re-ran the self-test and happily the self-test passed!
Unfortunately the buzzer still didn’t sound, so I need to troubleshoot the buzzer. Next time I’m at the hangar I’ll test the buzzer manually with a battery to verify that it’s actually functional. If it is, then there is likely a problem with the wiring – perhaps a bad ground somehow – or the buzzer power is configured to come from another pin on the ELT. The least likely problem is the ELT itself not supplying power. I really hope that’s not the case, as I don’t want to have to replace the unit at this stage.
Update: on June 1 I was able to successfully test the buzzer. The issue was a mis-pinned power wire (pin 15 instead of 8), and once corrected the buzzer worked as expected.
The buzzer. This thing is to alert anyone nearby that the ELT has triggered. This is actually super useful, because the ELT is hidden from view in the tail, and without this alarm you wouldn’t know if the ELT had activated (except for the flashing LED on the instrument panel).
Today I arrived at the airport early to prepare for and conduct a taxi test. In my test card, I planned to check steering operations on the ground – braking action, differential braking, and tail wheel steering. I also planned to calibrate the magnetometer on the compass rose. Everything went well, and I learned a few things along the way.
The compass rose is at the other end of the runway, and I am trying to minimize the amount of engine run-time at low RPM. Because the engine is brand new and needs to be broken in, the recommendation is to run it at high power for the first few hours of it’s life. Since I can’t do that on the ground, I don’t want to be running the engine longer than I need to before it’s flying. However, I need to calibrate the magnetometer, and it’s a sensitive instrument. Calibration requires slowly taxiing in a circle in an area free from any magnetic interference, which is what the compass rose provides. I decided it was worth the extra 5-6 minutes to taxi there and back to give the calibration the highest chance of success.
The engine started nicely, and I started the taxi test right away. Since the tower was operating by the time I was ready, I also had a chance to test my Com1 radio, tower gave me a 5/5 rating – loud and clear. With the cowling on I could check my visibility while taxiing. By sitting up straight I can see over the top of the cowl, although the first 200 feet in front of the plane is obscured. Moving my head to the left I had a nice view down the side of the cowl, and I found that to be the easiest way to monitor the taxi way ahead. Engine instruments showed the engine was running well – consistent CHTs, EGTs, and oil temp & pressure.
Arriving at the compass rose, I positioned the aircraft on the left side, facing due north, as described in the Garmin manual. I needed to boot the PFD into Calibration mode which required shutting down the G3X system, and booting it back up. I shut off the standby battery first, then the main battery. For a moment I was surprised that all the avionics were still operating. Then I remembered the alternator was still on, and was powering the system. I shut off the alternator, and the engine started to shut down! I immediately realized what had happened, but didn’t have the reaction time to stop it. I pulled the mixture to complete a normal engine shutdown. The e-mag ignition system depends on main buss power at low RPM. The mags have internal Alternators to provide power, but only about ~1200 RPM. I was idling the engine when I shut off the power, and the ignition system stopped igniting.
While the engine was shutdown I booted up the G3X into calibration mode and get the magnetometer test ready to go. The engine started right up again, and I followed the on-screen instructions to slowly turn in a circle to the right, holding position every 30 degrees. The whole process took about 10 minutes, and thankfully passed at the end. I taxied safely back to the hangar and shutdown. The total engine run time was 18 minutes, a bit longer than I hoped.
Analyzing the engine data again, everything appeared normal.
Post taxi inspection. Exhaust is starting to change color. No sign of oil leaks, a track of splatter from the exhaust joint lubricant.
Here is some of the engine and flight data. Note, the data between the vertical bars is interpolated, since I only had data from the beginning and end of the test, and not while performing the magnetometer calibration (the time between the bars).
Not sure why the airspeed is indicating 18 MPH before and after the test, and why it’s so much higher than the ground speed. There was approximately 5 knots of wind, and no other reason I can think of that would cause it to read high. Prop blast could be a factor after engine start, but the pitot tube is well outside the prop blast area.
Today was a major milestone – the engine ran for the first time! My friend Donnie met me at the airport at 7am to help advise and provide a second set of eyes on the process. It was early enough to be very quiet, allowing us to focus. After a few preparatory steps, we pulled the airplane out onto the taxi way and started the test. I had a test card written that covered everything I wanted to test – engine instruments, ignition, prop governor, alternator etc.
Everything performed completely as expected, and I was very happy and relieved to complete the test successfully. After verifying everything engine-related, I also took the chance to test the brakes. Donnie removed the chocks and I rolled forward a couple of feet and confirmed the brakes worked well.
After shutdown, I did find an oil leak. I traced it back to a loose oil hose fitting, and was able to torque it up and resolve the issue. It was one of the oil lines connecting the oil cooler, and I must have just plain forgotten to torque it. I added torque seal and checked all the other fluid fittings on and around the engine while I was at it.
The engine ran well. I let it idle for several minutes until the oil temperature reached 100 degrees F, then performed a run-up by slowly increasing RPM to 1800. That’s where I checked the mags, cycled the prop, and double-checked the alternator output. I collected the engine data from the G3X system and uploaded into an analysis tool on my laptop. There was nothing that looked out of the ordinary to my untrained-eye.
First start!First start from inside the cockpitCHT dataEGT dataFuel flow dataOil temp dataRPM data
Today I wired the Earth-X battery sense wire into the G3X system and configured an alert to trip if the sense wire goes active. The sense wire will go to ground if there’s a fault with the battery management system – if the voltage drops, or if the cells become unbalanced. The alert displays “Main Battery” in red font if there’s an issue.
There GEA-24 supports up to 4 discrete inputs, and they were all already configured (canopy unlatched, stall warning, smoke on, and… something else I forget right now). Since I not installing the stall warning, I decided to repurpose that for the battery alert. I used a spare wire I had run through the firewall and ran it to the GEA-24. The stall warning wire is still there, and I labeled it and insulated the ends. It runs from behind the MFD to the left wing’s small access panel ahead of the main spar.
Today I worked on calibrating the fuel tank senders. This involves booting the G3X system into Calibration mode, and then running the fuel calibration steps while filling the fuel tanks a few gallons at a time.
I built a stand to place under the tail wheel to keep the airplane in level flight attitude. Because it’s a tail wheel airplane, the fuel level reads differently between level flight and when operating on the ground with the tail down. The stand worked well for getting the plane into level flight attitude.
The first step is to perform the fuel calibration in level flight, then the process is repeated in the ground attitude. The system simply reads a set of datapoints and then computes a curve to interpolate all the in-between values. It’s up to the operator to decide what values to use, and how many points.
I coordinated with the fuel truck driver, and he helped by pumping in specific quantities of fuel while I entered the data points in the cockpit. We started with the right tank, filling in 3 gallon increments, and one final 1 gallon increment to the full 25 gallons. Everything went well, and the process was painless.
When we switched to the left tank, the sender value being read by the G3X was stuck on the same value, even as fuel was added. We stopped the process and I started troubleshooting. I quickly found the sender wasn’t grounding correctly. I was able to fix this by slightly torquing one of the sender screws.
I called the gas truck driver again, and we attempted to complete the process. Unfortunately the sender reading didn’t change when we added fuel, so we stopped again.
At that point I switched focus back to the right tank. I took the tail stand out, and set the tail down on the ground. I was startled to hear a splashing sound, and quickly realized the excess fuel was pouring out the right tank vent. I quickly placed a gas can under the vent and caught some of the overflow. Note: the tail low ground position means a practical limit of fuel capacity of about 24.75 gallons.
Rather than draining the fuel, I realized that I could calibrate the fuel level as I remove fuel, not just adding fuel. I decided to siphon gas out into canisters, as I could transfer the fuel much faster than via the vent opening. It also allowed me to control how much I was removing, which made the calibration process for the second time (the ground configuration) super easy. The gas cans were all 5 gallons, so I set points at 5 gallon increments. I left 5 gallons in the tank because I ran out of gas canisters, and the zero point reading will be the same as the in-flight value.
With that done I had run out of time, so will have to come back to troubleshoot the left tank sender.
Today I tested the fuel system for the first time, running the electrical fuel pump and testing fuel flow. Everything worked well, and I was happy to see no obvious leaks.
I started by disconnecting the fuel line leading into the fuel servo, and redirected it into a large measuring jug. Then I poured approximately three gallons of gas into the left fuel tank, and turned the fuel selector to the left tank. I wanted to make sure I had fuel in the fuel pump before starting it the first time, so I used the shop vac and some clear tubing to prime the system – drawing fuel through the fuel lines, all the way to the fuel servo.
I powered up the panel and switched on the pump, and was happy to see fuel flowing into the jug. I ran the pump for a minute and measured 92 fluid ounces pumped, which translates to 43.125 gallons per hour. I repeated the test on the right tank, measuring 94 ounces pumped, about 44 gallons per hour. During the second test the fuel flow gauge was measuring 43.7 gallons per hour, which is very close to the amount I measured.
I repeated the test in a level flight attitude, and then tested for unusable fuel. With the plane in a level attitude, I pumped all the fuel out of the tank, and then drained the leftover fuel from the fuel tank drain port. I measured 6 ounces of unusable fuel per tank, for a total of 12 ounces.
Note the fuel flow of 43.8 G/H. Fuel PSI is low at just 1.5, but I expect that is because the fuel line is simply draining into a jugDisconnected fuel line pouring fuel into a jugLeft tank test results after a minute of pumping fuelRight tank fuel flow testTotal unusable fuel left, just 12 ounces
Today I added a couple more blast tubes to cool the battery, and the voltage regulator on the back of the alternator. I also reconfigured the fuel lines in the engine compartment and moved the fuel flow sensor to the engine mount.
The Earth-X battery gets a lot hotter than the standard battery, and benefits from a blast tube to direct cool air at the battery while in flight. I used a length of 1 inch SCAT tube, some hose clamps, and an aluminum flange to make the blast tube. I used a step drill to upsize the hole to the right size. I also added a lock washer on the back side of the flange, along with RTV to make it secure to the baffling.
I had bought a 3D printed fitting for the back of the alternator that directs air at the voltage regulator, but never installed it. Today I mounted it and hooked it up with a blast tube from the air inlet ramp on the right side of the engine compartment. Lots of RTV, another lock washer, hose clamps, and it was done.
Then I moved onto the fuel lines. I had previously loosened some of the fuel lines to test that my new routing would work. By switching the fuel lines either side of the fuel flow sensor, the sensor moves aft several inches, placing it right above one of the engine mount struts. Today I removed some of the old fittings, added new ones to streamline to fuel line routing, and mounted the red cube to the engine mount. It’ll be much more stable, and cooler, in this location, which should hopefully make for a longer useful life.
Battery blast tube, forward side of bafflingBack side of the baffling showing the new blast tube Blast tube directed at the battery. The safety wire is holding the tube steady while a blob of RTV driesVoltage regulator blast tube entry on the air rampUnder side of the blast tube flange on the air rampThe 3-D printed fitting for the back of the alternator. Note, this is high-temp material with Carbon fiber blended inThe new blast tube The new location for the Red Cube (fuel flow sensor)The fuel lines aft of the fuel flow sensor. I removed the 45 degree fitting from the red cube inlet, as this is not ideal. The straight fitting I replaced it with will allow for a smoother flow of fuel, and is actually the recommended configuration by the manufacturer. A single adel clamp is holding the sensor.
Based on some advice from a few other builders, I relocated the “red cube” (the fuel flow meter) from the Vans recommended position to the engine mount. The Vans location is above the exhaust, and other builders have complained of excessive heat and vibration causing the units to fail. The fix is easy – switch the fuel lines – and the cube is perfectly positioned above the engine mount. By switching the fuel lines, I mean the line from the fuel pump to the red cube is switched with the line from the red cube to the fuel controller.
My mount is not the most robust, but I think it’s sufficient. A single bolt holds the unit onto a cushion clamp that holds the engine mount. If the clamp ever failed, the fuel lines themselves would keep it suspended. However, it would probably cause some damage in the process. A better solution might be designing a bracket that has two clamps, which is something I’ll look at down the road.
Today Donnie helped me with a weight and balance. The airplane weighed in at 1257 pounds, which is right about what I expected. The center of gravity is a very important datapoint too, and after crunching some numbers the empty CG was 81.54 inches aft of the datum. The datum is a point 72 inches forward of the wing leading edge.
The process started by loading up the airplane with all the panels, screws, interior panels, carpet, literally everything that will be part of the airplane. I didn’t bother installing everything, in some cases I just placed the parts in their correct position.
With that done, the next task was to level the airplane. The weight needs to be calculated on level flight condition, so I needed to raise the tail up about 3 feet. To facilitate this I brought my standing desk from home, and was able to easily raise and lower the desk to find level. Using a desk instead of a saw horse allowed room for the scale to sit under the tailwheel. A digital level on the canopy rail was sufficient to identify level.
With that done, I moved the desk out of the way and Donnie set up the scales – one under each wheel with a WiFi controller giving a digital readout for each scale. We then rolled the mains up onto the scales. Then we raised the tail and positioned the desk and scale under the tailwheel. With that done we wrote down the weights, and the weighing exercise was complete.
To get an accurate arm for each scale, I used a plumb bob to make a mark on the floor identifying the wing leading edge. Then I carefully measured the distance to the center of each wheel. The plum bob came in handle at the tail too so I could measure the distance while the tail wheel was still elevated.
With those calculations done, I created a spreadsheet to plot W&B and played around with a few scenarios.
Configuring the standing desk to attain level flight attitudeThe numbers. Interestingly, the right main wheel is a quarter inch aft of the left main wheel, relative to the wing leading edge.
