Flight Test: Card 2

As previously stated, there are five test cards to be completed in order to sign the airplane out of Phase I Flight Test.  The EAA has published a book on flight test with a far more elaborate and detailed set of test cards, but I'm choosing to use the test cards provided by Van's in the Production Acceptance Procedure (PAP).  This is a reasonable approach considering my airplane is licensed Experimental Light Sport (E-LSA), rather than the more permissive Experimental Amateur Built (E-AB).  E-LSA aircraft must be built exactly like the ASTM conforming prototype, so it's reasonable to assume they will fly just like the prototype.  Oddly, and happily, after the FAA issues the airworthiness certificate, a builder can change anything he or she wants on the airplane as long as it stays within the Light Sport rules which were recently totally revised (and improved immensely) in the new MOSAIC ruling.

Test card #2 involves measuring takeoff distance, rate of climb, 1-g stall speeds at various flap settings (again), and stall speeds in a 30-degree turn both power off and power on.  I also added some flow visualization studies using yarn tufts taped to the upper surface of the wing and the fuselage adjacent to the wing root.

The Spousal Unit (my brainy wife Karen -- Dr. KTH to her engineering students) will show the video to her students when she teaches about boundary layer separation, hence the sign in the window.

For takeoff distance, I used knowledge of the dimensions of standard runway markings to make a good estimate.  At approximately 7000 foot density altitude with little wind I'm seeing, on average, about 950 feet with 1/2 flaps and using Van's recommended technique: slightly raising the nose as early in the takeoff run as possible and letting the plane fly itself off rather than rotating at a particular speed.  The POH says I should see about 1300 feet at about 70 F, so I'm happy with this number.  My gut feeling is that I could fly it off shorter using a positive rotation at lower speed, but someone at Van's recently re-tested this and says no.

For rate of climb, amazingly, I've been seeing over 1100 feet/minute with no flaps at 75 kts with 7900 feet density altitude.  The POH says I should see about 400 fpm at that altitude!  All this should lead me to believe I've pitched the prop too fine, more toward a climb rather than cruise prop.  The cruise speeds I'm seeing say otherwise.

Cruise speed runs at 5500 rpm and 7500 feet density altitude are also measured on this card.  The book says I should see 114 kts with wheel pants.  I'm seeing 114 kts all right, but without wheel pants or any other form of aerodynamic clean-up, and this was at 8750 feet density altitude (engine power goes down with increasing altitude).  My wheel pants and nose wheel strut fairing are sitting in my hangar.  There are also many other areas on the airplane begging for aero refinement. I've often said that I think Van's left this airplane draggy in order to stay under the old Light Sport limit of 120 kts.

Another part of this test card requires setting the airplane up in slow flight at 55 kts and whatever power is needed to do this.  For me, this was 3480 rpm.  With the power left here, the airplane is stalled in level flight and in a 30 degree turn left and right with flaps zero, half and full, nine data points in all.  I do these strictly on the gages.  The flight director and slip-skid ball are superimposed on the synthetic vision, so I'm seeing everything I need, paying special attention to keep the ball centered.

With wings level and flaps zero, half and full I saw 37, 36 and 36 kts indicated airspeed, respectively.  With turning stalls, I saw little difference left and right (it's hard to hold exactly 30 degrees).  For the turning stalls for flaps zero, half and full I saw 40, 39 and 39 kts indicated airspeed.  This agrees reasonably well with theory.  Theoretically, stall speed should increase as 1/(sqrt(cos θ)), where θ is the bank angle.

Stall video can be seen here.

For the flow visualization, I wanted to get video of the initial flow separation at the wing root before the separation region had a chance to propagate out the wing in the spanwise direction.  I was surprised to see the large amount of separation on the fuselage adjacent to the wing root. To minimize this and the attendant interference drag, the max thickness of the fuselage should ideally be near the trailing edge of the wing rather than the leading edge as it is on the -12.  This airplane needs a wing root fairing -- hard to do with the full-span flaperons.  This explains the substantial pre-stall buffet prior to the actual stall break.

All flight-test videos and other videos documenting flights I take after I sign the airplane out of Phase I will appear at YouTube.com/@CC-Flyer.  CC stands for Carolinas and Colorado.  My dream is to fly this airplane back to my ancestral home in the Carolinas before my flying days are over.

Wingtip camera mount

 From the start of Phase 1 flight test I figured it would be useful and fun to make videos of the flights.  For example, one of the items on the third (of five) test cards says to station a ground observer by the side of the runway 500 feet from the departure end whose job it is to mark the point of liftoff.  This takeoff distance could then be measured and compared to the POH values.  One use of this number would be to make a judgement on prop pitch, too coarse if the number is greater than book value.  My chances of convincing the Spousal Unit to do this being zero, I decided to closely estimate the numbers using the standard runway markings visible in the videos.  All paved, public-use runways are supposed to conform to FAA specs for dimensions and placement ot the marks, and example being the centerline stripes: each stripe is 120 feet long with the space between stripes being 80 feet.

As shown in this photo, markings at KLMO are sparse compared to many other airports.  Runway 29 has only the "piano keys" at the end, the runway number, and the 1000 foot marker (two parallel stripes).  These, along with the aforementioned centerline, have specs.  I also checked the measurements using Google Earth.  The centerline is hard to see in the pic but is quite visible in my videos.

For cameras, I have the ever-present GoPro mounted behind me looking forward, and an Insta360 X5 on the wingtip.  As I'll show, the wingtip mount is easily removable and required no  modification of the wingtip itself.

The wings on the RV-12, as documented previously, are made to be quickly removable (Van's claims 5 minutes) in a manner similar to gliders.  The thought was, people could trailer the airplane to the airport, fly, then trailer it home avoiding having a hangar.  I've never heard of anyone doing this.  To facilitate this wing removal, a hand hold is built into each wingtip, providing a perfect spot for a removable camera mount.

My idea was to make a plate above and below the handhold, securely sandwiching the hole (I love to verb a noun) with camera mount attached.  And yes, before you say it, this could ideally be done with a 3-d printer.  My sources couldn't print with a material I'd trust for this application, however, so I ended up buying some Delrin sheets from Amazon.

The inner plate fits snugly inside the hole top and bottom with larger, outer plates on the outside.  Delrin is quite stiff and strong and great for this application.  I riveted the plates together since Delrin is difficult to bond.  I used the tip of a large bit to make relief divots in the surface of the Delrin which contacts the aluminum to clear the protruding rivet heads.

I secured the two sandwich halves together with two 1/4-20 stainless bolts and ran a middle 1/4-20 bolt up from the bottom to attach to the standard female tripod mount on the camera.  Of course, this tripod mount is nowhere near enough to secure the camera in a 120-knot wind, so I purchased a cage for the camera (Amazear metal protective case on Amazon) which has two additional female 1/4-20 holes on the side.  I made the bracket shown to attach to these.  Conveniently, the bottom surface of the wingtip is at a 30 degree angle to the top surface, allowing use of standard aluminum beveled washers (also from Amazon), shown in the pic.

The cage is ideal and seems very secure.

Video here.

Going forward, my flight test videos can be found at YouTube.com/@CC-Flyer.  The CC in the handle stands for Carolina and Colorado.

Flight test: Card 1

 Since more than 830 RV-12s are flying, quite a lot of knowledge has been accumulated regarding what to expect during flight test.  Van's provides at the end of the PAP (Production Acceptance Procedure) five flight test cards to accomplish Phase I.  It's a task-based approach rather than hours-based and I'm pretty much following what's laid out in the PAP.  It's done just like the rest of the PAP, that being check boxes followed by a description of the task (83 check boxes for the first test card alone).  Some tasks are simply yes/no, OK/not-OK things, others require numbers to be recorded.  The data logging system built into the Dynon HDX helps tremendously.  The system records all flight and engine parameters many times per second which are then downloaded to a thumb drive after each flight.  

All flight testing is set up to be done solo at a specific total weight.  Accurately observing and recording numbers would be difficult while also flying the airplane.  Thank you Dynon for providing this terrific data logging tool.  The FAA now allows, as discussed in the previous post, an additional qualified pilot to assist with Phase I, but so far I'm feeling the urge to keep doing it solo.  For some of the tasks I may want a helper, an example being calibration of the AOA gage which requires pushing a button on the EFIS at exactly the stall break while keeping the wings level with rudder.  We'll see.

Before getting into the actual flight test, here are a couple of additional thoughts about the first flight.  I've done eight flights so far (about 11 hours on the Hobbs) and I've gained some perspective.  Here's what I felt I had just done immediately after the first flight:

That's actually Chuck Yeager in a rocket-assisted F-104.  Here's what other people probably thought I had done on my first flight:

The first of the five test cards checks to see that all the sensors, instruments and systems are functioning properly.  Basic flying qualities are evaluated including trim and straight-ahead 1g stalls.  Frankly, considering the complexity of the systems, I'm surprised that everything worked perfectly and the airplane flies so damn well.

After reading the write-ups other builders have done detailing their flight tests, I was fully expecting to have a "heavy" wing as most people do to varying degrees. That's a tendency to roll one way or another with the ball centered and hands off the stick, requiring corrective stick force to counteract the rolling tendency.  No heavy wing.  The airplane also stalls straight ahead with no tendency to fall off on a wing.

Power-off 1g stalls were done with no flaps, half flaps and full flaps.  Because of the surprising numbers I am seeing in the cockpit (confirmed by the data logs) I repeated the stalls in each configuration multiple times on multiple flights.  According to a table in Section 5 of the POH, indicated airspeed and calibrated airspeed (corrected for instrument error) are the same.  Based on my real-world experience, I find this hard to believe.  Theoretically, a Pitot tube would produce about 3.5% error at the AOA seen at stall (~15 degrees) but I've seen larger errors in the lab.  Van's does bevel the end of the Pitot in order to mitigate this, so maybe it's true.  I just had a discussion with a Van's techs support guy at Oshkosh, and he agreed with me and said he'd check.  Strangely, just under the table in the POH showing no difference between indicated and calibrated airspeeds, the very next table shows a difference of about 4 knots between indicated and calibrated airspeeds but with indicated being higher than calibrated for stall speeds in all three flap configurations at a gross weight of 1050 lbf!  This would make my numbers even more amazing. Solo with full fuel, i'm right at that weight.

Anyway, even with a few knots error I'm surprised with the numbers I'm seeing.  Prior to flying my airplane, I had done stalls in other RV-12s, but it was always with a passenger.  For my airplane, built light (735 lbf empty weight), a passenger can easily be 25% of that, having a substantial effect on performance.

Solo, with about 110 lbf of fuel, I'm repeatedly seeing an indicated stall speed of 40 knots with no flaps and 35 knots (!) with half flaps.  With full flaps the number is a bit higher, probably because the nose is bobbing up and down and never reaching steady state.  The static ports are on the sides of the fuselage a few feet behind the aft edge of the canopy and may be in the turbulent wake from the separated flow at the wing root  

Landing, full flaps do add noticeable drag.  Half flaps are apparently magic in this airplane and that's the way I land unless I'm high on final -- something I prefer, slipping off any excess altitude.  This airplane will sink like a Cherokee (what I learned to fly in 57 years ago!) when slipped with full flaps.

The pic shows me sharing the runway with a bird on short final.  My hangar is at Vance Brand Airport (KLMO) in Longmont Colorado, elevation 5040 feet msl.  One of the tasks on the first card involved pushing the nose over until I had 120 knots indicated.  Starting at our field elevation and climbing a few thousand feet to do the tests, my density altitude has typically been over 10,000 feet.  Before doing this, I wondered how close I'd be to Vne (never exceed speed, no doubt based on flutter margins).  Many pilots I've known think Vne is indicated airspeed.  That's a good way to become no longer counted among the living.  Vne is true airspeed, 136 knots on an RV-12.  Fortunately, my Dynon HDX displays true airpseed (along with density altitude) so I kept a close eye on it as I did this task.  As I slowly approached the prescribed speed, at an indicated speed of 119 knots I bagan releasing forward force on the stick.  The computer screamed in my ear "Overspeed, Overspeed."  The data log showed I had reached a true airspeed of 136.1 knots.  I'd like to never hear that again.

Oshkosh 2026  I just returned from my 37th pilgrimage to aviation Mecca.  Although there was a record crowd (around 3/4 million people) it was a bit atypical.  As usual, I got there on Sunday (usually one of the best days for watching arrivals) but smoke from Canadian wildfires kept the airplane traffic low until Wednesday.  It was still, as always, a sensory overload.  Our NASA administrator, Jared Isaacman, flew his personal Mig 29 in.

His personal T-38 was parked beside it.  Looking at his accomplishments, I'd say he's an amazing guy.

First flight

 Incredibly, I started this build at my home on beautiful Badin Lake east of Charlotte, NC in October of 2011.  This bird flew the nest at my home airport in Longmont, CO on May 12th 2026, so I'm hereby claiming the record for the longest build time for an RV-12: 14.5 years.  Time flies.

First flights of aircraft licensed in the experimental category have historically had a comparatively high accident rate, and it's easy to see why.  In addition to physically flying the airplane, a solo pilot is keeping a sharp eye on the engine instruments (oil temp, oil pressure, fuel pressure. left and right cylinder head temps, left and right exhaust gas temps) and on the flight instruments.  After all, this airplane has never been in this situation before, pitched up at this angle at full power for this long.  Did I do all the wiring correctly?  Did I crimp all the fittings on wire ends well?  Did I adequately tighten all the fittings on oil, gasoline and water (yes water) lines?  It's a mental load which can distract from most important task: flying the airplane.

A few years ago the FAA actually did a wise thing by instituting the "additional pilot" program, whereby a second, qualified pilot with experience in that make and model airplane can accompany the builder pilot on the first flight, sharing the work load.  From what I've read, the first-flight fatal accident rate is now zero since the program was started.  My fearless friend and CFI Bill Snodgrass accompanied me.  Bill has over 1600 hours in RV-12s and is intimately familiar with how they should look, feel, sound and smell (and maybe taste).  I had planned to fly his RV-12 several times in the days leading up to the first flight in mine, but ended up not having flown at all for four months prior to the first flight.  The rust on my piloting skills was evident.  I had flown his plane for 20 or so hours in the previous year, so I was not completely uncalibrated to the RV-12, but I should have flown just prior to my first flight.

There are two schools of thought on first flights.  Some say an actual test pilot should be used because the builder is too heavily invested (emotionally and financially) in the airplane and might try to protect the airplane rather than save himself in the event of an emergency.  To me that would be akin to hiring a professional to stand in for you on your wedding night.

The pic above shows the airplane before the flight.  I know it's considered bad luck to photograph a flight before departure.  I figured if I made a smoking hole in the ground the NTSB could look at this pic and see that all the pieces were there.

The actual flight went well, no heavy wing, all instruments showing what I expeccted.  Rather than start on the first flight-test card, we decided to shoot four landings to a full stop, then pull the top cowl to look for leaks (none found).

Here's the after pic with the requisite RV grin.  It's been two days now and I think I'm still smiling.

I had planned to start the formal flight test yesterday but the wind didn't cooperate.  Phase I flight test will be entirely task based rather than time based as was typical in the past.

The Spousal Unit (my beautiful and talented wife, Karen) had given me strict instructions to not tell her when the first flight was going to occur in order to minimize her worry.  This was hard to accomplish without outright lying.  The day of the flight it went like this.  Her: "You're not going to fly today are you?"  Me: "Actually, I'm thinking about flying tomorrow."  That wasn't a lie.  I was, indeed, thinking about flying tomorrow but planning to fly today, also.

Off topic: Weird Colorado weather

We've had a dry year here on the Front Range, with snow totals and mountain snow pack about 1/3 of normal.  However, a few days ago we had eleven inches of snow over a two-day period, which then melted entirely within 24 hours.  Snow in May seems strange to this southern boy.

Airworthiness Certificate

 The final hurdle in this process -- a process which took 15 years and certainly involved blood (several #30 and #40 holes in various fingers and other blood-letting feats), gallons of sweat, and a fair amount of money -- is the Airworthiness Certificate.  Getting this certificate involves, among other things, an inspection by the FAA itself or by a person designated by them called a Designated Airworthiness Representative (DAR).  If the people from the FAA do it, it's free but apparently impossible to schedule.  If a DAR does it, it's expensive but can be done at your convenience.  I had an excellent DAR, Brad Roon, who charged $1000.  Horror stories abound from other builders I know who have gone through the process, so I consider myself lucky.

For the inspection, all cowls and inspection plates must be removed and certain bits of paperwork must be available. The required on-hand paperwork includes aircraft registration, weight and balance calculations, a complete list of service bulletins with dates showing completion, and a few other documents.  Brad provided a list of everything he wanted to see.  He did a very thorough physical inspection of the entire airplane, which made me a bit nervous, but was what I wanted.  I was confident that all was good because I had a superb EAA technical counselor (Dan Berry) doing periodic inspections as everything went together.  Dan was picky (to put it mildly) but, again, that's what I wanted.  After all, this machine, which started out as a collection of parts, will soon carry me into the sky.

Prior to the DAR visit, the on-line AWC application must be completed on the FAA website.  This is an unbelievable PITA, what you would expect, I suppose, from a government website.  Many documents must be uploaded there, including the Maintenance Manual, the Flight Training Supplement and the Pilot Operating Handbook.  Each of these is over 100 pages, and they want the actual documents, not a link to the documents.  I'll bet you $1000 to a dime that no one at the FAA looked at these other than to see that they were there.

One of the things you submit is a request for a test area, a box specified by lat-long coordinates, within which you're restricted for the Phase I flight test.  I photocopied part of a sectional chart for that and included it in my application, with my requested test area marked in red.

What the FAA granted me was much more generous: from Boulder north to Wyoming and from the Rocky mountains east beyond Fort Morgan.  Brad asked me how much flight time I needed for phase one.  For an E-LSA it can be as little as 5 hours rather than the usual 40 hours for EAB.  I told him 10 hours.  No doube it'll take longer.  Phase I flight test as it appears in the PAP provided by Van's is task based rather than hour based.  I'll describe it as it happens.

Another thing that is issued in addition to the Airworthiness Certificate is a document listing the Operating Limitations for the aircraft (OpLims).  These are requested by the DAR and approved by the FAA.  They give conditions under which the airplane must operate: pilot certificates required, what can and can't be done with the airplane, etc.  Interestingly, even though aerobatics are prohibited in the RV-12 by Van's, the OpLims say only that any aerobatics to be done must be demonstrated during phase one flight test.  Hmmm. Also, a lot of verbiage is devoted to talk about not having external things on the aircraft that can be jettisoned during flight.  That lets out any bombing runs.

So with Airworthiness Certificate in hand, all that's left is to re-install all the cowls, covers and plates, then go fly.

Off topic (sort of)

Here's what happens when you're old and you've hung around airplanes since grade school.  Hard to believe it's been 55 years since I got my license.  Time flies.

PAP Section G6-1: Dynamically balancing the propeller

After syncing the carbs I was less than pleased with the amount of vibration present, especially at low RPM. This being the case, I decided to deviate from my previous plan and do the DynaVibe deal before first flight.  I had planned to put this off until after Phase 1 flight test, as many people do, and fly the plane over to Akron, CO to a shop where others I know had done this with satisfactory results. 

The Mothership recommends doing it before first flight, however.  A member of my local EAA chapter 648, Rick Hall, owns the most expensive version of the DynaVibe balancer and offers it, along with his expertise, free of charge to chapter members (he did accept a bottle of Beefeater gin).

Using this device requires attaching three things to the engine: an accelerometer, preferably as close to the engine centerline and as near the prop as possible (shown here nearest the prop hub with the single black wire), a piece of special reflective tape to the back of one prop blade, and a laser pulse counter (the yellow device to the right).  The laser and reflective tape measure prop RPM.  The tach in the airplane shows engine RPM.

The engine is run at low RPM (2500 in the case of a Rotax 912), then up to cruise power (5200 with the airplane static, depending on how you pitched the prop) where the engine will live most of its life.  The DynaVibe then gives an angular location for the temporary stick-on balance weight and the mass of the weight.  The DynaVibe assumes the weight will be on the perimeter of the rear bulkhead for the spinner.  If for some reason the weight must be attached at a different radial location, a simple calculation allows the new mass value to be determined (r1 * m1 = r2 * m2).

I tied the tail to Rick's pickup truck since I'd had the airplane jump the chocks on a previous full-power run-up.  When that happened I feared I had set the prop pitch too fine, thereby producing too much static thrust.  My fears were allayed when I saw the correct static RPM (5200) at rull power.  On one full-power run I did throw off one of the stick-on weights, heard it hit something over the roar of the engine, but could never find it.

This process stretched over two days (due to starting mid-afternoon) but could have been done in two hours if I had been prepared.  I was not familiar with where and how the brackets for the accelerometer and laser attached to the engine.  Turns out I needed two M8 bolts which I didn't have.  A trip to the aviation aisle at a nearby Home Depot solved that problem but wasted time.

The picture at right shows the temporary stick-on weights on the aft face of the spinner bulkhead.  It was necessary to remove the paint to keep them attached at high RPM.

Van's warns against succumbing to the temptation to attach the permanent weights to the existing screws which attach the spinner to the bulkhead.  Instead, a new hole should be drilled in the bulkhead face and bolts and washers should be used to match the weight of the temporary ones.  Being an anal engineer, I calculated the weight of the aluminum removed by the drill (0.1 grams) and ground the required metal off the washers to get the matching weight.  The required weight was 13.1 grams (yeah, I know that grams is a unit of mass, not weight, but irritating though it may be they express weight in grams in Europe where the Rotax is made).

I was initially worried that the existing nut plate or rivets would interfere with the bolt head or nut of the permanent weight, but it turned out to not be a problem.

Shown first is the aft face of the spinner bulkhead, then below it the forward face.  I used large diameter washers for the AN3 bolt to keep the stack short.  Ideally, the DynaVibe would be used again to check the final balance with the permanent weights but this wasn't done.  I'm confident it's good.

PAP Section G-6 - Maintenance manual page 12-7: Carburetor Synchronization

When I sold my Honda CB-750 motorcycle I thought I'd never have to sync carbs again (that ill-handling monster had four carbs).  At least the Rotax only has two. The idea here is to ensure that the two cylinders on each side of the engine see the same throttle opening across the rev range.  This is done by starting with the throttle arm on each carb 0.004 inches from the mechanical stop using a feeler gage, then opening each carb 1.5 turns of the adjuster (clockwise).  The lock nut on each arm is then screwed down, attaching the arms to the two cables.  The two cables merge into one, which is the throttle in the cockpit.  Idle is now set for each carb and can't be adjusted individually with the throttle-stop screws once the cables are locked to the arms.  Adjusting the throttle plate opening of one carb relative to the other must be done with the ferrule adjusters on each cable.  The following didn't dawn on me initially and I learned it the hard way: the friction lock on the throttle plunger in the cockpit must be tightened down before adjustments are made to the ferrules.  This prevents relative motion between each cable and its sheath.  The ferrule only attaches to the sheath, but the cable and sheath must move as one at each carb during ferrule adjustment.  Otherwise it'll screw up the 2.75 inch plunger travel needed in the cockpit to ensure the achievement of full throttle on each carb with the plunger full forward. Francis Miller, who administered my private pilot check ride back in 1969, told me that the origin of the expression "balls to the wall" came from this: propeller, mixture and throttle "balls" all the way to the instrument panel.  I thought it meant something else.  Also, the RV-12 only has one ball.  Returning now to the topic at hand, the Rotax setup is different from anything I've seen in that the carbs are spring-loaded wide open.  The throttle cables allow the springs to open the throttles or pull them closed, a safety feature, I guess, in the event of cable breakage.

In the pic, the ferrule adjuster can seen on the cable at left, the lock nut which secures the cable to the arm on the right, and the throttle idle stop (with orange torque seal) just to the right and below.

The carbs are now mechanically synced and must be tested over the rev range, with adjustments made to the ferrules as needed.  This is accomplished by measuring the manifold vacuum on each side and making ferrule adjustments as needed to balance them.  The vacuum readings can be obtained with two individual gages or a single differential pressure gage (the bast way).

A couple of years ago when I was insanely optimistic about when I'd have the engine running I bought a kit from Aircraft Spruce which included two vacuum gages (which I later found for $20 each at Harbor Freight) and some tubing which was supposed to facilitate hooking everything up to a Rotax.  As I recall I paid a few hundred bucks for this.  Both gages were faulty, but I found them on line under a different name, still $20 bucks.  This rig would have worked, with poor accuracy, but probably good enough.

Fortunately, I was able to borrow a CarbMate, which uses an electronic differential pressure transducer (shown below).  Adjustments are made until the light is centered.  

After some puzzling attempts at a balance I started to doubt the device (couldn't be me, right?) so I rigged up a large syringe (below) hooked up to each leg of first the analog gages and then the CarbMate, exposing both legs to the same vacuum.  Both measurement methods passed the test, each analog gage showing the same reading and the CarbMate exactly in the center of its scale.  I went with the CarbMate.

I already had the syringe on hand since I had used it to test the pitot-static system earlier (discussed in a previous post).

With renewed confidence in the measurement system, I hooked it up to the Rotax.  The intake plenums on each side of the engine are connected by a balance tube which can be disconnected, allowing the two legs of the measurement device to be attached.  Although it's a bit crowded, the pic below shows the hookup.

The two black tubes in the center with brass fittings are the hookup.  The two brass fittings with the teflon thread tape go to the two legs of the CarbMate.  The initial confusing readings were due to a vacuum leak on one of the rubber hoses, easily fixed with a hose clamp.

Once all this was sorted out the process was easy.  Balance is checked at idle and at approximately 3000 rpm, this higher rpm allowing the carbs to get off the idle enrichment circuit.  Mine was perfect at idle (1800 rpm) and at high rpm but showed a slight variation (1/2 of a light on the CarbMate) part way between the two.  I called it good, figuring the engine spends most of its life either near idle or around 5000 rpm.

More Colorado wildlife: I had a few elk in my back yard.