Kent's Long-EZ project

[Kent Ashton]

1 hour ago, Jon Matcho said:

I'm betting that last picture up there is showing upwards of $1,000 in parts.

I think I saw about $550 for the kit

-----------------

I installed a Uavionix EchoUAT and SkyFYX-Ext in the Cozy. It was a pretty easy install but after five flights I have not been able to get a good FAA Performance Report out of it.  I am seeing targets in Foreflight, Greensboro Approach told me my UAT was working, and a couple of airplanes said they see my info.  The Uavionix App shows the correct monitor data.  Very frustrating because there is no good way to troubleshoot this.  Today I found this map of ADSB towers.  https://www.dropbox.com/s/v350zbnxlelowr9/user_waypoints.kml?dl=0   It can be opened in Google Earth.  I see that my airport is between towers.  Next time I'll keep real close to a tower.

BTW, the EchoUAT can "sniff" the transponder code and simultaneously transmit its own data.  My Garmin 327 has an RS-232 output that can be wired to the EchoUAT in lieu of the "sniffer" but it took a $180 software upgrade to the Garmin.  It's always something!  😞

[Kent Ashton]

I learned something about mold wax today:  I have made cowl molds, waxed them with a carnauba wax and noticed that PVA will clump up on the mold unless it is sprayed-on in very thin coats, which is rather tedious.  The purpose of the PVA is to keep the wax off the finished part.  A chap on the Cozy list says the PVA clumps because the carnauba-based wax contains silicone.  He says to use a "green" wax like this Partall   https://s3.amazonaws.com/cdn.fibreglast.com/downloads/00148.pdf  Good to know!

My prop is done, painted, and I made a cover.  After I fly it a bit, I will make a report.  I worry about flying an untested prop but it is the same layup schedule I used before which held together:  BID-UNI-BID on cambered side, UNI-BID on the flat side wrapping to the cambered side.  Painted in some "Porsche Green" motorcycle paint I had left-over.  Pic 3 is the balance.  It took a plug of lead  in the hub 3/4" long X 1/2" to bring it almost to balance, then a bit if extra paint on one blade.

Edited April 20, 2019 by Kent Ashton

[macleodm3]

Good Job Kent!  Not many of us patient enough to build a prop.... way to keep that homebuilding spirit alive!

After you get done testing you new prop, you are welcome to fly up here to Wise, VA and try out my Light Speed Engineering prop...  its a 68x80 that Klaus says should work well on a strong and clean Cozy.  I am curious if I'll get to use it as my test flying prop and I'd like to see some real world results before attempting first flight.

 

[Kent Ashton]

I like how my friend cut his baffle material so it will spread but keep a tight seal.  His AN bolts are a little overkill but those baffles are NOT GOING TO COME OFF!  🙂

 

[Ron Springer]

On 4/20/2019 at 4:09 PM, Kent Ashton said:

I worry about flying an untested prop but it is the same layup schedule I used before which held together:  BID-UNI-BID on cambered side, UNI-BID on the flat side wrapping to the cambered side.

Do you only wrap at the leading edge since the trailing edge is sharp? Or do you put a small radius at the trailing edge and wrap it there too?

[Kent Ashton]

47 minutes ago, Ron Springer said:

Do you only wrap at the leading edge since the trailing edge is sharp? Or do you put a small radius at the trailing edge and wrap it there too?

I made a "flox corner" at the trailing edge.   Grind a thin saber-saw blade so it it only about 1/8" wide and pointed (pic 1).  Mark a line along the T.E. about 5/16 to 3/8" in from the T.E.  Drill several holes big enough to get the saber-saw blade started in the holes.  Cut a line down the T.E. leaving the wood attached with small bridges where you drilled the holes (pic 2)  Apply a bit of contact cement on the cambered side of the T.E. and apply a strip of peel-ply tape (pic 3). Layup the cambered-side layers.  After cure, flip the prop and chisel-off the wood and peel-ply at the T.E which is pretty easy since it s only attached in 4 or 5 places (pic 4).  Fill the edge with wet flox and layup the flat-side layers (pic 5).

I have no idea how the professional builders do the trailing edges but this is about what I did on the last prop and it worked  well.   On the other prop, I didn't cut the line with the saw and it was lot more work chiseling off the wood.  

If this prop holds together I will make a full writeup.  If not, it was nice knowing you.  🙂

[Kent Ashton]

I wanted to fly the prop today but it was way out of balance.  The engine shook noticeably and I was afraid to fly it.  I confess I did not pay much attention to checking the tracking during construction and a quick check from the prop tips to a point on the wing T.E. showed one blade tracks 1/4" differently from the other.  Also when I inserted the lead slug in the hub, I did not split the weight on each side of the hub which might be giving me a coupling imbalance.  And it could be a combination of the two.  😞

They say a slight tracking error can be fixed by  torquing the bolts differently but I probably have too much error for that.

I will do a careful rebalance paying attention to hub (coupling) balance and check the tracking at the stations.  I might be able to shave the hub a fraction to correct the tracking.

[Voidhawk9]

Thanks for the update. Sorry that it hasn't gone too well today. But we are all learning here! 🙃

[Kent Ashton]

Just reading this Mk IV crash report from 2017   http://www.kathrynsreport.com/2019/04/cozy-mark-iv-registered-to-and-operated.html

It is rather inconclusive as to why the engine quit but sanding dust was found in the filter and flow-divider (fuel injected engine).   I am not sure why a 27,000 hour pilot needed a second pilot for Phase I.  They did not have the boost pump on when the engine quit so the co-pilot was not much help there.  I think a second pilot is just a distraction: he is talking to the pilot, the pilot is talking to him.   The engine quit at 400 feet.  That's on the ragged edge of a return-to-the-field-altitude.   Given the confusion and lack of experience in the aircraft, it's not surprising they didn't try a turn back, but at 500' AGL it is pretty doable with one pilot.  

The accident with pics was discussed here  https://groups.google.com/forum/#!searchin/cozy_builders/robinson$20cozy/cozy_builders/qEY53lRORCc/P0b05eIsCwAJ

My guess is that it was ultimately due to fuel starvation of some type.  I turn my Facet pump on for all takeoffs and pattern work and  I am careful to examine the gascolator screen and Earl's filter element every year.    One time I used too much Fuel Lube on the fuel valve and found gobs of it on the gascolator screen.  It was not enough to clog the screen but fuel does not dissolve that stuff!  Other than that, I have never seen more than a spec or two of anything on the screen.  Fingers-crossed.  🙂

 

[Voidhawk9]

No fault found, no reason for the engine failure.
I was a passenger in an Aztec that had an unexplained failure in one of its Lycomings too, leading to a forced landing on a parallel runway as it occurred at low level and below redline speed. Started right up afterwards. No fault found.

[Kent Ashton]

Prop woes:  Here's where a bit of poor craftmanship leads to a world of woe but it has been very interesting puzzling this out.  My first run of the new prop showed a bad vibration.  Checking the tracking I found I had a 1/4" to 5/16" tracking error at the blade tips (see diagram #1 in the first pic).  This was just careless building--not checking it while building.  So I had the bright idea to shave a bit off the hub to bring the tips into the plane of rotation (pic 2, halfway thru the process) (also see diagram #2 in the first pic).  I only shaved off a tiny bit to move each blade 1/8" to 5/32" toward the plane of rotation.  I rebalanced the prop through the center hole using both a static balancer and hanging balancer and ran it again using a Dynavibe prop balancer.  Still had a strong vibration.  The Dynavibe showed over 1.5O IPS which is off the chart.  A good prop will be .07 or less.  We added some weight to the heavy side but it was still very unbalanced.  Tracking was within 1/16".

After a bit of head-scratching I surmised that while my hub-shave moved the tips into the plane of rotation, it also moved the center of mass of each blade OUT of the plane of rotation which was, I suppose, was causing a coupling imbalance (see diagram #3 in the first pic).   I figured I could compensate for that by rebalancing the prop in relation to the shaved side of the hub, i.e., the side against the flywheel, using a hanging balance and a bubble-level on the hub.  The bubble-level is tricky to use and probably not as precise.  The .4 oz weight of the bubble-level has to countered with washers.  (pic 3)

BTW, The height of the ferule determines the sensitivity of the balance.  If the top of the ferule is below the hub CG, it will never balance--just flops from side to side.  As you move the ferule up, it becomes less sensitive.  When almost level, notice that the line no longer passes through the centerline of the center hole but the shaved face of the hub is a right-angle to the hanging wire.

OK, feeling pretty happy of myself at this point, then I realize that the prop bolts and the crushplate will be off center about the amount that the hang-wire is off center in the centerhole (see the diagram in pic 3).  It's not much but they are heavy and are likely making a big difference in balance.

So in the next few days I will put the crushplate and prop bolts on the hub and try rebalancing again.  What I would expect to see the prop tip-over in the direction of the bolt offset. Stay tuned to see if this prop can be fixed.

[Jon Matcho]

Impressive Kent.  Much of this is just short of magic for me and I can imagine how much research and learning you've had to do if you were not an engineer already (or even if you were)!

At worst the 3rd time will be a charm and you'll be a certified prop builder at that point.  

[Kent Ashton]

1 hour ago, Jon Matcho said:

Much of this is just short of magic for me

Yeah, me too.  I read that a lot of prop builders will slightly rework a prop (repitch it) for a customer but I committed to glassing both sides out of fear it might come apart so I am pretty much committed to the pitch and twist I used.  In the Dynavibe check, I got static of about 2300 rpm which might work for a spare prop if I can fix the balance.  It may have too much pitch but you can only really tell that by flying it.

Next time I would try to precut the layers (pic 1) so there is less carving and wood-removal required.  You could build them pretty fast that way.  It is easy to cut into the blade profile while bandsawing-off the excess wood so one tends to leave too much wood to be removed by hand.   CNC?  What's that?  🙂

 

[Jon Matcho]

What type of wood did you use?  I'll just have to re-read this topic.  Interesting stuff.  I just figured out now how you used the templates as a reference for your cutting machine.  My next question are:  How did you get the stations/positions for the profile templates?  Are you working from published plans of some sort?

[Kent Ashton]

Maple.  I started from a desired pitch at the 75% station and a blade length.  I will make a writeup on the process soon.

[Kent Ashton]

Lest you think that a hanging prop balancer is some sort of imaginary device conceived by an addled brain, I offer exhibit 1

https://www.mcfarlaneaviation.com/section/news/propeller-balancing-bushing-kit/

Ha, Ha, Ha, Ha, Ha

[Jon Matcho]

The very next email in my Inbox after this post's notification was from EAA about a discount for a prop balancer!

[Kent Ashton]

I saw this pic (#1) on a FB page and edited it below.  It is an old one from newsletters and the arrangement doesn't work but it got several "likes".  There is a fair amount of bad info on FB.  

The attach point for the master cylinder must be lower on the pedal to achieve sufficient leverage.  Matco says the leverage must be at least 2.5:1.  This pic caused me some grief in the past.  Pic 2 shows a laydown pedal arrangement that works pretty well but it could use a little more leverage.  

[Kent Ashton]

More bad ideas from the world of FB:   This chap (pic 1) is planning to put these intake holes in his NACA inlet.  The NACA flow, which is already swirling at the sides due to the design of the scoop, and turbulent, will flow toward that 3-hole arrangement.  He is expecting to get, say, 24 sq inches of airflow into three holes of about 9.5 sq inches total.  As a result, most of it will spill-over out of the NACA and create turbulence outside and behind the NACA.  That spilled air, slowing down as it encounters the 3-hole air-dam and then having to accelerate back to freestream velocity outside the fuselage, uses energy to slow it down and energy to speed it up again.  That's drag.

In addition, the molecules, which have momentum (air is heavy!)  are going to impact the flat surfaces surrounding the intake holes and cause a backpressure to incoming flow which will increase the spill-out tendency.  Then as the molecules bounce off the flat surfaces and try to enter those small holes, they will tumble (arrows added).  They will not enter his holes in an orderly non-turbulent stream, therefore, fewer of those little buggers will crowd through the holes as they bump and jostle each other.  The effect will be to reduce  his 9.5 sq inches to maybe 5-7 sq inches.

In addition, the choke-like design of his holes is attempting to accelerate air through the holes.  That will also cause backpressure and spillout.  And normally you would want to decelerate air entering a plenum to raise its pressure.

On my Long-EZ project, I had downdraft cooling so I used a smaller-size NACA on the bottom for carb and oil cooler air using two large square holes (pic 2).  These holes turned out to be larger than needed.  I closed them down with some aluminum as a test (pic 3) and eventually raised the floor of the NACA (i.e., made the NACA shallower) to make the holes smaller.    For this fellow, I imagine intake holes like pic 4 would work much better.

[Kent Ashton]

Dang!  Correcting the picture I posted above

.

[Kent Ashton]

Saw this Varieze scoop arrangement on a FB page.  I can't repost the entire discussion but the OP said

Quote

it cools really well and is turning in decent speed. We left the cowl alone until we could get a baseline on the cooling. Few more flights and we will have figured out the best way to cool it. We too have a cowl mold and per customer's request we built the top one and kept the existing bottom cowl. without wheel pants we are seeing 170ktas at 62% power.

Yep, 170 KTAS is a good speed but it's hard to say what's the reason without a lot more info.  It seems logical that more square inches of inlet is going to improve cooling but if it uses more air to do it, that would seem to increase cooling drag.  A mass of air enters the plenum, slows down through the fins, then is returned to the freestream in the low pressure behind the engine.  All that slowing and accelerating of the air mass uses energy.   Perhaps the greater air mass here does not slow down so much.  Perhaps the aircraft has more HP than usual,  or was run at full rich down low.  Heck, my airplane will go pretty fast if I run at full rich down low.  That doesn't mean much.

 I am not knocking the cooling idea, just saying that I don't understand it.

[Kent Ashton]

Been going backward on my project prop.  To review, the tips did not track in-plane--I had a 5/16" to 3/8" error at the tips.  I shaved the face of the hub to correct the tracking error.  My first try with the Dynavibe showed about 1.96 IPS vibration.  Real bad.  Then I had the idea discussed here https://www.canardzone.com/forums/topic/18661-kents-long-ez-project/?do=findComment&comment=63426   that shaving the hub might put the the crush-plate and mounting bolts a slight angle so I rebalanced the prop on my static and hanging balancer with the prop-bolts and crush plate installed on the prop.

I ran the prop again with the Dynavibe and it was even worse:  2.33 to 1.86 IPS at idle.  I was afraid to run it up it up with so much vibration.  Adding various bolts to the flywheel and washers under the prop bolts began to bring it down so that I was able to see .32 IPS at idle and .11 IPS at 2051 RPM.  (pic)   The "Good" to "Fair" range is .07 to .15.   It felt a lot smoother but all this weight was a puzzle.  After all, it is the prop that is screwed up; adding weight at the flywheel, some 12 inches forward does not seem to be the ideal solution.  One would think the weight change should be made at the prop.

(BTW, I put my trusty Performance Prop on for a check and it showed .09 IPS ("good" is .07 IPS).  I modified this prop some years ago and have only balanced it with the static and hanging balancer.)

So I have abandoned my idea about crush-plate and prop bolts being off center (although they well could be) and rebalanced it through the center hole like a standard prop.  The centerhole is not exactly on center anymore but rebalancing caused me to drill out a bit of the lead weight I had put in the hub from the last balance.  That lead was on the heavy side, according to the Dynavibe, so drilling it out should help.)

Next test, I will try to add weight at the prop bolts using large washers or maybe washers with extra welded weight on them.   We know that a one-blade prop can be balanced and run satisfactorily so I believe if I get the weights right, this prop will be usable.

 

[Kent Ashton]

By experiment, I ended up with a bunch of weights shown in the pic above to balance my prop and pondered how to distill that down to one or two equivalent weights.  I want to put my weights in or on the hub.  Here's an idea for using a CAD program to do that:  To illustrate, at the prop hub in the pic above I have two extra washers weighing 1.65 oz at 180 degree azimuth (from the Dynavibe reflector) on a 2.375" radius.   Multiply 1.65 oz. X 2.375" to get 3.92 oz-inches at azimuth 180.  Plot that vector on the CAD drawing.  Do the same for all the weights, connecting all the vectors.  You will see in my drawing below that the sum of all my weights came to 13.01" oz-in. at azimuth 207.71 deg.

Any weight you substitute that gives a 13.01" oz-in at 207.71 deg result will be equivalent.  Putting 4.33 oz. of weight in the hub on a 3" radius at 207.71 degrees is the same (4.337 X 3 = 13.01) .  Or if you wanted to place weights under two prop bolts, draw the azimuth of the two prop bolts, one azimuth from the origin and the reciprocal of the other azimuth from the 13.01@207 point and measure their length where they cross.  Prop bolts are on a 2.375" radius so divide the length (oz.-inches) of each vector by 2.375" to get the weight to place at each prop bolt.  Neat huh!   There is probably a way to do this with trig but the CAD program seemed like a good way.

I statically rebalanced this prop and started over with a new bunch of weights, trying to keep them at the hub.  (pic 2)  After much fiddling today, adding and moving weights, I could get the Dynavibe reading down to .13 IPS which is in the Fair-to-Good range and a big improvement over the 2.00 IPS I had initially.  My prop bolts were not long enough to accept a stack of washers so I made the angle seen at the bottom which is heavy by itself and added one more 1/4" bolt on it.  This arrangement of weights is about equivalent to the pic above.

You might ask "What is the radius of the angle + bolt?"   I will hang it from a string in two directions and where the strings cross, that's the CG of the part, then measure from the part-CG to the center of the prop bolt hole to find the radius.

BTW, I have Lightspeed ignition Plasma 1 trigger bolts in my flywheel that I never thought to balance.  They total about .35 oz so today I put an equivalent stick-on weight inside the flywheel opposite the trigger bolts.  I rebalanced my Performance Prop used one extra AN960-417 washer under a prop bolt and showed 04 IPS at 2490 RPM which is in the "Excellent" range.

Finally, it appears my new prop is not going to give enough RPM.  Static was only about 2300 but flying it will tell the tale.  It should certainly be usable as a spare prop if it holds together.

[Kent Ashton]

Just to clean up this weight discussion, here is the final adjustment and the visual depiction of the weight vectors.  I installed a 3 ounce rod (cut-off bolt) in the hub at a 3" radius and with four more engine runs, I added a .35 oz washer and a .75 oz, washer under prop bolts.  This gave an IPS reading on the Dynavibe as follows:

1073 RPM:   .07 IPS at 260 deg

2228 RPM:  .04 IPS at 302 deg, in the "Excellent" range

In theory I could eliminate the washers by adding a bit more weight in the bolt in the hub.  I flew the prop and it felt good.  I could get 2750 RPM at 8000 feet at peak power--an overspeed condition--but at the lean setting I usually use, it ran about 2650 rpm and did not throw a blade so it looks like it will be a good prop and I am calling this experiment a success.  🙂  I was surprised to see such a high RPM at peak power since the static seemed lower than I'm used to.

It is interesting that no amount of static balancing around the centerhole would have put this prop in balance.  It was only by trial and error with the Dynavibe that it could be worked down to a good vibration level.  I am happy to have most of the weight in the hub which minimizes coupling imbalance  For reference this is a 66" X 80" (26.87° chord angle) prop.

[Kent Ashton]

More on the ADSB install:  I showed my install here  https://www.canardzone.com/forums/topic/18661-kents-long-ez-project/?do=findComment&comment=63063     That seems to work well but you would not believe how many flights I have made piddling around with this thing and I still have one more test but I think I figured it out.    Below is my checklist and some gotchas.  Glitch #1 is changing the EchoUAT configuration in the App but forgetting to hit "UPDATE"  at the bottom.  I probably did that on several flights.  I hard-wired the Garmin 327 output to the EchoUAT thinking that would be more reliable than the Transponder Monitor (sniffer) mode.  However--glitch #2--the Garmin needed a $180 software update to output the data.  Even then, it didn't work.  That's when I determined Glitch #3: You can go into the Garmin configuration page and changed the RS232 to "REMOTE" [output] but if you fail to move the CRSR to a new position,  the Garmin does not accept the change.  This is likely why I had an  FAA performance report showing no BARO data transmitted by the Echo.  Will try one more time.  The setup file is at the bottom  if you want to download it.   Only valid for this combo of Uavionix and Garmin.

BTW I just saw a Jeopardy that said the word "glitch" was invented by John Glenn.    

ADSB SET-UP for Uavionix EchoUAT, SkyFYX-EXT and hard-wired Garmin GTX-327 [filename: ADSB Setup]

1.Connect to Wifi, see “Ping-XXX”

2. Launch Uavionix App

3. Configure as follows

Control  UAT TX Enabled

ICAO # [your # from the FAA registry]

C/S [your N-number]

Flight Pln -------

CSID Enabled

Anon.Mode Disabled

Emit.Cat Light Plane

VFR Code 1200

ADSB In Cap Both 1090/978 Mhz

Vse 30 (or 0 for testing)

A/C Len <15

A/C Wide <23

GPS AntOff 0

GPS AntOff 0

 

SetupSource App (wi-fi-stored)

Cntrl Source EFIS/PANEL (com1) or Transponder Monitor

GPS Source Ext GPS (Com2)

TrafUplink ---------

Com1 rate 9600

Com1 data 81N

Com1 Phy RS232

Com1 Prot. GTX

Com2 rate 115200

Com2 Input NMEA

 

-->DON'T FORGET TO HIT “UPDATE”

Confirm settings in MONITOR

 

EchoUAT Led indications

- Solid Red = fault

- Flash Red = transmit

- Solid Grn = powered

- Flash Grn = received

 

GARMIN 327 Setup:

[update software to V 2.12]

(RS232 OUTPUT must be set

to “REMOTE”)

w/ unit off, press FUNC + ON to

access Configuration page.

page thru with FUNC or STOP/START

use CRSR to select field

use 0-9 to change data [but not needed]

use 8 or 9 key to change field

move CRSR to accept data

 

 

ADSBSetup