The Gremlin as a Tiny Trainer

The Gremlin as a Tiny Trainer – How a 58 year old fixed wing pilot learned to fly FPV.

My Gremlin has served me well, but I thought I would share some of what I learned along the way. I am not a great FPV pilot, but I am flying under and around trees, occasionally through the branches, and doing flips and rolls now. The process only took about 50 years, plus 3 weeks….

I am a long time fixed wing pilot, though never a 3D pilot. I learned to fly quads line of site a few years back though not in a spectacular way. I also fly a lot of VTOL stuff that hovers, though usually with the aide of some sort of Autolevel. I tried FPV in a Bixler a year or so ago, and with the help of a Guardian flight stabilizer it was pretty easy, but it didn’t really grab me. Finding a suitable pair of goggles was a problem and dealing with the tangle of FPV video frequencies wasn’t much fun.

I eventually purchased the Quanum DIY FPV Goggle V2-Pro as well as a diversity receiver and some circularly polarized antennas. My intention was to use it with my glasses but being near-sighted it works best with the glasses off. I didn’t want to spend a hundreds of dollars on fancy Fat Shark style goggles or similar without knowing if they would work for me. The Quanum Goggle was a good choice as it allowed me to easily modify both the way it fit my face and the optics, and at $31.35 for the kit I wasn’t risking much. It may not be stylish, and I can’t say how the experience compares to more expensive goggles, but it works.

I tried flying in the house with an upgraded Inductrix AKA Tiny Whoop, but it wasn’t that much fun. The space was too small for flying in rate mode at my skill level, and the video quality through walls was pretty bad. Flying in AutoLevel mode was more successful, but it never grabbed me. The Inductrix could not handle much wind and I was beating it to death in the driveway, so I lost interest.

When the Gremlin came along I bought one mostly on impulse. The price was right, and the performance seemed impressive so I let the FT guys build one for me. It turns out I could have built it myself as soldering is not a problem, but the cost to let the FT guys do it for me was worth my time. I got the T.J. Williams Carbon Fiber frame so no complaints there. IMHO, some of the other frame designs block too much of the rotor disk and are vastly stronger than they need to be, especially when made from CF.

When I got it I was NOT a happy camper. Physically it was fine, but they forgot to send me the propeller mounting screws, so I could not fly it until I scrounged some from work. The part that made me frustrated and angry was the build video. It was great for all the physical build stuff, but really sucked in terms of how to deal with the computer and set it up. I eventually stumbled my way through it, but the level of information provided compared to what I needed was grossly lacking. They just assume you have all the necessary applications loaded and working properly, that you have the right drivers and cables, and they flew through the menus so fast it was hard to follow even with pause and rewind. I give them an A on the physical build and a D- on the computer setup portion.

IMHO the Gremlin frame designs are not great either. None of them provide any protection for the cloverleaf antenna. I made a few simple modifications which have served me well. I wanted to place the top plate above the cloverleaf antenna to protect it, but that meant that the CF top plate they provided was of no use as it would block the RF signal. I made a simple replacement top plate out of 1/16” plywood. I could not find nylon standoffs and screws that were tall enough but I found an even better solution. I made some standoffs out of blue Nyrod outer housing. I was also able to sand a slight bevel into the ends of the Nyrod so it was angled to miss the antenna. The top plate is held on with 4 small cable ties stretched tight through the center of the Nyrod. The result is light weight, plenty strong, but still a little flexible.

A 2nd less critical modification was a small 1/32” plywood plate over the bottom of the battery with a couple of small sponge rubber pads. This allows the Gremlin to sit level on the ground for easier takeoff, and protects the battery from direct impact and abrasion. The original O ring they provided was WAY too tight, so I replaced it with a rubber band, and later a pair of O rings that stay on the quad. Now I just lift the plate and push the battery in underneath.

I considered adding a hoop to protect the propellers, but I don’t think it is a good idea. Such a thing would allow it to get hung up in trees and it doesn’t seem to be necessary. The stock props are incredibly durable. I have bent them many times, but have yet to break one and I am still flying on the originals. More armor would also be extra weight, and with a 2S battery it doesn’t need any more weight. They say you can use a 3S battery for race quad performance, but the specifications for the motors and ESCs don’t support it.

Some other minor details… I taped down the ESC’s as they were unsecured except for the wiring and could shift outward to rub on the motors. I also protected the receiver antennas with a couple of pieces of inner Nyrod. I really think these frames would be better if the arms were just 1/8” longer. The CF frame is VERY strong. I doubt it will ever break. IMHO, the arms on the CF version could be made a bit thinner to block less of the rotor disk, but now I am picking nits.

Later, after flying a bit, I colored the top plate and the rear propellers red with a permanent marker. This made the Gremlin easier to find in the green grass. It is so tiny, it can easily disappear in even short grass so the red really helps, especially when I am not sure exactly where I landed (crashed).

More later....


IMG_2346.JPG
IMG_2345.JPG
IMG_2343.JPG
IMG_2347.JPG
 

PsyBorg

Wake up! Time to fly!
Nice write up detailing the standard issue build bumps and bruises we all go thru. Feel good you got the mods done on your own to suit your needs. The set up learning curve to get one of these little ones to fly good is a touch harder in my opinion then the larger ones.

I documented all the iterations and things I learned when building and setting up mine. If you read thru it there is stuff to learn. If you just want to try something that may make it easier to fly I have posted the settings I fly with like rates and such (don't go by my pids as that is set up for a heavier quad on 3s). The rates I use now on all of my quads and it makes for easy smooth flying yet still have the ability to hammer out tricks and things. I also did a lot of work tweaking the auto level settings for Horizon mode (don't bother using angle mode at all) so the Gremlin is docile enough on two cell to be flown in doors. Three cell if you have spent a lot of time flying quads.

The stock Gremlin power pack can indeed fly 3s and do it efficiently but it requires short flights like under 30 seconds each to keep motors cool when tuning from the stock Beta flight pids. Once rates are set and the tuning process done they rip on 3s and do not get hot enough to worry over. I was flying mine at a lardy 180+ g all up weight when I first started with it and it did well. Now it is on a diet as I change out to better suited gear for its size and it rips like crazy outside.
 
Liftoff as a flight training simulator for the Gremlin

At the same time I purchased my Gremlin I also bought a copy of “Liftoff” via the game access site “Steam”. At around $20 for a preliminary copy I had little to lose. My intention was to use it as a simulator, not a game. I think it has been useful but it is more complicated than that. I am not qualified to review Liftoff, so I can only speak to my experience with using it.

The game installed and ran without difficulty, something I don’t take for granted where PCs are concerned. It does some weird things from time to time but nothing that prevents me from using it in my limited way. The User Interface (UI) was a bit confusing but that probably says as much about me as about the game. They have a bunch of quads to choose from and I eventually settled on something called a “Barium” which is described as a simple sturdy quad.

I decided from the beginning that I would only fly in rate mode. Using Autolevel would have made things easier, but too easy, and it would have messed up my newly developing reflexes going back and forth between modes.

I briefly tried their tutorials but found them frustrating and not useful. Flying through a series of gates in a straight line should be easy, but the gates are small and the combination of holding altitude while dealing with the inevitable left to right drift was too much. I might be able to do it now, but it was too much at the time so I left and never looked back.

I started free-flying in the most beginner level there is, a rectangle fenced in field with hay bales and scattered trees around the edges. My initial attempts quickly resulted in a crash, and that was just flying around at altitude not trying to do anything. The quad behaved very strangely. Left rudder stick would result in a right roll and right roll would result in right yaw. Of course, quads have no requirement to go where you point them and they don’t naturally point in the direction they are going, so all of these things combined to make it virtually unflyable. I would basically crash, crash, and crash, for about 5 minutes until I will too pissed off to continue.

That was almost the end of it, but I came back to it a few more times and gradually my brain started to sort it out. Within 10 minutes of stick time I could wander about aimlessly without crashing most of the time. At that point I needed a goal to focus my efforts, so I set myself a task to fly around the fence line above the tree tops. This was incredibly sloppy at first with me wandering all over the place, and rounding the corners badly. It was still quite frustrating and I could only handle 5 to 10 minutes at a time. My results were so far below my expectations that it made me angry, though I wasn’t sure who or what to be angry at. Fortunately, progress was fairly rapid and within about ½ hour of stick time I was able to loosely follow the fence line without much crashing.


I tried a few different quads and settled on the Barium, mostly because it was described as not a race quad. I also thought about how the camera angle could cause the weird roll yaw coupling I was struggling with. I found the work bench section and set the camera angle to zero (straight ahead when the quad was level). This was much better, but made it difficult to fly forward with any speed without losing the horizon so I eventually settled on a camera angle of 10 degrees. The Gremlin has a camera angle of about 5 degrees so a lower camera angle better matched the Gremlin as well.

By this point, flying around the perimeter of the field was starting to get boring. I was at least able to fly for ½ hour at a time without getting frustrated though. I wouldn’t call it enjoyable, but at least I could see steady progress. I stumbled around the user interface and found the same hay field but with a simple set of race gates, which is where I have been ever since. I tried the forest simulation once, but with tree branches everywhere that lasted about 30 seconds and just pissed me off, so I will probably never go back.

Trying to fly the simple race course was and is a joke. I am sure my skill level is crap, but I suspect that there are also technical reasons as well. My PC probably does not have the video capability to provide the necessary refresh rate so the video lag is killing me. Also the physics of how the quad flies doesn’t seem quite right. It certainly doesn’t match the Gremlin which is more manageable in slow flight.

What I can do in the simulator is fly the course above the gates. Gradually over time I have brought the altitude down to where I am just missing the tops of the gates most of the time (OK, some of the time). It is also critical that I fly smoothly with no harsh stick inputs. In order to do this I have to try to hold a line that leads to the gate ahead of the gate I am about to go through (above). If I happen to go above the next gate, that is fine, but trying to force this issue and go directly above every gate is beyond me. It quickly results in oscillations about the ideal line until I am totally out of control. When this happens the only solution is to climb and slow down because bad just leads to worse.

All of this has built some specific skills which I can now name. Automatic altitude or throttle control is the most obvious. This is happening by degrees and I am not saying that it is great, but it is not all happening at the intellectual level anymore. Another automatic skill is left thumb rudder to “look” where I am going. At first there is a strong tendency to turn your head, which is pointless. At times I can have my neck craned at odd angles before realizing that I am doing it again. Transferring that reflex to my left thumb is another work in progress as is calming the rest of my body. A third automatic reflex is the mix of aileron and rudder to mix out the effect of the 10 degree camera angle. You might call this a coordinated turn.

Ultimately, flying in the game is not the same as flying the Gremlin, only similar. In the simulator it is virtually impossible for me to fly slowly through a gate. I have to fly through with speed on a predictable line to even have a chance. In real life with the Gremlin I can slow down and fly through openings while under control. I am not perfect by a long shot, but the difference is large.

In the simulator it is virtually impossible to fly around and object while looking at it. I end up oscillating back and forth closer and farther away until things get out of hand. In real life, with the Gremlin, I can do it, not perfectly perhaps, but reasonably well. It seems that real life has more damping on lateral velocity which helps me keep it under control. The simulator is more like flying on an air hockey table, with the added difficulty of holding altitude. It is just more “slippery”.

I suspect that there are some skills that are transferable from the simulator to flying the Gremlin in real life, so it has probably helped me, but the dynamics are enough different that it also messes me up a bit. When I go from one to the other it takes me a bit to acclimate. I have tried adjusting the simulator parameters to better match the Gremlin but I can’t really do it. It may be that I just don’t know how.

For those mathematically inclined, stick input for pitch and roll is one integral away from attitude, which is roughly speaking one integral away from horizontal position. That means the pilot is 2 integrals away from controlling the aircraft position, not to mention altitude or yaw. Now add in all the visual processing to determine altitude, attitude, velocity, and where you are in 3D space, and it's a wonder we can fly these things at all....

More Later…
 
Goggles and the Video Link

I am using the Quanum DIY FPV V2 Pro. I also bought a 5.8 GHz Quanum 40 ch race band diversity receiver, a pair of circular polarized antennas and a patch antenna. It works well in some respects, and not so well in others.

I modified the “face gasket” using a combination of the foam provided and some weatherstripping foam tape that I already had. I knew from experience that the tape would not stick indefinitely, so I coated some surfaces with a thin layer of Goop to tie it all together. This has worked better than I had hoped for. It is comfortable, and almost 100% light tight.

The goggle comes with multiple acrylic lenses that can be mixed and matched to work with your eyes. I used the 2 thickest lenses and mounted them as close to my eyes as possible for the largest possible view. This provides a sharp image with no eye strain.

My original thought was that I might need to remove the goggle and fly “line of sight” from time to time, or possibly for landing, but that turns out not to be true. I take off and land with the goggle on and it is not a problem. I tried wearing my glasses under the goggle, but it was less comfortable. If I removed the goggle then the glasses would come off anyway, and flying line of sight without my glasses isn’t much of an option. I tried mounting and old pair of glasses inside a previous cheap goggle but it turns out not to be necessary. Even with a strong astigmatism, my eyes are much better than the resolution of the screen.

The video screen within the goggle is fairly low resolution, 800 x 480. I gather that this is fairly typical for FPV goggles in general as they prioritize low latency over resolution and over video quality in general. There are other goggles with higher resolution but they are generally much more expensive. I cannot fairly compare the video of my goggle to other goggles because I don’t have much experience with other goggles. Also, and more importantly, the overall video experience involves more than just the goggle(s). My sense of it is that my goggle is pretty good, and that the camera and RF link are the limiting factors.

The Quanum diversity receiver is a different matter. It has various features that sound great on paper, but It does not seamlessly switch to whichever antenna has the stronger signal in milliseconds. It only switches after you have gone blind for fraction of a second or longer, and the antenna it is switching to may not be much better.

With the standard Gremlin 25mW video camera and TX in combination with the rest of my gear I have a range of about 100 ft., assuming I don’t go behind a tree. That is a reasonable and practical range for learning to fly but I don’t see how you could ever expect to chase airplanes with it. If my experience is not in line with others I would love to hear it.

This limited range combined with the patch antenna and circular polarized antenna that I am using can be a problem. The patch antenna might increase the range to 150 ft. but only if it is pointed at the Gremlin. If it is pointed more than about 30 degrees away from the Gremlin, which happens a lot, then the patch antenna is actually weaker. This results in a lot of switching back and forth between antennas with the frequent loss of video signal for a moment when the switch occurs. I should probably just use only the circular polarized antenna and live with the short but consistent range. At least I would avoid the loss of signal when switching back and forth.

More Later...
IMG_2355.JPG
IMG_2356.JPG
IMG_2357.JPG
IMG_2359.JPG
IMG_2364.JPG
 

PsyBorg

Wake up! Time to fly!
Nothing you have described is any different for the rest of us. We have all gone thru the same struggles.

As far as sims go I have not used liftoff yet. I use Freerider as my main, then hot props for different more race style flight practice. Both can be tuned over time to feel pretty close to the real quads I fly. I will fly DRL sim once in a great while but that seems way to cartoony and more video gamish then a practical flight simulator.

You just have to learn what the settings do in the sims and keep changing them until you can get the "feel" you want in the sim as close to what ever craft you fly in real life. It takes a lot of patience and time and practice. This September will start my third year of flying.

I will leave you with this.. a real live version of Liftoff after having a bad motor, beat up props and an FPV antenna with only two of the four lobes left. It will get better and easier the more you do it.

 
Where I Fly

I fly in a small corner of a local park. The part I fly in is about 120 ft. square with short green grass and scattered trees around the perimeter. It is rarely used by other people except occasionally to nap or read under a tree. Since my video range is only about 100ft. it is perfect for my needs. The park in general is primarily used for soccer and children’s activities. I have never seen another RC pilot there, which is also perfect. At only 25mW video transmit power I don’t want to be near other FPV pilots who are likely transmitting 200mW to 600mW.

Although my camera and receiver are both capable to transmitting on 40 different frequencies, they overlap like crazy so there are only about 5 usable frequencies at any given time. Avoiding frequency conflicts at a popular FPV site can be difficult, so I avoid popular FPV sites. I also avoid standard RC flying fields as they have nothing I want or need for flying the Gremlin. It turns out that all I really need is a small patch of grass and some shade to pitch my lawn chair. If it is surrounded by trees or other obstacles all the better. As I become a better pilot smaller venues and even my driveway is becoming an option. The swirling winds from surrounding buildings are a helpful training aide, even if they aren’t the best for making me look good.

I have also flown at school yards, parking structures, and just about anywhere that doesn’t have people around. The Gremlin is so quiet that I don’t worry much about bothering people. I also avoid dogs because it would not last long as a chew toy.

IMG_2353.JPG
IMG_2351.JPG
IMG_2352.JPG
IMG_2354.JPG
IMG_2349.JPG
IMG_2350.JPG
 
Nothing you have described is any different for the rest of us. We have all gone thru the same struggles.

Not claiming to be unusual in any way. Just honestly describing the process for the benefit of other older fixed wing pilots who might consider giving it a try.
 

PsyBorg

Wake up! Time to fly!
All good mate. If you are documenting and doing the book thing I will sit back and read of your journey. Most people post in here for dialog and comments that may help better them.
 
All good mate. If you are documenting and doing the book thing I will sit back and read of your journey. Most people post in here for dialog and comments that may help better them.

It might have been better as an article but I find the article entry interface painful to use and I have my doubts that they are very often read. I welcome friendly comments as they broaden the perspective.

Forum posts also offer some feedback on whether a topic resonates with anyone. In this case, not so much. I suspect that the older fixed wing pilots just don't hang out here. I also think a lot of them will never want to make this journey. Climbing the learning curve is more of a younger mans thing.

Finally, this branch of the hobby still strikes me a technologically immature. The video frequency management is still "bleeding edge" as are the computer programs like BetaFlight and the BlHeli programming app. That last part probably hurts the feelings of the people who make the applications, but they are too close to the problem, and too experienced to see it from an outsiders or beginners perspective. Eventually, it is easier to just buy something that already works.
 
Last edited:

PsyBorg

Wake up! Time to fly!
That is fully understandable. We older guys tend to stick with things that work. Similar to that old Chevy vs Ford Vs Mopar thing with cars. Balsa guys that have been in the hobby for long do not feel the guys working with simple foam and don't worry over crashes have not "Earned" their place in the hobby and the respect so there will always be a rift.

Only those open minded enough to at least take a look or even try new things will progress the hobby. There are many techniques taken from balsa building to make the foamies fly so the two are not so removed. If the Balsa guys do not wish to grow and move forward that is a choice they make and isolating themselves only limits the longevity of their existence. I think more foam builders will equate to the Balsa building growing faster to be honest. Many branch out to Balsa or incorporate some in their builds at some point.

Balsa will always be around in the hobby and Foam will / is taking over as primary people can either adapt or stagnate the choice will always be there.

AS for the quad thing they are here to stay. There is a learning curve for everyone young and old. Specially since the tech is growing so fast. The nice thing is there are very few "Elitist" types like you see in the old school Balsa world. There are more people here willing to take time and help / teach then there is in the Fixed wing community that I have found. Fixed wing clubs that I have come in contact with are nothing more then good ole boys clubs where if you do not fit a certain way they like you will never be allowed in or accepted. In the quad world pretty much evey one can come in until you prove yourself to be a problem person bent on doing the hobby wrong.
 
Last edited:
Just a quick update on my experience with Liftoff.

I finally found the Options - Graphics - Quality setting and set it to "Fastest". This didn't impact the visual quality at all but it made all the difference in terms of my ability to fly the quad. I can now fly the quad through gates at will, albeit slowly. No longer am I limited to long sweeping lines at high speed. I can now fly fast or slow, and make the quad go where I want. I am still not a great pilot but I am 1,000% better with this new setting and it is much more like flying an actual quad. I would say that the time I spent trying to fly and learn to fly with the previous setting was 90% wasted. They really should make the default setting something that works. I suspected that the video lag was killing me, but I didn't realize just how much. The difference is huge.