Hi all!

After building the orginal Mayfly 4" and flying it mostly on 2S, I felt there was still a lot of untapped potential if I could focus on a 3S build.
The original design was just a few grams over 250g with my 3S 18650 packs, so I'm going back to the "drawing board" to see what I can do to improve it.

I have 3 main goals for this build:

1. Reduce the weight (aiming for 240g AUW)
   
2. INAV instead of Betaflight
   
3. Reach 13km from transmitter!
   

I'm really excited about this build, because it'll be my first time ever using INAV!
I really look forward to trying some of the INAV features such as altitude hold and waypoint missions, because it'll make testing much easier.
Hopefully no more cramped fingers after flying circles around a park in acro mode for 40 min!

The motors will stay this same because I think the have good efficiency, but most of the electronics will change.

I also redesigned the frame slightly, in an attempt to stiffen it up, without increasing the weight.
As the motors started to get some wear, the previous frame seem to develop a bad oscillation around 35% throttle that would send it into a fly-away.
I orginally tried a 2.5mm carbon plate with my first frame, but eventually switched to a 3mm plate, to improve the stiffness and flight characteristics, but this still didn't eliminate the oscillations.

So I've reduced some weight under the FC mounting area, where it's not needed and added some triangles to the rear arms to hopefully improve the stiffness there. My hope is that the increased stiffness will allow me to go back to using 2.5mm carbon.  By switching back to the 2.5mm carbon, using titanium hardware and 20mm standoffs instead of 25mm, I've got the assembled frame weight down to 19g vs. 22g of the previous 3mm thick frame that I was flying.

Bill of Materials:

• Frame - Custom 4" Deadcat
  
• Motors - T-Motor F1404 3800 kv
  
• AIO FC/ESC - Flywoo GOKO GN722 HD EVO 40A 20x20
  
• Camera - Caddx Baby Ratel 2
  
• VTX - TBS Unify 5G8 V3 800mW
  
• RX - IRC Ghost Atto
  
• GPS - Flywoo Goku GM8 Mini v2.0 GPS w/ Compass
  
• Props - Gemfan 4 x 2.4 biblade
  
• VTX Antenna - TrueRC Singularity
  
• Buzzer - Vifly GPS-mate (I may eliminate this if I'm desparate to save weight...) JHE20B Quad Finder LED Buzzer Beacon (lighter at 2.7g)
  
• All titanium hardware
  

I've been collecting parts over the winter and I'm going to start building today, so more build photos to come soon.

Hi Hamped_FPV

13km should be easy... :-)

Good luck

Hey Amped,

Great to see Mayfly 2.0 in progress.

Look forward to following the journey.

(27-Mar-2022, 08:21 PM)The Red Baron Wrote: Hi Hamped_FPV

13km should be easy... :-)

Good luck

I hope you're right!

It begins...

(27-Mar-2022, 09:29 PM)roteron Wrote: Hey Amped,

Great to see Mayfly 2.0 in progress.

Look forward to following the journey.

Hi roteron,

I hope to implement some of the great ideas you've put into your build!
Especially related to the battery packs.
I even picked up a spot welder to make things a little cleaner and lighter.

Actually.... question for you... do you have any tips for soldering the balance lead wires to magnets?
In my first attempt I overheated the magnets and they lost their magnetism.  Did you run into that issue?
Those were just magnets I had on hand and they were only about 3 or 4mm thick so I've since bought some longer ones with more mass in hopes that they won't heat up as much (and I'll be quicker with the soldering iron).  Did you do any prep work to the surface?

Thanks.

Hi,

Didn't do anything special. I was using nickel plated Neodymium magnets. I put a dab of flux on the end with a flux pen and then a drop of solder on the magnet to tin it, After tinning the wire I just brought the two together. It only takes a few seconds of iron each time. The difficulty I had was the iron sticking to the magnet...must have some iron tip or plating, so I put it in a metal vice that held the magnet well so it wouldn't dislodge when I pulled the iron away.

(28-Mar-2022, 03:45 AM)roteron Wrote: Hi,

Didn't do anything special. I was using nickel plated Neodymium magnets. I put a dab of flux on the end with a flux pen and then a drop of solder on the magnet to tin it, After tinning the wire I just brought the two together. It only takes a few seconds of iron each time. The difficulty I had was the iron sticking to the magnet...must have some iron tip or plating, so I put it in a metal vice that held the magnet well so it wouldn't dislodge when I pulled the iron away.

Good tip on the vise!  I had the same issue with the little ones, which is part of the reason I heated them too much.

I should be good to go now for soldering to the larger magnets.


Thanks!

(28-Mar-2022, 04:03 AM)amped_fpv Wrote: Good tip on the vise!  I had the same issue with the little ones, which is part of the reason I heated them too much.

I should be good to go now for soldering to the larger magnets.


Thanks!

You're welcome. May I also suggest another tip...3d print some small cylinders to push over the magnets to space them apart and minimise their attraction to one another. This will lower the chance of shorts when you are attaching them to the cells...the benefits of weight reduction come at a cost of needing to be a bit more careful compared to plugging in a balance lead 

So the "getto" prototype is built and hovered...


Patience isn't my strength, so I've already crashed it a few times, which was scary because it's not build for hard crashes.

Tested it first inside, and then once in the backyard.  Both ideas were pretty dumb and we had really strong gusty winds today.
As hard as I tried to destroy it... luckily it seems to have survived so far.

So far, I've had to do a few INAV configuration changes that I didn't expect, like flip the board on the roll and yaw axis, even though it's mounted top side up and the directional arrow points forward; so I don't really understand why that was required, but it worked.

I also had a hard time calibrating and getting the compass to work.
In the end I also had to do a mag alignment of CW 90 instead of the default CW 270 Flip.

Today was -8C with high winds and snow on the ground, so hoping for some decent weather soon to do more testing.

I had the compass mounted on the plastic container just to get it further away from the electronics, but that's when I was having compass issues, due to the wrong configuration settings.  So now I'm hoping that it doesn't need to be mounted up that high.  I'm going to try remounting it on a 3d printed plate on top of the frame above the camera and see how it performs.

Oh ya, and so far the dry weight is 90.1g, with way longer wires than what I need going to the GPS/Compass.

I wanted to see what the lowest weight possible is for this build, so I installed the Caddx Ant camera instead of the Caddx Baby Ratel 2 and I didn't install a buzzer yet.

During testing and my first long distance attempt I'll have the buzzer and Caddx Baby Ratel 2 installed, but I may swap those out for the lighter weight configuration if I really want to push it.

I also decided not to use the VIFLY GPS Mate.  I was planning to use that to save battery while I waited for a GPS fix, but I didn't realize this FC has 4.5V pads that are power by the USB port, so I'm just going to use a USB powerbank to power up the GPS and RX while waiting for a fix and save a couple of grams, with a normal buzzer instead of the GPS mate.

Looking good on the weight!

Do you mean you had to physically flip the board after the photos were taken?

(29-Mar-2022, 10:27 AM)roteron Wrote: Looking good on the weight!

Do you mean you had to physically flip the board after the photos were taken?

No I meant software flip.
The board is installed with the arrow on top and pointing forward. Is that how you installed yours?
I think our FC's are similar except that mine's a 722

Gooooood, on the good way !!!!

(29-Mar-2022, 01:12 PM)amped_fpv Wrote: No I meant software flip.
The board is installed with the arrow on top and pointing forward. Is that how you installed yours?
I think our FC's are similar except that mine's a 722

Yes, the boards are very similar, and physically/electrically seem the same from the outside…installed the same way as yours.

I had to flip the rc control on the roll axis. I used the rxrange command in the cli to flip the 1000/2000 to affect this. I wanted to leave my dji radio in factory default and limit mods to only one side.

I set my compass to CW90 which I thought was odd considering the mag chip is on the underside of the gps board, but it works this way. Yours is spun around 180 relative to me with it on the front compared to the back.

You did the accelerometer calibration, yeah?

Great concept. Looking forward results !
Btw, I have few questions:
1. Is this AIO board compatible with Arducopter?? If not, is it possible to stay sub250 by using ardu compatible boards like Kakute H7/Matek F405 MiniTE with separate 4in1 ESC?
2. How does this build handle winds? (I m really impressed and curious how some companies managed to make sub250 mil. grade quads that resist 60km/h winds and 80km/h gusts like UAVTEK Bug nano)
3. Have you considered new HGLRC 1603 2800kV motors too ? (Dave_C quad designer claim that its wider stator handles better 4" props; more efficient than 1404)

(09-Apr-2022, 12:27 AM)versus Wrote: Great concept. Looking forward results !
Btw, I have few questions:
1. Is this AIO board compatible with Arducopter?? If not, is it possible to stay sub250 by using ardu compatible boards like Kakute H7/Matek F405 MiniTE with separate 4in1 ESC?
2. How does this build handle winds? (I m really impressed and curious how some companies managed to make sub250 mil. grade quads that resist 60km/h winds and 80km/h gusts like UAVTEK Bug nano)
3. Have you considered new HGLRC 1603 2800kV motors too ? (Dave_C quad designer claim that its wider stator handles better 4" props; more efficient than 1404)

Sorry, I've never used Arducopter, so I can't answer that question, but from what I've heard, I don't think the F722 has enough flash memory...
It sounds interesting though.  I might have to try it someday!

It does not handle high winds very well.  I generally try to fly when the winds are less than 15 km/h.  Part of that is because it's so lightweight, but also because I purposely don't have very a agressive PID tune to maximize efficiency.

I had not considered the 1603 motors, but I took a look at the spec sheet and I just don't think they'd be a good fit for this build.
They are quite heavy at 12g each vs. 8.5g for my current motors.  So that would add another 14g to my AUW which would put me very close to 250g.
I plugged the specs into ECalc and added the additional weight and in theory the flight time and range is less than my current 1404's.

Thanks for the questions!