Ultra Budget Knuckle build

by jhitesma | February 22, 2014 | (15) Posted in Projects

So I've wanted a quad since I first learned about them a few years ago.  But not only is my budget tight I'm also really cheap.  I also know I'm not a very experienced or disciplined pilot so I knew I'd be crashing a lot which meant more expenses.  I picked up a Syma X1 and had fun with it...but it just didn't hold my interest because it was too limited.  When the anycopter was announced I seriously thought about it as a platform to scratch build a quad that would make me happy...but again the expense of doing it right scared me off.  Then the knuckle was announced and I started seriously thinking about a build because it looked more robust and easier to build out of stuff I had laying around.

This is a summary of the build.  If you're interested in the full story be sure to check out my build thread in the forums where I chronicled the whole build and continue to chronicle my experiences with this quad.

So how ultra low of a low budget was I thinking?  My initial goal was $100.  I knew it wasn't a very realistic goal, but I already had a Turnigy 9X TX and a spare RX so I thought I'd see if I could pull it off.  After a stressfull weekend I attempted some retail therapy on the HobbyKing website to relax myself.  At the time KK2.0 boards were completely out of stock and the few times they restocked they were gone before I could get one in my cart.  But MultiWii boards always seemed to be in stock so I started reading up on them to see if they would be worth trying.  The more I read about MW the more I liked what I saw and I soon realized that I actually had everything I needed on hand to build my own controller from scratch - suddenly my $100 budget looked realistic!

I pulled out an arduino megaI had picked up when I learned HK has knock off arduinos - I originally bought it for a 3D printer project but since that project got put on hold due to the expense of motion control parts it was just sitting on my desk.  I grabbed the motion plus and a nunchuck off my wii..tossed it all together on a breadboard and within an hour I had a functional (if not flight ready) flight controller on my desk!

With a flight controller figured out it suddenly looked possible to pull this off for <$100 so I placed my order:

4 - 1300kv 24g motors $32.38
4 - 10A multistar ESC's $24.96 (These were a mistake!  Flashing them with simonk eventually proved all but impossible so I'd suggest either Afro ESC's or if you're setup with a USBASP for flashing somethign cheaper like the HK F20's and flashign them yourself.  I wound up using the HK Blue Series 20A's but only because I didn't know about the F20's which are easier to flash and cheaper.  20A is overkill but I was planning on flashing the 10A multistars and using the 20A's on a bigger build later.  10A is plenty if you use teh 24g motors.)
1 - XT60 to 4x3.5mm bullet breakout cable $2.94 (this WAS a mistake, the HK website listed the 10A multistars as having 3.5mm bullet connectors but they actually have 2mm. But the breakout cables with 2mm bullets only have JST connectors and since I plan on using my 1500 or 2200mah packs which have XT60's on them those don't help me much...so I'm going to end up doing my own thing instead, will save this for a bigger quad project in the future.)
2 - Sets of 8045 quad props $5.18 (I know I'll break props...probably should have ordered more than 2 sets...and really kicking myself for ordered both sets the same color.)
1 - 8ch 9x RX $8.99 (I had planned on using one of my existing RX's...but my order came in low enough I was able to splurge..good thing too because the next day I lost one of my RX's and no longer had a spare!) 
1 - MultiWii Bluetooth module $7.49 (this was a bit of a risk since I wasn't sure if it would work with my homebrewed controller...but I didn't see any reason it shouldn't and it was cheap enough to risk. Not to mention the idea of being able to program the controller from my phone was one of the things that REALLY impressed me about the MultiWii. Turns out it was worth the risk as it took me all of 5 minutes to set it up and it seems to work great!)

Even with shipping my order was under budget enough I was able to toss in a few spare props for some of my other planes and still keep the total under $95. Woo Hoo! 

Unfortunately I quickly realized I forgot a few things...and reading up on MW more so I also ordered some extra sensors off ebay:

1 - BMP085 baro sensor breakout board - $2.95 shipped (Still waiting on delivery, by mail from Hong Kong...I'm not holding my breath.)
1 - HMC5883L 3 axis magnetometer compass - $2.31 shipped (Still waiting as well...but like the baro I'm not in a hurry since it's not needed and is just for me to experiment with after I have this thing flying.)

And an order from HK's USA warehouse for other things I forgot:

4 - 3mm prop adapter $4.72

2 - 10pc 2mm bullet connectors $4.86

 

With all the parts I should need on hand I was ready to start building!

 

I was kicking myself for only ordering one color of prop...but I knew I'd break them and planned on ordering more soon so I didn't kick myself too hard.

 

Now it was time to get really cheap.  Like stupid cheap.  I love FT, and I would have loved to just buy the knuckle hub kit but my budget was at it's breaking point already so I had to scratch build.  I had some Monogram/Revel 1/8" ply on hand from before I fell in love with FT's laser cut firewalls which are made from much nicer ply and are cheap enough that even for me they're worth buying.  So I started on making my own knuckle hub set.

 

First I used some spray adhesive to attach the plans to the ply and center punched each hole:

 

 Next I got out my brad point drills since I wanted to be as accurate as possible:

 

And proceeded to drill all of the knuckle holes:

 

With all the holes cut it was time to separate the knuckles...but how? I have a nice old 1940's Delta bandsaw that a friend gave me for free because it was a rusty mess on the of his house. I disassembled it, cleaned it, rebuilt it and it works...but I haven't found a new tension spring for it and as a result it tends to break blades and I'm out of silver solder to repair the blades right now. Plus it's not the best for straight accurate cuts. My table saw is overkill and has way too big of a kerf. My xacto razor saw has a nice thin (.015") kerf and would do it...but it's time to replace that blade as it has almost no teeth left so it would take forever to use it. 

I decided to do it the hard way. And this is one of those points where I effectively spend more to do something the cheap way. $12 to order a set of knuckles from FT would be a lot cheaper than the time I'm spent cutting these by hand with this method (WAY cheaper if I was to charge myself the hourly rate I charge clients when doing my day job!) I got out a steel straight edge, some clamps, and one of my utility knives along with some new blades:


A half hour later I had them cut out enough to test how accurate they are:

 

For more details about this process see my build thread in the forums bottom line is it worked and my frame came out nice and square.  But if I had it to do over again...I'd probably spend the $12 for the laser cut parts because while my primitive methods worked they were not much fun!

After some paint I had my basic frame assembled

 

I went with hot pink because my daughter likes it and I figured it would be nice and visible.  It also contrasts nicely with the black I used on the knuckles.

After some assembly that I won't rehash again but is covered in detail in my forum post I had things together enough for a test!

Holy cow, this is looking more and more like it's going to actually fly!

After a bit of cleanup to the wiring I was ready for a test flight and had a taste (a brief taste) of success:

I proceeded to make better landing gear out of ply instead of brittle acrylic, and started to have slightly longer flights.  But pretty quickly had a crash hard enough to break a boom which gave me an oportunity to do some major rebuilding.  You may have noticed that my center booms were closer together than on most knuckle builds - that was due to me misunderstanding the video instructions from FT.  So in my rebuild I corrected for that and moved my center booms out, added a top plate to mount things too and made some zip tie landing gear because even the ply was too fragile for my beginner pilot skills.

 I continued to make minor incremental improvements, my flights weren't getting much longer but that was mainly my skill and not the quad at this point:

Then I finally got to try in a bigger area and tested auto level which worked great.  In fact it worked so well the next day I attached my old cell phone with some rubber bands and DT foam and made my first attempt at aerial video from a quad:

Not great, but enough to get me excited.  And that evening I did a bit better once I realized the white balance was off on the cell camera:

By this point my investment had reached about $150 but that was mostly due to things I had broken and replaced learning to fly a real quad :D

Next I added the baro and magnetic compas sensors I had ordered off ebay - just stuck them on with a bit of hot glue at first to test them:

 

 Hey that was easy!  They worked!  I did need to use a level converter from Sparkfun (which cost way more for S&H than the board itself did!) since the new sensors were 3.3v while the wii parts and arduino were 5v.  

At this point I was hooked.  The quad hit a few roadbumps and the project took a few detours chronicled in that forum post I've linked so many times already I'm not going to bother to link it again.  But overall things just kept getting better.  I ordered a UBLOX NEO-6M GPS off ebay but before I could get it wired in I had to fall back from my Mega board to an arduino pro mini (which wouldn't support the GPS) while I waited on a replacement mega board.  I also ended up ordering a MPU-6050 gyro/acc breakout board to upgrade my primary flight sensors.

Once the new gyro/acc came in I cleaned up some of the wiring and made myself a separate "IMU" (Inertial Measurement Unit) board to replace the wii parts I had been using and get all of my sensors inside the frame where they'd be safe.

 

 

 

 

I also got the GPS wired up and working at this point.  My first test flights with the new sensors blew me away!  The wii parts got me flying..but the modern gyro/acc made a HUGE difference and made the quad much more manageable:

 

A friend loaned me his Emerson 720 action cam and even with it just zip tied right to the frame things were looking better and better...still a lot of jello but with no vibration dampening on the camera I wasn't surprised:

 

At this point the budget was still pretty good.  The GPS was <$20 off ebay (though I did spend $12 on a $1 battery for it because the ebay seller said it wasn't included and because it was lithium digikey charged and arm and a leg for shipping...and of course when the GPS did arrive it turned out it actually HAD the battery despite the seller saying it didn't.)  The MPU-6050 breakout was only $6 delivered off ebay.  All in all not counting my TX/RX and parts I broke or bought but didn't need what I had could be replicated for about $150 at this point.

Then I got an openLRS from my wife for Christmas.  At this point the project really wasn't budget focused anymore!  The openLRS has been great.  I had to make a few tweaks to it (explained in my forum post) but it's worked out very well and cleaned up the quad considerably since I now run a single PPM wire instead of 8 PWM wires for control between the RX and flight controller.  I also enabled telemetry and moved my bluetooth module from the quad to my TX so I get full telemetry on my phone while flying and can easily adjust PID's and other settings right from my phone.  I can even do full logging and import flights into google earth now:

 

Next thing you know I bought myself a 5.8ghz video setup and put it together on a budget with a camera and screen I had laying around - but that's still a work in progress and a story for another article!

I did do one more upgrade at this point and finally switched to simonK.  I tried to flash my 10a multistar ESC's but quickly learned it really is all but impossible to flash them.  They don't have programming pads and the processors on them are the ultra small leadless packages.  Even with a fine tipped soldering iron and magnification and tiny wires from an 80 pin IDE cable I wasn't able to get them reflashed.  So I used the method RC Hacker shared and flashed the Blue series 20A ESC's I had bought for a planned second multirotor.  

 

Upgrading from the original wii sensors to the MPU-6050 made a huge difference...upgrading from the multistar ESC's to simonk made just as big of a difference again!  At this point my quad still looks like something pulled from the back of Dr. Brown's Delorean but it's flying great.  With simonK I'm able to fly much more aggressively and have finally done a few flips successfully as well.  

 

Since I just had to edit this article because all of the videos stopped working here's a little bonus update. I've been practicing my flips and getting better at both flipping and crashing. Last weekend I was able to get my wife to record me showing off a bit:

Of course the one time she records I go for a double flip and flub it. Only damage was a few broken zip ties though. I've gotten to the point now I can even do flips in my backyard and don't loose much altitude doing them anymore either.

 

I still need to make a proper antenna.  And I've probably spent closer to $250 than my original $100 budget at this point.  But with what I've learned I'm now sure that a great flying quad can be built for my original goal of $100.  

Again, lots more details (LOTS) in my forum thread and this is still an ongoing project - heck my RX is still just held in place with a rubber band!  But my goal has been achieved and then some and I've been putting off writing this article for way too long.  If you've also been putting off trying multis for budget reasons hopefully this will convince you it can be done on a budget.  

The keys to pulling this in on budget are to keep your expectations realistic and try not to get carried away.  The 24g motors and smaller cheap ESC's (just be sure you can flash them or spend a few extra dollars for the AFRO's or other pre-flashed ESC's) are a bit part of keeping costs down.  The homemade flight controller helped keep costs down as well - but the MW boards from Ready to Fly Quads are a great budget choice if you don't have an arduino on hand or aren't up to a total DIY flight controller.  

That said...I would not build a quad with the 24g motors again.  They work really well on the knuckle, no complaints there.  It's just that you have to epoxy the wires on them or you will break them (in my forum post I detail rewinding one after learning that lesson the hard way) and it's way too easy to bend shafts in a crash. But for the price they're tough to beat.  If I had a bit more of a budget (like I'm trying to plan on my next multi) I'd go with a short shaft style motor like the NTM series from HK.  I've bent a prop mount on an NTM motor before - but it's a lot quicker and easier to replace one of those than to replace a shaft in a 24g motor!  

Hopefully sharing this will help someone else avoid the mistakes I made and tackle their own ultra budget quad project and actually keep it under budget! 

COMMENTS

CJGFX on March 6, 2014
Just to let you know that none of the YouTube videos seem to be working.... The Knuckle Quad is going to be my next build... Looking forward to watching the videos
Log In to reply
jhitesma on March 17, 2014
Not sure what happened to the videos, but they're all fixed now. (Just re-pasted them and they started working again.) Even added a newer one while I was in there :)

The knuckle is a blast. This thing takes a beating and just keeps going. I fried my FPV camera last week so this weekend I was flying acro again and having a bast. Even with the little 24g motors I'm doing multiple flips pretty well now that I figured out the best settings for it.

Even after a little miscalculation that resulted in a bent prop and broken wooden boom I still put it back up (just pushed the boom into placed and the pressure of the knuckles seemed to hold it well enough to fly.) Amazingly even with a broken boom it still flew great and I was able to pull off a few more flips before knocking the broken boom off :D

Log In to reply
3DMad5 on March 28, 2014
Any particular reason you went with the Arduino Mega? Would you choose something else if you were to do it again? The reason I ask is that your article inspired me to try something similar, and the HobbyKing website has a bewilderingly large selection of Arduino related products... Also, any chance you could do a build log/instruction log on the custom FC?
Log In to reply
jhitesma on March 31, 2014
I went with the Mega mainly because I happened to have one within arms reach the day I started really reading up on multiwii :D

That said I would go with the Mega again, the ProMini is also popular for MW because of it's small size and lost cost ($5) - and I did run a promini on mine for a week or two after smashing my first Mega board in a crash. But for my plans the extra serial lines and connectivity of the Mega are well worth the extra $10 since Mega boards can be found for around $15. The promini only has one (actually maybe two IIRC) serial line - which is REALLY limiting. With the pro you can't have USB and GPS connected at the same time let alone telemetry. With my mega I use 3 serial connections - 1 for GPS, 1 for Telemetry, 1 for my OSD - and I still have USB for programming since that's built in on the Mega.

I could write up an article on how to use a Mega board like this. The only real tricks are getting the arduino environment setup on your computer which can sometimes be tricky, and then figuring out which pins to use for what - which is mostly documented on multiwii.org but I did occasionally have to dig into the code to figure out how things needed to be connected. The most difficult part was wiring up the level converter since the Arduino runs at 5v but the sensors run at 3.3v. And even that wasn't hard it was just annoying to have to spend more on shipping for the converter board than the board itself cost :D

Log In to reply
saitek93 on April 13, 2014
Nice work! The first run of my projects usually look messy, but I go for functionality first, and then presentation. Great stuff, and thanks for sharing.
Log In to reply
jhitesma on April 27, 2014
One of these days I really want to finish a project and make it look nice :D I tend to get too carried away with my projects though and just keep upgrading and tweaking making them uglier and uglier. When I do spend the time to make them look nice it seems that they get cursed and then die in flames...so while I love building something beautiful doing so tends to mean it becomes a "trailer queen" and never gets used. (Like the 3" Sandia Sandhawk model rocket I built in 1996 and have yet to fly because while it's built better than anything I've ever built before it's also prettier than anything I've ever built so I'm convinced if I ever do fly it it will blow up on the pad :D)
Log In to reply

You need to log-in to comment on articles.


Ultra Budget Knuckle build