Blackout Mini H Quad Build Log

Tritium

Amateur Extra Class K5TWM
@jhitesma

Why does the call sign have to be displayed? I read that the station has to ID every 10 min and at end of transmission. I do not read that it has to be a visual display. I am working on a "voice" ID box to use for FPV that uses the sound channel on the FPV transmitter. So far I have spent $1.98 for a 10 sec voice recorder from China. All it needs now is a 10 min cycle timer.

Thurmond
 

jhitesma

Some guy in the desert
Mentor
@jhitesma

Why does the call sign have to be displayed? I read that the station has to ID every 10 min and at end of transmission. I do not read that it has to be a visual display. I am working on a "voice" ID box to use for FPV that uses the sound channel on the FPV transmitter. So far I have spent $1.98 for a 10 sec voice recorder from China. All it needs now is a 10 min cycle timer.

Thurmond

You're correct - it doesn't have to be displayed visually the station just has to ID. It's perfectly fine to ID by voice or morse. I choose to do it by display since:

1) It's supposed to be a built in function of the OSD I'm using so it shouldn't require anything additional (Freebie!)
2) To me it seems most in spirit to the rules to to ID visually when the purpose of the transmission is a visual signal.

I would say using an audio carrier is equally in spirit - but since I don't even have anything hooked up to my audio on the RX end I couldn't even confirm my own ID through that method so to me it just seems like a slightly dishonest way to do it.

So bottom line - personal choice :D

I have considered simply using an extra pin on my arduino and a bit of code slipped into MW to generate a CW ID string that could be used...but decided it was more work than just typing my call into the field that's already there in my OSD :D

I do still want to see openLRSng modified so it's beacon can send a morse callsign instead of the tones it currently sends...I've looked into that a few times but haven't gone very far towards making it happen.
 

FinalGlideAus

terrorizing squirrels
For a number of reason's I've not had a chance to continue on with this build until now. Working 65 plus hours a week at the moment doesn't give much fiddling time. Plus I was still waiting on a few key parts but now it's all here (I think....) I can get cracking. I hope to be flying it this weekend but work hours will dictate if that happens or not.

Last night and tonight I got some of the electronics sorted. Because I'm trying for light weight I'm giving it the "warpquad' treatment and trying to shed weight where I can. Most of these quads are at least 550-600grams with a few prokies going well over 600. I'm hoping for just over 500grams flying weight.

First up was the Naze 32 pin connections. I always order my Naze without pins. I have a selection of styles of pins which I choose to customize the build for the copter it's going in. The usual Soma low profile setup was installed under the Naze giving a very clean and tidy look without wires hanging everywhere. Although I'd love the simplicity of using a satellite receiver I need long antennas and diversity to counteract the blind spots caused by the Carbon body so a an AR600 is being installed with the standard PWM pins on the side of the Naze. If left as is those connectors would stick out the side of the body so I intend to turn the Naze 90 deg so they face backwards and then in the CLI reset where the Naze thinks forwards is. Pretty clever little FC that thing is. This also means easy access to the USB plug when connecting to the computer.

The PDB on the mini quad has solder points to power the FC but it is 12v so you would normally power the FC from an esc. I plan to change out the BS12a esc's for another kind soon that doesn't have a BEC installed so I instead wired up a little Pololu 5v step down regulator which will live under the Naze and be wired up to the PDB and power the Naze. It is possible to attach it to the Naze but since I have a bit more room on this build than the Warpquad I will just stick it on the PDB.

I finally got my ImmersionRC VTX last weekend so I set about looking at what I could do to it. It has its own regulator built in so can run off most voltages which means simply soldering it to the PDB. It also has a 5v ouput that is filtered to power the board camera. I want to make use of this noise filter to get a clear image so I don't have to make another filter but the camera I am using on the quad (sony 600tvl super had) runs off of 12v. The simple answer is to wire in a Pololu 12v step up regulator. While inspecting the VTX I noticed the plugs were overly large so figured I could save some weight and space by removing them and hard wiring everything. Here is what the original board looks like. The plugs are those big black things on the end.
http://www.hobbyking.com/hobbyking/...d_NexwaveRF_Powered_Video_Link_Fatshark_.html
After I did that a whole heap of real estate opened up on the board and I decided to mount the step up regulator in the VTX. I'm quite happy with the results and now have a lighter vtx which is 12v camera ready and all in one.

Well that's it for tonight. Another 11 hour work day coming up tomorrow so time for some sleep.
 

Attachments

  • IMG_1876.JPG
    IMG_1876.JPG
    212 KB · Views: 83
  • IMG_1871.JPG
    IMG_1871.JPG
    293.5 KB · Views: 121
  • IMG_1874.JPG
    IMG_1874.JPG
    287.1 KB · Views: 131

trigger

Senior Member
On the other side of the planet, my Mini H Quad build went quite smoothly after fixing my video<->+12V short-circuit problem... Until yesterday.

The shop I bought the ESCs from sent me an email, explaining they had a slip-up and one batch of BS12a wasn't flashed properly to SimonK, and was still running the stock firmware.
I had previously (but not thoroughly) tested the ESC/motor combos, and it seemed to work, but now that they mentioned it, I realize that the startup beep sequence of the ESCs isn't the usual SimonK one.
Which means I probably need to flash them to avoid troubles with the Sunnysky X2204's. Which is the one thing I didn't want to have to do...


To make things worse, I won't be home for the next two weeks (I'll be on vacation, and I had hoped to be able to fly the Mini H during that time...). I ordered the socket flashing tool from HobbyKing, shipped to my father's where I'll be staying at for a few days.
But since there's no guarantee it'll be delivered on time, I'm thinking about flashing the ESCs "manually" (i.e. soldering wires to the Atmega pins) this week-end. But, it looks like a lengthy and risky process, and makes me quite uneasy (and I don't have spares) :-/

On the bright side, It gives me the opportunity to try BLHeli out...

Anyway... First world problems ;-)
 

FinalGlideAus

terrorizing squirrels
I will update tomorrow with some info and pics but the frame is almost complete with only some setup in the Naze to go after work tomorrow. Pretty happy with the weight coming in at around 430 grams in racing mode and just under 500 grams with a larger battery and Mobius attached. It's amazing how much weight you can shed when your picky about the build. Time for some sleep.
 

bucket183

Member
Finalglideaus,
Great build log keep it up. A few questions if I may

Can you post a picture of the pcb what goes where etc, from the Bo picture I'm confused about the wiring of the camera and vtx. The video pad is that just for signal?? So if using a möbius or gopro I would need to wire the video signal to the pad and then have the power go through a 5v reg to the power pads?? For the vtx (immersion tx) connect power to power pads and video to video pad?? Which pads are best to connect the the battery lead to?

You said you got the rubber rings for the flight controller from Soma, who is this? friend?? Shop? I would definitely like to source some plus the purple addons once back in Oz.

Assuming I had all the gear, how long from unpacking to first flight?? I fly back into oz on a Sat, and fly out to Perth on Mon and would love to take it with me, do you think a day is enough to get it together??
 

FinalGlideAus

terrorizing squirrels
Apologies for the lack of final updates. Work is crazy busy right now and 65hours a week leaves little flitetest time. I will set aside some time to add the post tonight after work. I will include pictures to answer some of your questions.

I have put about 20 packs through the BO quad now and getting it flying nicely. Also done some mods to improve performance.

The rubber orings are not 100% needs but help with isolating vibrations to the FC. To get some just find an oring supplier.

You could build it in a day but it depends on your previous experience. If this is your first mini quad, first Naze 32 and first FPV machine I think it will drag out longer. Don't rush things on the build.

All the power pads on the PDB are 12v. The receiver and Naze 32 run off 5v. You normally run these off a bec from an esc but I used a 5v stepdown regulator seperately. Not sure what voltage the GoPro or möbius run of but you want to be flying through a board camera anyway. The immersion vtx has an inbuilt regulator so you can wire that to the PDB direct. It also outputs a filtered 5v feed that you can plug direct into with a fatshark board camera that runs off 5v. The sony board cameras that people use run off 12v hence why I installed the 12v step up regulator in the immersion vtx.

The signal pad is just for the signal line although I didn't even use it. I will update with pics.
 

FinalGlideAus

terrorizing squirrels
Next up was to get the components ready. The S22024's had the motor wire exits epoxied and because there may be the occasional spray of water depending on where I fly I made sure to water proof the FC and FPV camera. I just used some liquid electrical tape. After I bought the tin I realized it was red but I don't mind the extra bit of color now. The esc's were prepped by flashing them with BLHeli FW and removing all wiring that wasn't needed. I only need the signal wire form the servo cable as I'm not using the bec's and the ground wire isn't needed either because the FC is grounded and all ground cables are common.

It was also a good time to solder some of the components to the PDB. To help balance the quad better I put the VTX up the front so used the front middle pads to power it. Because I am not using the esc bec to power the FC but instead using a Pololu step down regulator, I soldered that to the center PDB pads. The extra wires seperate from the Pololu board are the ground wire for the esc and a 12v wire for the Naze 32 battery sensor.
 

Attachments

  • IMG_1878.JPG
    IMG_1878.JPG
    340 KB · Views: 74
  • IMG_1879.JPG
    IMG_1879.JPG
    251.8 KB · Views: 193

FinalGlideAus

terrorizing squirrels
Here is the finished product. Final weight came in at 430grams in race config and just under 500grams with a larger battery and Mobius. By being careful about the build I think I saved 50 grams. I decided to leave out the OSD as it just be a distraction and I only wanted it for voltage monitoring and I have the Naze voltage beeper for that. I have friend who have OSD's in theirs and they never look at them.

I borrowed Boris B's PID's
http://forum.flitetest.com/showthread.php?8966-Naze-32-vs-CC3D-vs-Flip-Pro-Help-my-brain-hurts/page2
and it got me airborne without any hitches but I have since tweaked it somewhat more to liven it up and because of the new props I'm running. The very first thing I noticed was just how pathetic the thrust was with the Gemfans. Most people use these and I really don't understand how! They are the worst pieces of trash out there. If you are descending at any decent rate you will never pull out in time before you hit the ground. I guess if you're a complete beginner then they are something cheap to learn on but I couldn't get them off quick enough. The Blackout carbon props are not much better but are also very expensive. The HQ 5x4 props are much better but shatter on impact because of the carbon blend and are more expensive as well. When the nylon versions come out it will be better but they're not here yet. Anyway, apon talking to Soma I had a whinge about the fact I was driving a tank around the sky and 24hours later I got the props in the picture shoved in my hands and was told to test them out. They are the nylon FC 6x4.5 props but have been cut down to 5". I tried them the next night in the dark and WOW, that is much better. Although the quad could do with 6" arms, with these on it is at least fun to fly. I haven't done a thrust test on them but I would guess they have increased thrust by half again. Needless to say, these are my go to props from now on. Aparently Soma is talking to HQ prop about making these cut down blunt nose props from the mold which will make things easier. What can I say..... Soma is the MAN!!! :D

To help with landing on hard surfaces I got some 12mm x 6mm sticky back rubber and put in various spots underneath. It actually works really well, weighs next to nothing and is simple. It also hold the antenna in place on the bottom. As I stated earlier, I used a full receiver in this build to make use of the long antennas and diversity. The one on the bottom faces backwards and top one goes vertically. I used the inside of an old Biro pen to make the mast. With this setup there is very basically no position that covers both antennas, improving reception. I have also setup failsafe on both the the RX and Naze 32 so have twin failsafe for extra piece of mind. I still have to do some more in depth range testing but the initial tests proved very positive.

The esc's were mounted "Warpquad" style. The esc's are stuck on double sided foam tape and then the the whole arm and esc is covered in heat shrink which gives a cleaner look.
 

Attachments

  • IMG_1892.JPG
    IMG_1892.JPG
    172.9 KB · Views: 81
  • IMG_1898.JPG
    IMG_1898.JPG
    181.7 KB · Views: 76
Last edited:

FinalGlideAus

terrorizing squirrels
With the top off you can see my wiring. On the PDB the four corners have the esc's soldered. The front pads power the VTX and the vtx powers the Sony board camera via the Pololu 12v step up regulator I installed inside the vtx earlier. This leaves a simple and minimal wiring system that is easily removable with two plugs. The centre solder pads power the the Pololu 5v step down regulator which in turn powers the FC and RX. I soldered the battery lead on the inside pads of the two at the rear because the relief strap (ziptie) would not do anything if it was soldered to the outside pads. Kind of a silly place to put those pads. The Naze beeper is sitting under the Naze and secured by friction alone.
 

Attachments

  • IMG_1893.JPG
    IMG_1893.JPG
    178 KB · Views: 92
  • IMG_1894.JPG
    IMG_1894.JPG
    230.3 KB · Views: 212
  • IMG_1895.JPG
    IMG_1895.JPG
    194 KB · Views: 72
  • IMG_1896.JPG
    IMG_1896.JPG
    190.4 KB · Views: 93
  • IMG_1897.JPG
    IMG_1897.JPG
    227.2 KB · Views: 82

FinalGlideAus

terrorizing squirrels
All in all I'm happy with the finished product. The new props have really made the quad come alive. I've got some new batteries of various sizes coming to test out but I think based on others experience I'll settle for some 1500 and 1800 packs.

Positives:
The quad is extremely well made and easy to assemble.
It has a good following now thanks to the Flitetest guys.

Negatives:
The build manual is hopeless. The current manual is for the older version that has different sized bolts and no PDB. With the price you are paying I would expect something a little more up to date. With a little online research I managed to find the info needed but people with less experience may struggle. If not done correctly you can make some important mistakes with the PDB in particular. Some people have mounted the PDB on the bottom which will damage it. another guy put his on upside down and shorted everything out on the carbon arms. Others have asked what pads do what. An up to date manual will fix this.

Lastly, the quad really needs 6" arms. Considering what you can get out of these mini motors, 5" arms are not doing them justice. I was considering using my connections to get some 6" inch arms designed and cut but there are plenty of other great mini quads out there to choose from that already have this option. The Minion is one example and is a nice simple little frame.
http://patsminihquad.com/the-minion/


None of that really matters because today at a secret location a new FPV racer had it's maiden flight that will make people re-think what performance really is. I will be having a look at it tomorrow myself but get ready to be impressed. Think acro quad agility with tons of power. I will update when I can ;)

My next build will be either one of those or another acro quad that has enough power and speed to make you wet the bed just thinking about it. My motors should be here tomorrow. Get excited people...
 

FinalGlideAus

terrorizing squirrels
Looking nice!

What's the conformal coat your using?

Use smaller words Crafty. I'm only a simple guy :)
Are you talking about the sealant over the solder pads and FC? That's called "Star Brite" Liquid Electrical Tape. It dries with a rubbery finish.

BTW: heard your name mentioned in the latest FT video :)
 

FinalGlideAus

terrorizing squirrels
I don't have video of mine flying yet as my Mobius is still in the mail. I can however show you some of the fun had during my groups flying meetup on Sunday where I maidened the BO quad. I also got to have a look at the new 330 BO quad prototype thats not for public viewing yet.

This video is from Strepto's footage on the day and is coming from a Minion mini quad (with 5" arms). Cranial will enjoy the music selection I think :).

https://www.youtube.com/watch?v=i7iCfFyP8BI
 

jhitesma

Some guy in the desert
Mentor
I'm curious...how do you have the two failsafes setup?

I considered going that route at a few points on mine...but was worried it may actually make things less safe. Let's see if I can adequately explain my thought process as to why. (And this may also be different on how Naze handles failsafe compared to MW...I haven't fully investigated that.)

The failsafe on MW kicks in when one of the first 4 RC channels goes beyond a given set point (defaults to something in the 960 range IIRC.) So I figured setting up failsafe in my RX to hold set values would keep the MW failsafe from ever kicking in and could in fact allow for a flyaway since the quad wouldn't necessarily hold itself in place. (doubly so with no gps)

I figured I could set failsafe in my RX so it enables the GPS hold as well as fixed RC values...but really what made the most sense to me was to just go with the "semi-controlled fall" of the MW failsafe and get the thing out of the air if control is lost.

So I setup failsafe on my RX to cut all channels when the link is lost allowing the MW failsafe to kick in and let the quad "land".

In the case of my flyaway if I had had my RX set to hold a low throttle I probably would have still had a flyaway as 1) on the new motors it takes much less throttle to stay airborne than I expected, 2) since it happened at the start of a pack unless the throttle defaulted to a low enough value to let the quad fall it still would have drifted away before the pack went dead.
 

Craftydan

Hostage Taker of Quads
Staff member
Moderator
Mentor
FGA,

yeah, that was me :) nice to get the shout-out, but far more fun hanging out and enjoying the festival! They were right -- the combat would have been totally crappy if we weren't having so much fun. That and while multirotors were everywhere in pockets, it was very underground. The organizers had *no* idea what to do with us. In the end, if we kept to the pattern or off to the side, they ignored our lack of wings . . .


jhitesma,

have you dug much into the code around the onboard failsafe?

I'm toying with the idea of tying it to enble baro/GPS hold/angle mode, then slowly lowering the baro setpoint at about 0.5 ft/s -- that's a healthy descent, but not frame shattering. The GPS should cancel lateral movement, angle to keep it flat, and baro will set it down wherever it is. Give it 100' worth of descent (200s) or baro error maximum(set point below the ground), and disarm.

Any thoughts?
 

jhitesma

Some guy in the desert
Mentor
I've reviewed the code enough that even calling it a "semi-controlled crash" is a bit generous :)

Basically all it does is start a timer once the RC value goes below the set point - once that timer reaches a setpoint it goes into failsafe and just locks in at the throttle point defined in the config.h. If no vertical acceleration change is detected then another timer starts which eventually disarms the board. There's no attempt to control descent rate or use any of the advanced sensors if available to do anything intelligent.

There are patches people have come up with to do smarter things - but almost all of them are based on the 2.2 firmware (like e_lm_70's version) and the patches I've seen based on 2.3 aren't really tested. EOSbandi has some improved failsafe in his nav version which is supposed to be incorporated into the next official release - but who knows how long that will be. His beta version does seem to be stable though and I've been wanting to experiment with it. Just got distracted by FPV and then the issues relating to my flyaway and now vacation :D

In general it's one of those things that sounds simple "Just go into angle and use the baro to limit the rate of descent" but guys like EOS who've spent a lot more time than I have thinking about it are quick to share laundry lists of reasons it's not really all that simple ;) The main reason the "improved" failsafe methods are still just unofficial patches are because there are too many potential ways things could go wrong with them for the maintainers to accept them into the official release. What if it goes into GPS hold...but GPS doesn't have a lock or isn't tuned properly and sends the copter off instead of holding it. What if the baro isn't setup correctly or angle isn't tuned right. On a properly setup and tuned copter those "improved" failsafes would work great...but if things aren't setup just right they could make the copter do things worse just fall out of the sky.

Now that summer is here and I'll be spending more time on the bench and behind the monitor instead of at the field I hope to spend more time getting to understand the MW code internals and looking into some of the improved failsafe options. But...I've promised myself I'll finish my android game project and get it live in the play store before digging in on any other software issues. (And that project is CLOSE...it's in private beta and playable. Just have an issue with a font and a few menus to finish cleaning up and a few gameplay tweaks I want to make. Except the libGDX library I used has changed so much in the 10 months I've been letting my project gather dust I've got a lot of updating to do!)
 

FinalGlideAus

terrorizing squirrels
Jhitesma. My failsafe is just a simple one....cut the power to zero. I have that on both the RX and FC. In a normal failsafe situation the RX will failsafe first everytime but if the power comes loose from the RX, without the failsafe on the FC it would just hold last postition (I think) and if you're in the wrong attitude and throttle setting it's gooooooone. This just covers basically all the bases. I don't care about it falling to the ground as it's strong enough. I just want to get it back. :)

A silly guy on RCGroups set his failsafe to hold last position and it kicked in during a punch out with a fresh battery. Sounds like it was climbing past 1500ft when he lost video signal! Through dumb luck an anonymous person dumped it on his doorstep a day later! The mobius didn't capture the fly away because the crash dislodged the memory card corrupting that flight but there was another flight on there of him flying in front of his house and there were some well know businesses in the video so they must have worked it out themselves. Very lucky man. Needless to say he is re-configuring his failsafe now...
 
Last edited:

cranialrectosis

Faster than a speeding face plant!
Mentor
I think the simple solutions tend to work the best the most often.

What I would really like to see is how to test it. I know my FS works. I can see that when I turn off the transmitter the copter motors stop spinning. Testing the altitude hold failsafe in the traditional sense could get harsh on the wallet.

If for no other reason that that, I think this particular nuance of flight control is vague today.

Still, I will relish reading your findings jhitesma and Dan. :)
 

Craftydan

Hostage Taker of Quads
Staff member
Moderator
Mentor
jhitesma,

Complex systems are always fraught with unintended consequences . . . but it's good to know the field has mines, even if you don't know where they are. I'll tread lightly ;)


FGA,

YIKES!


CR,

There's a lot of merit to the "glide like a rock" fail safe strategy. While I've got the RTH+baro+angle set to an oh-crap switch on my radio, I also have the magic "fall from the sky" switch . . . the transmitter's power switch. we'll see if experimenting can find some . . . alternative.