My first quad was HMB235 with 1806/dys 20a (simonK)/naze32/betaflight. I never flashed the esc's and everything seems to have worked ok - but i'd suspect i don't know what i'm missing. I actually assumed i'd be able to flash esc's through the F3 board like the naze, and didn't realize some of the F3 boards weren't able to do this.
Under 250, Under 250, Under 250!!!
- Thread starter Craftydan
- Start date
In most cases the default ESC flash for an ESC sold for multirotors is fair to good -- not particularly holding the airframe back from doing well, but might be holding it back from becoming excellent. Most people are happy with "doing well", so it's generally not a big deal.
In this build, however, I'm recommending it for 3 reasons:
- Rotation -- reflashing in BLHeli will allow me to make up for the fact all 4 ESCs are wired the same. The motors need to be reversed on two arms, and it's a simple task while reflashing. It can be done via stick commands, but if you pick the third option (reversible), stick programming locks up and the ESC MUST be reflashed to recover -- a design hole, IMO, which leads me to . . .
- Turning off stick programming -- if you can easily reflash, stick programming will only create problems. It needs to go off.
- Tightening -- if you went for the littlebees, then why aren't you using damped light (active braking)? Because it's not the default in the flashed ROM. Since these motors are such high kV, they are running a bit on jittery side as it is. Being able to tighten the control, move to a slightly better ROM, and turn on advanced features the ESC supports will make a difference on the overall tune. This is a "well-to-excellent" move, but it's also making the best of what you've paid for.
BTW, as for your question about 4S . . . that's one of the reasons I picked these motors
Keep in mind, you will have to drop to a shorter prop, and a 4S pack will probably break the 250g weight, but from the STATIC thrust tests you're over the 5:1 ratio at 15% less current.
It's not particularly confidence building to find 3S on a 4" prop performs just as well (albeit much closer to max current) as 4S on a 3" prop, but there still is a bit of wiggle room. Is the standard 4045 prop near ideal for 3S, where the 3030BN is a bit off for 4S? Will that difference separate a bit more as you pick up enough airflow for the prop to work efficiently? Is there enough margin on max current on 4S for a 3045 or 3045BN? Dunno . . . not sure if I'll hunt out those answers in this project, but the airframe *can* do it, so the option remains open for it
Didn't get to work on her last night, but did get to look at some diagrams and poke at the part layouts around the broken trace. I've come to the opinion that whoever laid out this mini Naze board was doing it by hand, and did so . . . inefficiently. Absolutely space saving, so kudos for that, but odd, for sure. There are multiple paths for a few of the same traces (to my good fortune, the trace I pulled up is one of them). The shape of the traces has an odd organic hand-drawn feel to it . . . between the un-optimized routing (even the most basic of net routing tools would have cleaned out the looped paths to reduce the traces and vias) and the wavy, groovy paths taken, this has to have been done by hand.
Not complaining -- It's paying off for me in this case -- this torn trace should have killed all power to either the main processor, the IMU or at least two of the LEDs, but I've plugged it in and everything on the 3v side still works So far all it's done is ripped out a bootloader pad, which I was about to cover in liquid electrician's tape anyways . . . we'll see if I can get some time to solder in the jumper-wire tonight, connect up the Sat, and I'll show you what I mean . . .
In this build, however, I'm recommending it for 3 reasons:
- Rotation -- reflashing in BLHeli will allow me to make up for the fact all 4 ESCs are wired the same. The motors need to be reversed on two arms, and it's a simple task while reflashing. It can be done via stick commands, but if you pick the third option (reversible), stick programming locks up and the ESC MUST be reflashed to recover -- a design hole, IMO, which leads me to . . .
- Turning off stick programming -- if you can easily reflash, stick programming will only create problems. It needs to go off.
- Tightening -- if you went for the littlebees, then why aren't you using damped light (active braking)? Because it's not the default in the flashed ROM. Since these motors are such high kV, they are running a bit on jittery side as it is. Being able to tighten the control, move to a slightly better ROM, and turn on advanced features the ESC supports will make a difference on the overall tune. This is a "well-to-excellent" move, but it's also making the best of what you've paid for.
BTW, as for your question about 4S . . . that's one of the reasons I picked these motors
Keep in mind, you will have to drop to a shorter prop, and a 4S pack will probably break the 250g weight, but from the STATIC thrust tests you're over the 5:1 ratio at 15% less current.
It's not particularly confidence building to find 3S on a 4" prop performs just as well (albeit much closer to max current) as 4S on a 3" prop, but there still is a bit of wiggle room. Is the standard 4045 prop near ideal for 3S, where the 3030BN is a bit off for 4S? Will that difference separate a bit more as you pick up enough airflow for the prop to work efficiently? Is there enough margin on max current on 4S for a 3045 or 3045BN? Dunno . . . not sure if I'll hunt out those answers in this project, but the airframe *can* do it, so the option remains open for it
Didn't get to work on her last night, but did get to look at some diagrams and poke at the part layouts around the broken trace. I've come to the opinion that whoever laid out this mini Naze board was doing it by hand, and did so . . . inefficiently. Absolutely space saving, so kudos for that, but odd, for sure. There are multiple paths for a few of the same traces (to my good fortune, the trace I pulled up is one of them). The shape of the traces has an odd organic hand-drawn feel to it . . . between the un-optimized routing (even the most basic of net routing tools would have cleaned out the looped paths to reduce the traces and vias) and the wavy, groovy paths taken, this has to have been done by hand.
Not complaining -- It's paying off for me in this case -- this torn trace should have killed all power to either the main processor, the IMU or at least two of the LEDs, but I've plugged it in and everything on the 3v side still works
So far all it's done is ripped out a bootloader pad, which I was about to cover in liquid electrician's tape anyways . . . we'll see if I can get some time to solder in the jumper-wire tonight, connect up the Sat, and I'll show you what I mean . . .
I'm sure you don't need this reminder Dan...but for others reading along. Don't forget that if you disable stick programming on the ESC's you can't calibrate the ESC's either since the calibration uses stick commands. So after disabling stick commands you'll have to manually set your low/high points through blheli to match what you have configured on the Naze.
Looking forward to seeing these little guys in the air. Really want to make something similar for myself but can't quite justify it If I could convince myself it's small enough I can fly FPV in my backyard...I'd go for it ASAP. But I'm not at all convinced I have enough room on my 1/4 acre for that.
Looking forward to seeing these little guys in the air. Really want to make something similar for myself but can't quite justify it
If I could convince myself it's small enough I can fly FPV in my backyard...I'd go for it ASAP. But I'm not at all convinced I have enough room on my 1/4 acre for that.
Cyberdactyl
Misfit Multirotor Monkey
Beautiful progress Dan. Very well thought out and implemented.
The same ol' same ol' grumpy Cyberdactyl comment I would have for these small racers, is they tend to keep the boom plate close to, or as wide as, their bigger brothers. That inevitably hurts these little guys three fold. Because of the much smaller prop, the boom-width-area/prop-disk-area ratio is even closer. Also the thrust column velocity is higher, enhancing the already higher drag. And last, the prop is a good 4-8mm closer to the boom, causing even more pronounced splash.
I suspect the reason we are not seeing robust, extra thick, yet thin in planform sub 250's, is because the designer is stuck with a frame/boom uni-plate. At these sizes, individual booms, which fasteners are necessary, adds too much weight. So main body thickness and boom thickness has to be a compromise.
I suspect, at this size, vertical booms would add a good 10-15% extra thrust.
The same ol' same ol' grumpy Cyberdactyl comment I would have for these small racers, is they tend to keep the boom plate close to, or as wide as, their bigger brothers. That inevitably hurts these little guys three fold. Because of the much smaller prop, the boom-width-area/prop-disk-area ratio is even closer. Also the thrust column velocity is higher, enhancing the already higher drag. And last, the prop is a good 4-8mm closer to the boom, causing even more pronounced splash.
I suspect the reason we are not seeing robust, extra thick, yet thin in planform sub 250's, is because the designer is stuck with a frame/boom uni-plate. At these sizes, individual booms, which fasteners are necessary, adds too much weight. So main body thickness and boom thickness has to be a compromise.
I suspect, at this size, vertical booms would add a good 10-15% extra thrust.
heh . . . Reminds me of a recent quote from Timecop: "Who the f**k calibrates escs in 2015?"
Who indeed? Everyone who wants to eek past "flies OK", but for those who can't easily flash, they have to use the radio/control board. Plenty of reasons that re-flashing ESCs would be beyond trivial, but most of those depend on the hardware selected. This gear can do it -- it was carefully selected so it could -- so I'll be going the better route by "calibrating" it by setting the values with the flash tool.
Adam, Patience. The slow boat comes at it's own maddeningly slow rate (Gives me time to keep ahead of you in the log anyways)
Cyberdactyl, The arms are thinner than a comparable 250, but I'm sure you're right the percentage of column lost is higher. Vertical booms would be nice, but the added weight would steal a lot from it. I like the concept of vertical booms, but there's just no getting around the added complexity
Next two sections are all done but the write-up . . . Jhitesma, I dunno how you do it Finished up too late last night, so I'll write up one tonight.
Who indeed? Everyone who wants to eek past "flies OK", but for those who can't easily flash, they have to use the radio/control board. Plenty of reasons that re-flashing ESCs would be beyond trivial, but most of those depend on the hardware selected. This gear can do it -- it was carefully selected so it could -- so I'll be going the better route by "calibrating" it by setting the values with the flash tool.
Adam, Patience. The slow boat comes at it's own maddeningly slow rate
(Gives me time to keep ahead of you in the log anyways)Cyberdactyl, The arms are thinner than a comparable 250, but I'm sure you're right the percentage of column lost is higher. Vertical booms would be nice, but the added weight would steal a lot from it. I like the concept of vertical booms, but there's just no getting around the added complexity
Next two sections are all done but the write-up . . . Jhitesma, I dunno how you do it
Finished up too late last night, so I'll write up one tonight.
Next two sections are all done but the write-up . . . Jhitesma, I dunno how you do it
I've found that learning to type in my sleep helped a lot
PsyBorg
Wake up! Time to fly!
I see some discussion over thrust and the arms interfering. Have you thought about possibly mounting the motors on the underside like I have seen on several quads? Or possibly flipping the whole build putting the arms as a top plate and have them push up instead of pull up the quad. To me that would have a dual effect. It would get the arms out of the prop wash causing far less air disturbance AND it would lower the center of gravity. Now I'm not sure how that would effect high speed turning but I imagine straight runs would benefit as well as flips I would think.
Cyberdactyl
Misfit Multirotor Monkey
Cyberdactyl, The arms are thinner than a comparable 250, but I'm sure you're right the percentage of column lost is higher. Vertical booms would be nice, but the added weight would steal a lot from it. I like the concept of vertical booms, but there's just no getting around the added complexity
Yea, for cut-plate designs, I can't deny it's a predicament, especially when the size is under 280.
That's an interesting idea, but I think might have similar inefficiencies that affect the top/bottom motor configurations in that the air coming to the prop is still disturbed / not smooth. I suspect it'll be sort of like when a multirotor is descending in it's own wash, but maybe not as bad.
Cyberdactyl
Misfit Multirotor Monkey
Blocking flow is blocking flow.
That said, there might be an optimum configuration for much reduced drag for a certain cross-prop disk area. For example, if there was a design that separated the boom from the prop, either over, or under, by two or more prop diameters, the drag might drop considerably. The catch of course, would be the added weight of such a gangly and awkward structure, not to mention the added drag in other aspects of flight.
That said, there might be an optimum configuration for much reduced drag for a certain cross-prop disk area. For example, if there was a design that separated the boom from the prop, either over, or under, by two or more prop diameters, the drag might drop considerably. The catch of course, would be the added weight of such a gangly and awkward structure, not to mention the added drag in other aspects of flight.
PsyBorg
Wake up! Time to fly!
I know there will still be effects from pulling air around a boom but at least you wouldn't have the air bouncing off them fighting back. Kinda like a zone oriented ground effect? Im new to all this so don't take me seriously. Just trying to think outside normal parameters as I have all my life. Once in a while I can pop out a viable thought.. maybe not this time but it has happened more then once over time.
Either way its good to have discussions on all levels as learning is learning whether from success or fail.
Ok, so dragging this thread back on topic (hey, it's not like I've had my share of dragging threads off) . . .
We last left off with a tested power deck so . . .
. . . We all need a little more control in life
and I mean little:
This tiny guy has most of the features of it's bigger brother, minus a lot of weight
and minus attachment points too . . . tricky. we'll use this side, keeping everything flush with these connectors.
Now flip it over and test fit again:
Perfect. plenty of space in front of it for the RX.
Speaking of receiving . . .
Now one of the issues with this particular board is it's designed to work with ppm and PWM, but no accommodations for a Sat or S.Bus port. What's not stated is this is a native Naze32 clone, which means the RX inputs naturally remap to Sat and S.Bus inputs, the only thing missing is 3v power for the sat. Never fear, I have a trick up my sleeve . . .
. . . the bootloader pin doesn't short to ground. It shorts to 3v. find that pad, and I've got the power I need!
a bit of poking-and-prodding and there it is. this pad is shorted to the STM32F103 chip's power pins *and* the output pin on the voltage regulator over in the far corner. so I add on a milled socket for the sat's power connection . . .
And now for Spaghetti . . .
. . . and I'm set for the RX. Now to wire in the ESC cables before I mount the board . . . no snese in letting the wires I extracted form the ESCs go to waste, let's use those -- wire in #1:
through #4
now here is where I found 1-4 didn't lay out well for which direction the cables went. no worries. just lay them out around each other. best way I found for these was swapping the middle 2 first (2 & 3), then overlaying the swapped outer two over top of those (1 & 4).
Add on a lead for power (I used some more scrap of the ESC wires . . . I've got a lot left over )
I can tell you where to stick it . . .
. . . and now it's time to consider mounting. I'm going to go with mounting tape, but I need to make sure the lights are unobstructed. there are 3 LEDs on this board, one in the bottom left corner, and two just forward of the white connectors.
so cut a piece of foam mounting tape too big, and make sure the LEDs aren't covered . . .
. . . then cut it away from the connector
. . . and repeat -- took 3 layers to get a full layer above the connector.
Then shave the foam down to a flat surface, level to thee connectors.
and that's a nice flat surface, ready to mount.
add on one more layer of extreme mounting tape (good stuff!), and . . .
Stick it on!
So now it looks more like Cthulhu than a control board . . . but we'll trim some tentacles before we're through.
Whew, whata ride!!!!
yeah, tonight's was a long one . . . split this in half, so it should be more readable . . .
next we'll install the RX, deal with repairing evil pads and make a first pass at setting up the board
We last left off with a tested power deck so . . .
. . . We all need a little more control in life
and I mean little:
This tiny guy has most of the features of it's bigger brother, minus a lot of weight
and minus attachment points too . . . tricky. we'll use this side, keeping everything flush with these connectors.
Now flip it over and test fit again:
Perfect. plenty of space in front of it for the RX.
Speaking of receiving . . .
Now one of the issues with this particular board is it's designed to work with ppm and PWM, but no accommodations for a Sat or S.Bus port. What's not stated is this is a native Naze32 clone, which means the RX inputs naturally remap to Sat and S.Bus inputs, the only thing missing is 3v power for the sat. Never fear, I have a trick up my sleeve . . .
. . . the bootloader pin doesn't short to ground. It shorts to 3v. find that pad, and I've got the power I need!
a bit of poking-and-prodding and there it is. this pad is shorted to the STM32F103 chip's power pins *and* the output pin on the voltage regulator over in the far corner. so I add on a milled socket for the sat's power connection . . .
And now for Spaghetti . . .
. . . and I'm set for the RX. Now to wire in the ESC cables before I mount the board . . . no snese in letting the wires I extracted form the ESCs go to waste, let's use those -- wire in #1:
through #4
now here is where I found 1-4 didn't lay out well for which direction the cables went. no worries. just lay them out around each other. best way I found for these was swapping the middle 2 first (2 & 3), then overlaying the swapped outer two over top of those (1 & 4).
Add on a lead for power (I used some more scrap of the ESC wires . . . I've got a lot left over
) 
I can tell you where to stick it . . .
. . . and now it's time to consider mounting. I'm going to go with mounting tape, but I need to make sure the lights are unobstructed. there are 3 LEDs on this board, one in the bottom left corner, and two just forward of the white connectors.
so cut a piece of foam mounting tape too big, and make sure the LEDs aren't covered . . .
. . . then cut it away from the connector
. . . and repeat -- took 3 layers to get a full layer above the connector.
Then shave the foam down to a flat surface, level to thee connectors.
and that's a nice flat surface, ready to mount.
add on one more layer of extreme mounting tape (good stuff!), and . . .
Stick it on!
So now it looks more like Cthulhu than a control board . . . but we'll trim some tentacles before we're through.
Whew, whata ride!!!!
yeah, tonight's was a long one . . . split this in half, so it should be more readable . . .
next we'll install the RX, deal with repairing evil pads and make a first pass at setting up the board
Last edited:
Looking good. How well does that connector stay seated with only two wires in place? I did the same on my Nighthawk 280 Pro (though on there I still have 3 wires connected) and ended up putting a bit of tape over the connector because it was so loose I was worried about it coming loose in aggressive flight. I had always assumed the housing provided most of the physical connection on those connectors but apparently the pins do more of the work than I gave them credit for.
Ok, yeah . . . TL;DR. In the light of day, I need to split these. I'll break it at the RX -- not half-way but a clean spot to separate them. should be able to play some games so it's just copy and paste, but I might have to futz with attachments . . . again.
Edit: Hmmm . . . not so easy I'll have to clean this up tonight.
Jihitesma, they hold pretty well, but I suppose a lot of it can be up to the tolerance of the molds. Haven't tried it without any pins holding, which would end up being the acid-test for plug v. pin, but 2 pins hold reasonably tight. in either case, both sides this harness is connecting to are hard-mounted to the frame -- even if they were just a touch loose they should stay connected. Not quite true for the RX power pin, but that's why it got a drop of liquid Electricians tape to "bridge the gap". Still removable, but not accidentally.
Edit: Hmmm . . . not so easy
I'll have to clean this up tonight.Jihitesma, they hold pretty well, but I suppose a lot of it can be up to the tolerance of the molds. Haven't tried it without any pins holding, which would end up being the acid-test for plug v. pin, but 2 pins hold reasonably tight. in either case, both sides this harness is connecting to are hard-mounted to the frame -- even if they were just a touch loose they should stay connected. Not quite true for the RX power pin, but that's why it got a drop of liquid Electricians tape to "bridge the gap". Still removable, but not accidentally.
Last edited:
So moving on, we've got the Naze mounted ready to give flight . . .
. . . but sometimes it's good to receive too!
So on to the Sat RX. first things first . . . hard plastic shell?!? this is a lightweight build! that's gotta go.
Remove one micro screw and *POP* out of the shell it comes.
huh . . . just like shellfish . . . looks kinda fragile. let's put on a coat . . .
so I pull out some 1/2" shrink-tube, cut to length, and to prep, I poke two holes just a 1/8" form the top for the antennas.
Pull the antennas through . . .
. . . and shrink.
add on some extreme mounting tape and feed the antennas through some holes in the CF frame -- need to get them clear so they can hear what's going on outside.
So now that it's mounted, let's add to the spaghetti . . . wait . . . let's subtract a bit first.
On this connector (the same as the Naze32 breakout connector) we only need ground and RX #4 (why 4? because that's the one that UART2's RX line is mapped to, which we'll set that port in cleanflight to be the sat RX).
so we're down to 2 wires . . . but the RX needs 3.
That's what that power pin we installed is for. I soldered on a red wire to it and then routed the RX cable and trimmed it short:
trim strip tin and solder . . .
. . . and the RX is connected
Splitting once more . . .
man this one was too long
Next up . . . Disaster strikes once again . . .
. . . but sometimes it's good to receive too!
So on to the Sat RX. first things first . . . hard plastic shell?!? this is a lightweight build! that's gotta go.
Remove one micro screw and *POP* out of the shell it comes.
huh . . . just like shellfish . . . looks kinda fragile. let's put on a coat . . .
so I pull out some 1/2" shrink-tube, cut to length, and to prep, I poke two holes just a 1/8" form the top for the antennas.
Pull the antennas through . . .
. . . and shrink.
add on some extreme mounting tape and feed the antennas through some holes in the CF frame -- need to get them clear so they can hear what's going on outside.
So now that it's mounted, let's add to the spaghetti . . . wait . . . let's subtract a bit first.
On this connector (the same as the Naze32 breakout connector) we only need ground and RX #4 (why 4? because that's the one that UART2's RX line is mapped to, which we'll set that port in cleanflight to be the sat RX).
so we're down to 2 wires . . . but the RX needs 3.
That's what that power pin we installed is for. I soldered on a red wire to it and then routed the RX cable and trimmed it short:
trim strip tin and solder . . .
. . . and the RX is connected
Splitting once more . . .
man this one was too long
Next up . . . Disaster strikes once again . . .
Last edited:
um . . . I'm not sold on mine at the moment. I'm tracking an overheating issue that I haven't licked yet . . . or given up on
I've got a couple of options left, but need to explore them first before I declare "Junk" and move on or declare "cheap but workable if . . .", and layout what needs to be fixed. More on that later . . . but not much later.
EDIT: Wait, what? antenna? right. sorry. caught me speed-reading again
For the moment, I'm sticking with the naked rubber ducky. Not a high end solution, to be sure, but functional, light, modestly well protected (hard to damage a straight floppy wire) . . . and cheap. it'll be more than enough for use, and that leaves it as an upgrade to the new owner's taste.
Mine will eventually get a cheap light-weight cloverleaf or skew-planar soldered into the ducky harness . . . but even the vertical pol antenna will perform acceptably well, just at 60% full range. I can live with that for now.
I've got a couple of options left, but need to explore them first before I declare "Junk" and move on or declare "cheap but workable if . . .", and layout what needs to be fixed. More on that later . . . but not much later.
EDIT: Wait, what? antenna? right. sorry. caught me speed-reading again
For the moment, I'm sticking with the naked rubber ducky. Not a high end solution, to be sure, but functional, light, modestly well protected (hard to damage a straight floppy wire) . . . and cheap. it'll be more than enough for use, and that leaves it as an upgrade to the new owner's taste.
Mine will eventually get a cheap light-weight cloverleaf or skew-planar soldered into the ducky harness . . . but even the vertical pol antenna will perform acceptably well, just at 60% full range. I can live with that for now.
Last edited:
Similar threads
