Electrohub CG

joshuabardwell

Senior Member
Mentor
Actually . . . putting your CG in the wrong spot will make it fly equally bad in an acro or a self-level mode. More so, if the board is off the CG, but even if it's on the CG the motors won't care if the board is keeping it level or you are when two are taking it easy and the other two are struggling to keep balance.

You're right about the motors working unequally. But I don't think that having the FC off the CG matters in Acro mode, as described above, because the gyro reads the same regardless. Am I wrong?
 

Craftydan

Hostage Taker of Quads
Staff member
Moderator
Mentor
Nope. The gyros really care about the CG. Why? What do the gyros measure? Angular motion. What is the airframe twisting about? The CG.

Always, everything acts through the CG.

If you place the gyro off the CG it will still read angular motion, but the farther out you get more of that angular motion gets traded into lateral motion the gyro can't see. For instance, look at a see-saw. The closer you are to the center the more you feel the angular motion, and the less you move up and down. the farther out, the more you move up and down and the less you feel the change in angle.

In effect moving a gyro away from the CG, the poorer the picture it sees . . . and the stronger any vibration noise appears.
 

joshuabardwell

Senior Member
Mentor
If you place the gyro off the CG it will still read angular motion, but the farther out you get more of that angular motion gets traded into lateral motion the gyro can't see.

OMG! What an insight! Of course! The lateral motion comes from somewhere! Brilliant!!! The larger the radius of a circle, the fewer degrees of arc you go through per linear unit of movement!

EDIT: Nope, hang on... that's not quite right. Ten degrees is ten degrees regardless of radius. But it occurs over a longer time. Have to keep thinking...

EDIT: Nope... it has to occur over the same time, as long as the body is rigid. Still thinking...

EDIT: Sorry... I'm still not convinced. If you assume a rigid body, then when one part of the body undergoes an angular change, all other parts of the body must undergo the same angular change, over the same time. I don't see a way that this isn't true. Therefore, a gyro at any location on the (rigid) body should read the same regardless. As mentioned previously, copters are not rigid, and vibration will be worse when the FC is not at the CG, but the displacement issue should be moot. I think.
 
Last edited:

RichB

Senior Member
Nope. The gyros really care about the CG. Why? What do the gyros measure? Angular motion. What is the airframe twisting about? The CG.

Always, everything acts through the CG.

If you place the gyro off the CG it will still read angular motion, but the farther out you get more of that angular motion gets traded into lateral motion the gyro can't see. For instance, look at a see-saw. The closer you are to the center the more you feel the angular motion, and the less you move up and down. the farther out, the more you move up and down and the less you feel the change in angle.

In effect moving a gyro away from the CG, the poorer the picture it sees . . . and the stronger any vibration noise appears.

The further off the CG you are, the more lateral acceleration you will feel when rotating. This is why FC near the CG is advised for angle modes.

Rotational motion, as sensed by a gyro, is not affected by distance from the CG. There is no tradeoff.
 

joshuabardwell

Senior Member
Mentor
The further off the CG you are, the more lateral acceleration you will feel when rotating. This is why FC near the CG is advised for angle modes.

Rotational motion, as sensed by a gyro, is not affected by distance from the CG. There is no tradeoff.

I think you're agreeing with me: that FC performance in Rate/Manual/Acro mode is not affected by distance from CG, except in as much as vibration is increased. Yes?
 

JourneyFPV

Junior Member
Hello JourneyFPV,

There was a couple of things I had to do to resolve this.
One of the mistakes I made was putting the propellers on the wrong way.
I notice this when the back would lift and the front would not causing it to tilt forward. so double check they are correctly installed and your motor is turning in the correct way.

The next thing was to balance the quad by moving the battery towards the front of the electrohub until it balance. By lifting the quad with one finger on each side should lift evenly. My electrohub was back heavy when I would lift it so by moving it more towards the front it balanced it out.

The next thing I did to help is program the flip using the MultiWiiConf to auto level. I use the Aux1 to set the Angle mode. I plugged the flip to channel 5 on my receiver. Then I programed my Radio to use channel 5 for my landing gear switch to be able to turn it on or off by that switch on the radio.

As a beginner quad pilot the Angle mode really helps to control it better.

Also I was using a 4s battery in which it gave the quad a lot of power. Too much for me as a beginner. When I just give it a little power it would take off so fast and high.

One of the suggestion in the Quad for dummies video is to adjust the throttle in the MultiWiiConf called the "MID" the default was set to 0.50 I adjusted it down to 0.35.

The last thing is, the other wonderful people that replied to my post suggested not to give it slow power to take off. I agree,..
You need to give it enough power to get it to jump up in the air, about six feet or more then practice keeping it level.

Hope this helps.

Sidney

Thanks Sidney, Got it working great now it's time to HOVER!!.....and then hover some more.
 

Craftydan

Hostage Taker of Quads
Staff member
Moderator
Mentor
Hmmm . . . Pondered this for a bit . . .

Did I mention Math hurts :p

You're effectively right about the gyros. There's a little bit of acceleration cross-talk between the modern MEMs gyro and the accelration axies -- on the MTU6050 it's actually not as good as I'd expect it would be, but it's still a single-digit percentage affect . . . but those affects aside, torque is axial, not positional, so reading a torque on the frame (via correallis force) will only strongly depend on the axis. Not quite sure how I got wrapped the wrong way around that axle . . .

Effectively, orentaion is critical (or the correction gets applied in the wrong directions), Motor placement to CG is helpful to have right, and accelerometers prefer being on the CG, but the gyro doesn't care . . .

That being said . . . careless placement for an acro mode might not stay a non-issue. Some of the newer flight modes coming out now include an accelerometer stabliized rate mode -- this isn't horizon where it snaps back to level when you release the stick, but instead the mode mixes in the accelerometer's readings for absolute angle reading in with the gyro's angle rate readings to produce a far tighter acro control with no drift in angle . . . since angle is no longer being dead-reckoned, but measured directly.

Haven't seen this mode appear in any of the lower end contorllers -- I've only heard of it in the Tau Labs ROM -- but from the reports I've heard it's tight and super precise. Would be neat if it got ported down to the Naze boards.
 

JourneyFPV

Junior Member
Way to go JourneyFPV!!!

I would love to hear what resolve the issue?

Me too! Ha Ha! well my problem was (as most people I think who go down the path of building their own quad with the Electrohub and RTF parts) that I rushed into flying the thing without reading enough. Really, after I read everything on the RTF website and followed a few instructions from Crash's video Flip 1.5 for dummies it worked perfectly. Oh, and the help from you guys of course. There was a few things I had to figure out with my Radio (Frsky Taranis) like the setting of the end points to adjust the rudder range.

Turning on Angle mode and calibrating the ESC's were probably the best things.

Now I just feel kinda stupid as all my questions were there I just didn't look hard enough.

I think because I used the Flite Test Electrohub build video as my guide only I was led to believe it was plug and fly - no fault to the Flite Test guys, they are awesome.
 

cranialrectosis

Faster than a speeding face plant!
Mentor
Nicely done!

I like how you put the foam over the flight controller. :)

You will do better if you keep the air below the rotors clean. You may want to move the skids so they aren't directly below the rotors if you can help it.

Great job. Welcome to piloting your own work.
 

makattack

Winter is coming
Moderator
Mentor
Nice going JourneyFPV!

Good to see success after some setbacks. I kind of wish I had found this thread earlier. I just finished my tricopter and didn't think about cg and motor placement. This thread motivated me to search the articles base and I found this:
http://flitetest.com/articles/symmetrical-tricopter-and-t-copter-design-theory

Makes sense, but now I realize my rear motor is too far back for an optimum symmetric setup. It still flies well, at least in acro mode, but I think my flip 1.5 might be mounted too far forward for the accelerometer modes to work well. Right now, it drifts to the right in angle mode, but I suspect there are other problems once I calibrate and trim it. Oh well, it flies well in acro, so maybe I just have to leave it there.
 

joshuabardwell

Senior Member
Mentor
Don't over-think needing to have the FC directly over the CG. There is no reason why the copter shouldn't be able to hover in autolevel mode regardless. If it drifts, you just have to go through the ACC trim process. With the motors disarmed (be 100% sure of this), push the throttle stick all the way up and then move the right stick to full deflection in the direction you want to trim, and back to center. Each time you do this, the blue LED on the FC will blink one time. So if the copter is drifting right in angle mode, you would push the right stick to the left two or three times. The put the throttle to zero, arm the motors, take off, and see how it hovers.

Before you do that, make sure that your transmitter trims and subtrims are zeroed and that the channels show approximately 1495 to 1505 (ideal is dead-center 1500) in the MW Config GUI. If your trims are off, this will cause drift in Angle mode. But it will also cause drift in Acro mode, so you're probably okay there if it flies well in Acro mode.

If the FC is not over the CG, what will happen is that you will get some coupling of inputs. This will manifest as the copter tending to pitch or roll whenever you yaw. But given the size of our copters, it is almost certain that you are within a few inches of optimal placement, pretty much no matter where you put the FC, so this effect ought to be small, and any undesirable effects are probably due to other factors.
 

makattack

Winter is coming
Moderator
Mentor
As usual, some great advice from JoshuaBardwell. I do get the sense I'm over-thinking/getting OCD about this new multirotor stuff. Agreed. It flies stable in acro mode so far. I just need to calibrate the ACCs and trim for angle/horizon mode in case I need to switch into them... otherwise, I'll learn to fly in acro in a progressive / low-risk rate.

BTW, fascinated by engineerguy's youtube channel: https://www.youtube.com/channel/UC2bkHVIDjXS7sgrgjFtzOXQ -- also thanks to Joshua.
 

joshuabardwell

Senior Member
Mentor
Thanks for the kudos. I thought of one more thing. With a quad, and especially if you have larger props (like 10" instead of 8" or 9") there will usually be some coupling of pitch/roll to yaw, regardless. This is because the copter must yaw by spinning up one pair of motors and spinning down the other pair. The motors spin up as fast as the ESC can push them, but they only spin down as fast as friction can slow them down. The net result is that the force from the motors spinning up is not exactly equal to that of the motors spinning down, and the quad gets somewhat disturbed. Also, the motors spinning up may not spin up at exactly the same rate, and depending on the PID gains, the FC may not compensate quickly enough. Bottom line: if, when you yaw aggressively, your quad becomes somewhat disturbed in the pitch and roll axes, this doesn't necessarily indicate a problem with your FC placement. You may just be experiencing one of the reasons why some people prefer V-tail quads or tri's.

If you do have a problem with your FC placement, you should only experience it in autolevel modes. With a rigid body, all parts of the body will experience the same angular change over the same time, so the gyro should be unaffected by placement on the body. Since the Acro mode only uses the gyro, FC placement shouldn't affect performance in Acro mode. In autolevel mode, FC placement off of the CG will result in pitch/roll coupling to yaw input, as previously mentioned.

A final thing: when I say that FC placement doesn't affect performance in acro mode, that is ignoring the effect of vibration. The CG of the copter is where vibrations will be minimized. The further from the CG your FC is, the worse vibration it will experience. But in terms of "false input" (e.g. yaw coupling to pitch/roll), the gyro is not affected in that way.
 

Craftydan

Hostage Taker of Quads
Staff member
Moderator
Mentor
The motors spin up as fast as the ESC can push them, but they only spin down as fast as friction can slow them down.

Heh, Heh, Heh . . . That doesn't have to be true. (it usually is, though).

And you don't *have* to spring for the pricy KISS ESCs to do it (but by al repots they're worth the cost).

There's quite a few ESCs capable of Regenerative breaking or Damped control -- same difference: the ESC is actively slowing the prop. It can make a HEAUGE diffence for the reasons you've stated.

the downside . . . you need to flash a ROM capable of it on an ESC capable of it. It's not activated in any stock simonK or BLHeli (simonK will need a rebuild of the ROM, where BLHeli it's a config option if the ESC can do it) so it's something you have to *know* how to turn it on, and indications of wheterh a particular ESC is capable can be hard to find.
 

joshuabardwell

Senior Member
Mentor
There's quite a few ESCs capable of Regenerative breaking or Damped control -- same difference: the ESC is actively slowing the prop. It can make a HEAUGE diffence for the reasons you've stated.

the downside . . . you need to flash a ROM capable of it on an ESC capable of it. It's not activated in any stock simonK or BLHeli (simonK will need a rebuild of the ROM, where BLHeli it's a config option if the ESC can do it) so it's something you have to *know* how to turn it on, and indications of wheterh a particular ESC is capable can be hard to find.

Ah, yeah. I left that out for simplicity. Do the KISS ESC's come with firmware to make the feature active? I thought they were the only ones that could do it and didn't realize anybody else could even do it.

Is it true that the ESC basically charges the battery in regenerative mode?
 

Craftydan

Hostage Taker of Quads
Staff member
Moderator
Mentor
Depends on how it's done, but if they're wired correctly, yeah.

It will, if the back EMF is allowed to pass current back throgh the transistors into the battery, draining the energy and braking the prop.

The back EMF can also be routed into the other legs, casuing one motor winding to inject power into braking on another winding. while that forces the prop to a stop, you bleed off the fly-wheel energy in the prop into stopping the prop, instead of bleading it off into the battery. It works, but responsiveness is all you gain from it -- no small thing.

As far as what's default on the KISS ESCs . . . dunno. Never played with them to know what setup is required.
 

cranialrectosis

Faster than a speeding face plant!
Mentor
Regen braking is on by default on KISS. The motors are quieter on KISS and my flight times went up by <> 1 minute on KISS.