CAL Multirotor Blog

When you have as much "stuff" as I do, packed in (mostly) category-labeled shoeboxes, you forget everything you have.  This evening, I remembered that when I built my last gimbal (from scratch), I bought an extra motor as a spare.  That motor, while not identical to the defective motor in the gimbal, is close enough to fit.  Anyway, the gimbal motor that I ordered from Hobby King is the same type, so even that one would not be an identical replacement and would require the same modification.

So I did some measuring and drilling and Voila!, I now have a working gimbal. Now the motor coming from Hobby King will be a spare, and if I get a replacement from the Chinese guy that sent me the defective gimbal in the first place, I'll have two spares.

All good!

I believe the danger of the hobby LiPo batteries is underestimated.  The batteries in laptops and other consumer equipment is not designed for very high currents (discharge rates), but the batteries used in electric-power models are.  When you see a rating of "30C", it means that the batteries can output 30 times the amp-hour rating (which is milliamp/hour divided by 1000).  So, if you have a 3000mAH battery, with a 20C rating, that means it can output 3A X 20 = 60Amps on a short-term basis - but that is their LOAD rating.  They can output several HUNDRED amps during a short-circuit condition.

Since commercial batteries (like for laptops and phones) are not required to produce very high output currents, the cells are designed differently and they usually cannot deliver huge currents - even during short-circuits. The cells are usually small (a bit bigger than AA size) and are arranged in a series-parallel arrangement to produce the required current and voltage.  The fact that each cell is not too large means that a catastrophic failure in one cell is not likely to produce enough heat to cause serious damage.  Also, laptop and phone batteries have a fusible link that blows if a cell shorts or the temperature gets too high. This keeps the energy from the adjacent cells from feeding power INTO a shorted or overheated cell, further decreasing the risk of fire or explosion.

Hobby batteries have NONE of the protection mechanisms mentioned above. The cells are usually put in series only, which means that every cell must put out 100% of the output current.  This requires a large cell. There are no fuses, no temperature limits, no "smarts" of any kind.

A charge cycle is especially dangerous. The material inside the battery actually expands with time - which is why older batteries get that "pillow" look.  Sometimes the insulator which separates the (+) and (-) plates inside the cell breaks down and the cell shorts. At that time, the full output current capability of the cell appears across its internal resistance.  My Turnigy charger has an internal resistance tester, and I find that most cells have an internal resistance of  0.010 ohms (10 milli-ohms).  If a cell has a voltage of 4V, then 4V across 10 milliohms = 400A.  400A at 4V = 1600 Watts!!  That heats the battery up within seconds. Note that this sequence occurs even if the charger stops immediately when the short is detected!  There is NO WAY to stop this.

And let us not forget that lithium (and each cell has a lot of it), is reactive.  It burns!  The end result is a very, very nasty fire with flames shooting out of the battery for several feet.  Anything nearby WILL catch fire if it is at all flammable.

But is this likely?  YES!!!  I have had this happen TWICE.  I was not charging in an ammo can, but this morning I had the battery on the charger on my concrete garage floor.  I heard a POP, and within seconds, flames shot out of the battery by more than 3 feet!  The bottom line is NEVER, NEVER, NEVER charge a battery in anything that is flammable - or within 10 feet of anything that is flammable.  I normally charge batteries in a metal "AMMO" can, and I'm going to get a second where I store charged batteries. I'm not certain that a battery sitting by itself, not on the charger will not do the same thing.

I do know that if I charge a battery in an ammo can with a bunch of other batteries, ALL of them will be destroyed if one goes up in flames. I don't know if the lack of oxygen will help quench the flames or not - but I'm not going to test any theories.  As far as I'm concerned, the only safe approach is to charge in one ammo box and store in another.

One of my bigger quads, and my little 150mm copter had a really high-pitched squeal at certain RPMs of the motor.  At first, I thought it was the bearings, so I changed the bearings in the motor that made the most noise, but the noise persisted!

I put an oscilloscope on the motor leads and found that the frequency the ESC fed to the motor was around 3Khz.  But that is a square wave, meaning that there would be substantial harmonic power at the 3rd harmonic - 9KHz.  I think that is the squeal that I heard.  I can still hear 9Khz if it is very, very loud.  I can imagine what some youngster would think! 

So why do some motors produce noise at high audio frequencies, while some do not?  Is it the motor "bells". I put some silicone seal (a pretty good sound dampener) on the outside of the motor and did my best to balance the weight to see if that made a difference.  It did not. If it wasn't the motor bell, what was it?

So I took the motor apart and figured out what was going on - the windings were loose!!!  The copper wires in the coils themselves were moving ever so slightly back and forth in response to the driving current and making the noise.  I realized immediately that this is a VERY bad situation, since the wires will constantly bend back and forth (even if only a fraction of a thousandth of an inch) and eventually work-harden and break.  That could put your expensive toy on the ground in a hurry.

I wanted to hold the wires in place, and thought epoxy might be a good choice, but I didn't want to block the space between the windings - that is needed for cooling airflow.  I was looking for something really thin that would adhere to the wires via capillary action.  I found something - superglue  (ethyl cyanoacrilate).  It worked perfectly when dripped over the top of the windings. It wet the wires and embeded itself into them, then hardended.  The only downside is that 150C is the maximum temperature. I don't know if the insides of my motors ever get that hot, but maybe they do. 

So, after super-gluing the coils, I put the motor back together and let the glue harden for an hour or so.  Then I powered it up.

NO SQUEAL!!

So far, so good. And I'm looking for some suitable glue that has a higher temperature rating.  If anyone knows of one, I'd like to hear from you.

Next summer, my wife and I are considering hiking in the Swiss Alps.  

Over the past XMas, we went to the Palm Springs/Palm Desert  area and I was drooling over the thought of flying my toys in all that open space - but I had no aircraft with me because nothing else would fit in the car. I thought it would be great if I could take a 'copter with me on the European hike - but I can't imagine taking anything as large as a Cheerson CX-20 (or a DJI Phantom) hiking.  Besides, by the time you load up one of those quads with a camera, gimbal and all the extras, and then try to fly at a base altitude of 9,000 feet, they just don't have much 'go'.

So, while I'm waiting for new motors to arrive for the "Santa Cruz Project" (from Hobby King), I decided to rekindle my mechanical engineering skills and design a quad that could lift everything I wanted (at altitude), have good battery life, and be small enough to carry in a backpack. I have most of it planned out.  The design - as it stands, can carry an APM-style controller with GPS, Lidar Lite laser rangefinder, full 900Mhz telemetry, 2W 5.8GHz FPV video transmitter and 2-axis gimbal  capable of handling a GoPro.  It can carry up to a 5000mAH 3-cell LiPo battery and should have a flight time of about 25 minutes.  

The 1100KV motors spin 9" propellers. Calculated all-up weight is about  2 lbs (battery included).

The frame is aluminum and carbon fiber and is extremely strong, rigid and light.  The overall (flight ready) size is 12" wide X 13.5" long X  6.5" high.  The gimbal is mounted far enough forward and low enough such that that the props will only come into view if the GoPro is set to the max wide angle and the craft is pitched more than 30 degrees forward.  If the GoPro is set for a slightly less wide angle, the props will never come in to view.  

But... the design is such that it will easily fold down to a package that is 8" wide X 10" long X 5" high.  It should take about 60 seconds to fold/unfold the unit (no tools required).  And it will easily fit into a backpack!

I'm seriously considering building and selling these "Little Giant" quads.  Let me know if you are interested.  No price has been set yet, but it will be less than a DJI Phantom with camera and gimbal - and the Little Giant will certainly have higher performance, longer flight times and be much more customizable.

When I realized that the SunnySky 380KV motors were not powerful enough to power the Santa Cruz Project, I ordered one of each of 3 other motors.  That way, I could test them without spending a fortune.

One of the motors I ordered was the Turnigy AeroDrive 4240 - 530KV from Hobby King.  When I first looked at this motor, I noticed that it was big - bigger than I had imagined.

It wasn't too heavy, and it looked nice.  Best of all, it looked to have much better construction quality than most of the hobby motors I have bought.  The bearings were very tight and smooth. The windings looked good.  

So, I bolted on a 16" X 5.5" propeller and screwed it to one of the frames I have.  I connected to a 40A Turnigy PLUSH ESC, which can handle batteries up to 6-Cell.  I have a 5000mAH 4Cell battery (on bigger machines, I like to go with 4 Cell = 16.8V  batteries because for the same amount of power, the current is less and all the wiring has less loss). 

I used a "Servo Tester", an invaluable little accessory.  It is about 1" X 1", has a knob on the top and can be connected to any ESC and send the proper signals.  The knob allows you to adjust the PWM from 1ms to 2ms - full range.

So I held onto the frame tightly, and slowly turned the knob on the servo tester.  The motor started winding up.  There was very little motor "whine" - only the sound of the big 16" prop slapping the air.  I turned it up more and was astonished at how little throttle I had to give it before it REALLY started to pull.  So I connected it to my digital fish scale and put a DC current probe on the ESC input wire to measure the current.  By the time I got to 10 Amps, the motor/prop combo was pulling about 2lbs. I turned the knob more and the motor just kept winding up.  By the time I got to 30 Amps, the scale read 6lbs 5oz !!!  At 40 Amps, it was 6lbs 13oz!  4 of those could lift 27 LBS !  

40A at 16V is 640 Watts.  Since there are 745 Watts in one Horsepower, the little motor was probably putting out a little more than 3/4 HORSEPOWER!!! I admit that it could not do this forever, but I ran it at 30A for several minutes and it barely got warm. Best of all, it had a 4 ounces more pull when drawing 10A than my 28mm X 16mm NTM 800KV did (the NTM was spinning a 12" prop).

I ordered 3 more.  They should be here by next Monday.

Just to think... If I powered it with several 4-cell batteries in parallel (I would need 120 Amps),  I could probably lift 14 pounds of payload with a quad - and still fly decently.

I got the new motors in, and bolted them to my Santa Cruz 'test' chassis.  The all up weight with a 5000mAH 4-Cell LIPO is 7.4 lbs.

When I was testing it out, I had an 8 lb weight tied to it. Before I was 3/4 throttle, it was trying to lift the weight! This thing is as strong as I thought it would be.

So I took it to a local park to give it a 'spin'.  It was not easy to fly.  The PID parameters of the controller were set for something with much less power.  I suppose the experience could best be described as having a Toyota Yaris driver take a spin in a Bugatti Veyron.

The big quad was very 'touchy' and wanted to climb at the slightest provocation. Still, I'm skilled enough (now) to fly it around without crashing.  One thing that I noticed - as powerful as this craft was, the combination of the inertia of the over 7 pound weight and the inertia of the big 16" props gives it noticeable throttle lag, kinda like turbo lag on a car. .  When it is hurtling toward the ground from some high altitude, you had better hit the throttle hard before it gets below 20-30' or you will have a pile of junk very, very soon. 

But when it gets going !!!!!  

I flew for 20 minutes and had quite a bit of battery left - YAY!  I'm on the right track.

Below is a picture, along with a 12" hammer so you can judge the scale.

I have 4 camera gimbals, and I'm building a 5th.  Most of the motors have a common problem:

"Normal" brushless motors have a few turns of wire on each core.  For current-carrying reasons, the wire has to be fairly large diameter.  This makes it relatively strong.

Gimbal motors, on the other hand have hundreds of turns on each core.  In order to get that many turns, the wire has to be very fine (it looks like 32Ga or 34Ga to me).  This wire is not strong!  Even when 3 or 4 ends are combined to attach to one of the lead wires, the arrangement is weak.  Pull on the lead wires with more than a pound or two, and you will have a lead wire in one hand, and a motor in the other. Don't ask me how I know this.

So,  use a substance like GOOP (available at Home Depot) and use a tie-wrap end to force a small amount into the motors right at the point where the wires exit the motor case. If the wires go through some heat-shrink as they leave the case, put a small tie-wrap tightly around the heat shrink to hold the wires in place.  After all, your GOOP made certain that the heat shrink tubing wouldn't be pulled out of the case, but it didn't prevent the wires from being pulled through the heat shrink.

And then there are crashes.  After a crash, I have had props, batteries, cameras, arms, etc.  winding up in places where they should not have been. You should design for crashes - that is, make certain that the expensive (or hard to get) stuff is NOT the part that breaks in a mishap.  In some cases, simply pulling on a wire to re-route it can break it or the circuit board the wire is attached to.

Use tie-wraps to tie the motor leads to the frame close to the motor - assuming you trust your motor mounts.  Use another tie-wrap to tie the wires exiting the ESC to the frame near the ESC.  If something bad happens, hopefully the bullet connectors will pull apart before too much damage is done.

When planning the wiring of an aircraft, always consider what will happen if it crashes.  Simply tying everything down is not always the right approach.  It depends a lot on just what you tie the wires to.

I have attached an EXCEL file that should help those wanting to know approximately what the effect of a larger battery will have on their "airtime".  This calculator can be used several ways:

You can put in the mAH and voltage of your current battery (use 11.5V for 3-cell and 15V for 4-cell) and change your "inefficiency" number up or down until you get the airtime  you have measured on your present setup.

I call this number the "inefficiency" because as the number goes up, the efficiency goes down. So, since it represents 1/efficiency, I call it IN-efficiency..

Once the inefficiency number for your craft has been determined, figure out what your craft will weigh with the new battery, change the maH number and enter those into the appropriate cells.  The airtime number should be very close to  the new airtime you can expect with the new battery.

Keep in mind that if your motors or props are heavily loaded, the inefficiency goes up rapidly, and the approximation will no longer be valid.

I was testing a new vibration damper for the camera gimbal of my little 6" prop "fun flyer".  This little beast has been so much fun to fly because it behaves pretty much as you would expect - always.  It is also extremely fast. 

It would climb out of sight in about 12 seconds. I figure that is about 500-600 feet, giving it a climb rate of over 2500' / minute.

It would also go "sideways" faster than anything I have ever seen fly.  The park that I normally go to is simply not big enough to gauge its true speed, since it will traverse the park in a few seconds.  With a different quad, I once had a greyhound (dog) race me across that same park.  They were 'even'.  But that quad wasn't nearly as fast as little one. The craft has a KK2 controller.  KK2's are easy to set up, easy to fly, and - since they don't have a lot of extra features like GPS, autonomous flight, or even a magnetometer, generally more reliable.

Controllers like APMs will go NUTS if they loose the signal from the compass, or if you accidentally mount the compass upside down, so it reads North and South correctly, but East and West backwards. I have done this.  The craft would fly perfectly if you flew directly North or directly South. But if you tried to fly more than a few degrees off the N-S axis, it would veer one way or the other..

Back to my story - I was testing the gimbal vibration damper on my street.  I did't want to take the time to drive to the nearby park for my test.  After all, I only wanted to see whether or not there was "Jello" in the video that came from the camera. I ran up the street and down the street, then I took it up fairly high - probably about 200'.  It was hovering at a steady altitude, then I started bringing it down. Not real fast, but not real slow, either.  All of the sudden, it flipped over and started falling.  I hit full throttle, thinking that it would have a better chance of catching itself if the props had some RPM.  I heard the motors wind up - then it flipped again!  This happened about 3 times 

And then it hit!

Hard pavement. Sad. Big mess. 4 Broken props. 3 broken motors. Broken frame. Sad. Broken gimbal .1 broken gimbal motor (of 2).Broken SJCAM HD camera. Bent (and therefore useless) battery. Cracked case on the radio receiver. Sad. It would have been in better shape if a car had run over it. Maybe 2 or 3 cars. I was really sad! The only other thing that could have created more damage is if the wiring had shorted and the whole quad burned up in a fireball fueled by lithium. 

So, I salvaged one motor, one gimbal motor and the main wiring harness - and the controller (amazingly, the only damage was 12 bent connector pins (which I fixed with a needle-nosed pliers). And I resoldered two tiny wires and used a bit of epoxy and fixed the SJCAM camera. I used more of that same epoxy to fix the radio receiver case.

The rest of the metal and plastic went in the trash.

I ordered a new frame ($8.99 from HK's US west warehouse) and a new gimbal and one gimbal motor.  Fortunately, I had 3 new (or slightly used) "main" motors that I can use to rebuild the craft.

I should have a complete replacement running by next weekend.

And I STILL don't know what happened.

I wanted to get my Santa Cruz project quad in the air so on Sat I took it to the park. It flew great and got to approx 200ft. Then it spun around on its axis and came crashing to the ground. Examination showed that one of the prop nuts came loose.  Moral of this story... USE LOCKTITE.