Friends with benefits

Every member of my "fleet" is custom - built by me.  That means that there are a lot of custom parts.  Many times, I use parts I find on THINGIVERSE (3D printer files).  But most of the time, the parts I find there are not exactly right for my use. They don't have a hole in the right place, or they need  a piece cut off. 

Rather than go in an re-engineer the part and generate the proper STL file, I simply drill a hole, or file this, or cut that.  The end result is that nearly every part is one of a kind, and even if I can print a new one, it requires modification before I can use it.

Recently, I built a little quad, a little '250 class' fun flyer.  It didn't have any fpv or gimbal, but it flew nicely.  And it handled well - it was easy to fly. And it was very fast.  I was with a guy that I work with (and a personal friend) and I decided to go to a park and show him how the craft flew.

After 'throwing it around' in the air for awhile, I asked my friend if he wanted to try.  He took the controls and got it into the air.  About a 30 seconds later it was obviously out of control.  This was not a control failure but a controller failure (my friend being the controller).

Before I could grab the transmitter, the quad was into a tree and fell to the ground.  It broke 3 arms, one propeller, the main body, bent a motor shaft and fried one ESC (I suppose because the motor couldn't turn).

If I had a 'factory bought' unit, I could have ordered new parts and put it together quickly.  But since I didn't -. I had to print new parts, modify most of them a bit, cut down a propeller, and modify an ESC (with shorter than normal wires).  That took a long time.

There are two morals to this story:

1. Friends don't let friends fly their toys.

2. In many cases, a factory-bought unit is better - especially if you crash one.


For all you "racers" out there that are using EMAX MT1806, or SunnySky X1806 or something else similar on your 250-class quad or hex:

You have probably noticed that when using 5X3 props, your motors run fairly cool with 3S batteries, but when you move up to 6X4, the motors run too hot.

I have a solution for you -

FPV Model MC2204 motors (2300KV).  These little motors are larger diameter than the ones mentioned above, and aren't as tall, but they are significantly better than the ones they replace - and they have more power, my thrust stand confirms it. 

The FPV Model motors have a larger bolt-pattern - a full 16X19MM, and have M3 holes, rather than M2, but if you can fit these on your craft with little modification, do it!

These motors spin a 6X4 prop on 3S batteries just fine, and they barely get warm.  They are well-made and look to have good bearings.  The leads are better as well.  Instead of simply providing strands of fine copper wire like some motors, the FPV Model units have "real" wires. 

They don't have shafts that extend past the motor, instead they have CW and CCW adapters that bolt onto the top of the motor (with four M2 capscrews).  The motors cost a bit more - about $16 each, but they are worth it.

Battery Voltage and Capacity

When I talk to others about what they are flying, I sometimes get the remark that "A higher voltage battery (such as 4Cell vs 3Cell) results in longer flying time.

That is not always true.

It is probably true that if you take your quad that is now flying on a 2700mAH 3Cell battery, and replace it with a 2700mAH 4Cell battery, you will get a bit more airtime.  But the difference is not because the battery has a higher voltage.  It is because a 4Cell 2700mAH battery has 33% more stored energy than the the 3Cell 2700mAH battery.  But both are rated the same - 2700mAH.

In order to get true battery capacity you have to multiply the voltage by the mAH rating.  It takes 1000 milli-Amp Hours to give one Amp Hour (since there are 1000 milliamps in one amp).  A Watt-hour is a measurement of energy, so I like to use WATT-HOURS to measure the capacity of the batteries I use.  A 4 Cell battery produces about 16V, so the 2700mAH 4-cell battery holds (2700/1000) * 16 = 43.2 Watt-Hours, while a 2700mAH 3-Cell battery has a capacity of 32.4 Watt-Hours. 

A 2000 mAH 4 cell battery has almost exactly the same energy storage capacity as a 2700 mAH 3 cell battery.

It IS true that to produce a given amount of power, the current in Amps is lower for a battery with a higher voltage.  That is true. Lower currents result in lower voltage losses in wiring (assuming that the wire sizes are the same), so it is possible that there will be less system loss if a higher voltage battery is used, but again, that isn't always the case.  ESCs have losses as well, and depending on the design, those losses may be lower or higher at higher voltages.  Lower because of lower currents, higher because of higher switching losses.

I have generally found that - unless you are drawing more than 60A from your battery, you should try to use 3 or 4 cell batteries.  If your current demands are greater than that, or if you wire runs are very long, then you should use a 5 or 6 cell arrangement.

Of course, choosing the cell count depends heavily on the prop size, the KV rating of the motor and the motor maximum current, but if you are designing "from scratch" and have total control of the system, I would start with a 3 or 4 cell setup and work from there.

Also note that many ESCs are not designed to work with more than 3 or 4 cells.  You may be able to violate that spec for awhile, but if you do, you will have an unreliable piece of hardware in the air above your head.


I have had two ESCs catch fire.  One was the result of a failed ESC, the other was the result of a crash.

A little background -

Each ESC has at least 6 FETs (Field Effect Transistors).  And there are three windings in the motor it controls.  One end of each motor winding is connected to a common point, while the other end of each winding goes to a pair of FETs in the ESC.  3 windings =  6 FETs.  Some ESCs put these FETs in order to handle more current, and I have seen ESCs with as many as 18 FETs, but the principle is the same.

The FETs are used in pairs because one FET pulls one end of a motor winding to the (+) battery terminal, while the other pulls that same end toward the (-) battery terminal. By changing which polarity (+ or -) each of the 3 motor wires is connected to at any time, the motor rotates smoothly. The polarity changes several times each revolution.

If both FETs are ON at the same time, then the battery is shorted through the two FETs - a very bad condition. Fancy control systems have several fail-safe mechanisms to insure that both FETs are never on at the same time.

But the $15 ESCs do no qualify as a "Fancy control system".

If anything goes wrong, the FETs can fail in the "open" condition, or in the "shorted" condition. Failing "open" isn't a huge deal - the ESC simply stops working.  Failing in the "short" condition is another matter. If one FET of the pair shorts and the controller (that controls the operation of the FETs) isn't aware of that fact, the other member of the pair turns on. The result is burned wires, or worse - a fire.

I am advocating the EVERY multirotor have a fuse to limit the current during such instances.  I have contacted two fuse manufacturers and am trying to convince them to build special fuses - ones that are small, handle large currents and can be installed in-line easily.  The ones that are available now (mostly for trucks) are too large and heavy.

Flight time/battery life

The Santa Cruz project will take a lot of flying time, so I'm trying to maximize that.  I have looked at several posts about how to increase that time.

Some websites claim that to get maximum airtime, the weight of the batteries should be half the weight of the craft.  Others have similar answers. 

I disagree with most all of them!  I think I know how they arrived at their answer, but it is just plain wrong.

The efficiency of a propeller/motor combination is not linear at all.  At a low RPM and light load, the efficiency is good, but at heavier loads, that isn't the case. To get long airtime, you need all the efficiency you can get.  You have a certain weight, and you need to keep the number of Watts consumed at the lowest possible level.

For example:  Using a Turnigy 3536 motor (910KV) spinning a 12 X 5.5" APC prop and powered with a 4 Cell LiPO, I get 1KG of lift when the motor is consuming 173 Watts (5.78g/Watt) .  But when I increase the speed of the motor to max, the lift is 1.8Kg while consuming 450 Watts (4g/Watt). Note that I did not account for any losses in the ESC or wiring.

When I use the same motor and spin a 11 X 4.5 prop, to get 1Kg of lift, the motor consumes 184 Watts (5.43g/Watt), at full RPM, I get 1.5Kg of lift while consuming 284W (5.28g/Watt).  The decrease in efficiency isn't so dramatic with the smaller prop because the smaller prop can't get to the 1.8Kg level that the 12" prop managed.

Consider a battery that I just bought:  A Turnigy 4 cell 10Amp/hour (10,000mA/hour) battery.  It weighs 820g.  If we assume that the battery averages 15.5V during its discharge cycle, then the energy in the battery is 10 X 15.5 = 155 Watt/Hours. If my motor and my props and the rest of my aircraft weighed ZERO grams, and using the example above, one 12" prop/motor combination would consume 820/5.78g = 141 Watts.  My battery has a capacity of 155 Watt/Hours, so with a 141 Watt draw, the prop could hold it in the air for 155/141 = 1.09 hours = 66 minutes.  Not bad.   But wait!  We have not accounted for any weight except the battery. 

The motor weighs 105 grams and the propeller weighs 20 grams.  The speed controller weighs 25 grams and 2' of 14GA wire weighs 30 g.   That is 180g total.  Since we can lift 5.78g/W and we just added 180g, we now need 31 more Watts to lift those items.  141 Watts + 31 Watts = 172W, and we are very close to 1Kg total weight.  At the 172W level, we can stay in the air for 155/172 = 54 minutes.

So far, we have added no airframe, no flight controller, no camera, no landing gear, no ....

Also, we are producing only as much lift as we have weight.  We are hovering, not going up,or sideways.  If we have to do those things, we will be consuming much more power.  You might think that going sideways is not much harder than hovering.  Think again!  if you are flying at a 45 degree angle, half the motor thrust is (effectively) going sideways, while half is going down.  It now takes twice as much thrust to hold altitude as it does when hovering.  And you may wish to climb.  That will take more power than hovering as well.

So kudos to those who have obtained more than 1 hour of hovering flight.  Their airframe was toothpicks, and their batteries were special.

Adding batteries doesn't always help, since my measurements show that propeller/motor efficiency drops as they get loaded more heavily.  What good is twice the power if the efficiency is only half as much?

The secret is that larger propellers and motors, spinning slowly are not only efficient at light loads, their efficiency doesn't drop off until the load gets quite heavy.  So if you have a lot of batteries, you can lift them efficiently only if you have big motors and big propellers. You might think that such an arrangement would be good even for small craft, but you would be wrong since the larger motors and larger propellers weigh more.  The fact that the propellers are larger means that the airframe must also be larger and heavier.

So, if you like airtime, get the lightest frame you can find. Use big propellers (like 16") and big motors that have a rating of no more than about 510KV if you are using a 4 Cell battery or 380KV if you are using a 6 Cell battery. Keep weight to an absolute minimum, and it shouldn't be too hard to stay in the air for more than 30 minutes - as long as you aren't doing any stunt flying.

Santa Cruz Project - restart

I have been very busy, and have not spent as much time on the Santa Cruz Project as I should have.  Instead, I have been working at my "real' job designing data collection systems and power supplies.

I recently came upon the realization that I can't "get there from here" literally with my present design.  No matter how much I try, I can't get the current airframe light enough in order to be able to make it to Santa Cruz. The problem is the motors.  The more weight I have to lift, the bigger motors I need.  Bigger motors weigh more.

This is what I call the "Saturn V problem" - Putting a little more weight into orbit takes a bigger engine and more fuel, which weighs more, so you need a bigger engine and more fuel... To put 10X as much weight into orbit, you need something like 50X as much rocket weight.

I just need a little more.  My idea of using 16" propellers is not going to give me the efficiency I need.  By 'efficiency', I mean grams lift per Watt. I need a little more than 8 grams/watt if I am to make it. 

In order to guarantee that I get that kind of efficiency, I will have to use an 18" propeller.  But larger propellers spin more slowly.  That means that I'll travel slower horizontally, and I'll need that extra airtime just to get to Santa Cruz.

Challenges, challenges....


After much thought and calculations, I realized that my original design for the Santa Cruz Project was faulty.

I had built a very lightweight frame from 3D printed plastic and carbon fiber tubing.   It was light and strong.  The rather large motors had 16" (now 18") propellers sitting on top, and the 4 big battery packs hung vertically underneath the center platform.  The 4 batteries (together) weighed more than the entire rest of the craft. That is just what I wanted.

But I soon realized one thing:  In order to go fast horizontally, the craft has to tilt up to 45 degrees, and in order to tilt, the 'copter has to fight any weight that is below the center of gravity (CofG).  With so much battery weight below the CofG, I would waste a lot of energy just tilting the craft. What I  need is a center of gravity that is on the same plane as the center of lift (CofL) - the propellers! 

I need to re-do my frame to put the batteries much higher in the craft.  And with the CofG on the same plane as the CofL, when I drop a battery to the ground, the craft's CofG will not change much, making the control system more stable and easier to 'tune'.

So... back to the drawing board.  I'm still going to use my "BOX - X design", where 4 thin arms connect the central core platform to each of the motors, and 4 very thin rods connect the ends of the arms - in other words, a box with an 'X' connecting the apexes of the box.  Only this time, my battery 'box' will extend upwards such that a portion is above the plane of the propellers. 

Now, I need to find a good place to put the flight controller - which should also be near the CofG.  This is precisely the area used by the batteries.  I may have to make a square tube opening in the middle of the batteries just to hold the flight controller. 

More challenges!

Locktite, ABS and magnetometers

I was working on a GPS/magnetometer mounting bracket and wanted to insure that the screws holding the assembly to the (printed) frame of one of my quads didn't come loose.  Most of the GPS units you buy for mutlitirotors also include a magnetometer as part of the assembly.

Magnetometers (electronic compasses) are, of course, sensitive to magnetic fields.  Whenever a wire carries current, it develops a magnetic field around it.  For that reason, it is vitally important that you keep magnetometers as far away from current-carrying wires as possible.  Since multirotors have lots of current going to the motors, it is difficult to keep the magnetometers far away from those wires.If the magnetometers are too close to the wiring, the multirotor will turn (YAW) during periods of high-throttle. The problem can get severe enough that the mutirotor will go "nuts" in autonomous mode, since the GPS says the craft is going in one direction, and the compass says it is going in another. The GPS is not affected by the magnetic field.

In order to improve the situation, I often use a piece of magnetic steel or else "mu-metal" under the magnetometer.  Thin steel is an effective magnetic shield, mu metal is even better, but is more expensive and harder to find.  I also support the platform on standoffs or long screws to place the assembly well above the rest of the mutirotor frame.  In this particular case, I used long #6 screws for that support.  

I normally use nylok nuts on the screws, but I had no #6 nylok nuts, so I decided to use Lock-tite (blue).  



Quick update

I have been doing a lot of testing recently.  My focus has been on long-range autonomous flights and how the craft behaves in winds. As the props get larger on a quad, the stability in wind actually decreases.  There are two reasons for this:

The props have a larger swept area, and therefore have more area exposed to the wind, and 

The props are going slower.  Bigger props don't spin as fast as small ones.  Consider a car that is sitting on a highway and not moving. The wind is rapidly changing direction and is gusting from 10-20MPH.  The car 'feels' every change in the wind.  Now, move the car down the road at 60MPH.  The car is still affected by the wind, but since the car is moving and is creating its own 'wind', the actual wind is a small percentage of all the forces on the car.  So a little 6" quad is quite happy flying in a 25MPH wind, while a quad with 16" props has a much harder time.

Another thing to consider is maximum speed. A small quad 'screws' through the air faster than a big one.  This has to do with prop pitch and RPM (more about that later).  If a craft can fly at a maximum speed of 30MPH, it obviously cannot fly properly in a 35MPH wind!  Small quads often do have a maximum speed of over 40MPH, while a big quad with 16" props may have a maximum of 20-22MPH.  It would struggle to hold its position in a 15MPH wind.

Anything flying over the mountains (towards Santa Cruz) is likely to encounter some decent wind.  Any craft that makes the trip will have to take that into account.