By Mark Davis
There is much discussion on RC hobby bulletin boards about BEC’s. It is popular to remove stock BECs and replace them. Is this necessary? Is it helpful? Can I hook up two and have redundancy? When do I need to disconnect the red wire from the ESC? Well, it depends on the situation. This article is to give a high-level model of the two main kinds of BECs, so you can understand the considerations that might apply to your situation.
What is a BEC?
One option is to run your receiver directly from an auxiliary battery, as in Figure 1. The other option is to steal a little power from the high voltage batteries that run your motor, and use that to run the receiver instead, as in Figure 2. Since the motor typically uses a high voltage, this requires a voltage regulator, to regulate the high voltage down to a lower voltage that doesn’t blow up your receiver and servos. Such a regulator is called a “Battery Eliminator Circuit” or BEC, because it “eliminates” the need for a receiver battery. The term “BEC” is used only in RC. To the rest of the world, it is simply called a “voltage regulator.”
Types of BECs
There are two main types of regulators: Switching and linear (sometimes called “LDO”).
The switching regulator can output more current than it takes at the input. Of course output power cannot be higher than input power. But power is I x V, and so the switcher can drop V and raise I without “creating power”. In practice, the converter is only about 85% efficient, meaning that IxV at the output is only about 85% of IxV at the input. The rest of the power is converted to heat, but that’s only about 15%, so not too bad.
The linear regulator or LDO passes current through, while dropping voltage. So I remains constant while V drops. The efficiency of this depends on how much the voltage is being dropped. If you start with a 6S battery at for example 24 volts, and regulate down to 6 volts for the receiver, that means V
At the output is only 25% of V at the input. Since I is constant, that means I x V at the output is also 25% of I x V at the input. That means 75% of the power is lost as heat, and only 25% of the power passes on to do useful work. So the efficiency of this type of converter is highly dependent on the percentage that you are dropping V. It is a good choice when Vout is not too much lower than Vin. It is a bad choice when Vout << Vin.
So why do you use one or the other? The advantages of an LDO are : (1) it is usually cheaper; (2) it usually has a lower minimum overhead (i.e., it can have Vin only slightly above Vout); (3) it has lower standby current, that is current wasted when there is no load; (4) it is quieter (the switching frequency of a switcher can bother your radio or FPV setup). The big disadvantage of an LDO is inefficiency. All the voltage drop is accomplished by converting input power into heat. This makes it almost completely inapplicable to high-overhead situation (such as dropping from 24 to 6 volts). Generally I use a switcher, unless there is a specific reason not to.
Maximum voltage and current.
As any real-world device, there are also limits to the voltage, and current they can handle. Read the input voltage limit, and never exceed it, or you risk smoking your BEC. BECs will usually also state a current number on them, however, there is a trick here. That is usually the maximum current the BEC can supply, and both switching and linear regulators have the property that their current capacity decreases when the headroom (Vin/Vout) increases. For example, if a BEC says “10 Amp”, that means that it can supply 10 Amps when (for example) the input is a 2S battery (charged to around 7.4V-8.4V) and the output is 6.0V. In this configuration, the overhead is very low. Both types of converters operate very well here. However, if you take that same converter, and connect it to a 6S battery (around 22-25V), then it cannot output 10A anymore. Its current capacity will be much lower.
The reason for this effect is completely different in a switcher vs. linear regulator (see the Appendix below if you care to understand more). But the end result is approximately the same: the number on the package is NOT always available, especially in a high-overhead situation. One difference is that the limitation for a linear converter is due to heating. Therefore, the maximum output current for a linear BEC is also dependent on cooling and airflow.
Vendor specifications are not always very clear on this point. For example the ever popular Castle 10A BEC says on the webpage:
So it clearly states a maximum voltage (25.2v), and gives two output points, 7A and 5A, depending on the input voltage. When does it supply 10A? Well, it doesn’t exactly say. We can assume it is in a very low overhead situation such as a 2S input. The point is, you can’t blindly assume you are always going to get 10A out of it just because it says “10 Amp” in big letters on the package.
Another common point of confusion is about tying the outputs of two BECs together. Does it work? Is it advisable? The short answer is “no”, but the long answer is more complicated. A lot of things can go wrong when you tie two outputs together. No two devices coming off the production line will be the same, and one might be trying to maintain 5.8v while the other is trying to maintain 5.9v. They will then “fight” each other. This can lead to heating, wasted current, and reduced device lifetime. With switchers, it is even more complicated – they can get confused and oscillate when two outputs are tied together.
Having said that, people do this often, and seldom get bitten. It is possible to design a converter to tolerate this situation. I have seen twins RTF planes that use two regulators with outputs tied together (e.g., the Flightline 1600mm twins), and in that case I would assume/hope that the engineer responsible for designing it carefully checked that the particular regulator works in that configuration. Some people also add additional output protection such as diodes, but if you are competent enough to do that, you probably don’t need this article, so I will skip that topic. Ask me at the field if you want to jaw about it sometime.
Some ESC’s have a BEC built in. They take in the main battery voltage, and output a regulated voltage on the red wire on the throttle cable. If you want to use a separate BEC with this kind of ESC, then it is advisable to cut the red wire on the throttle line to the ESC. Otherwise, you are effectively tying the output of two BECs together (the separate BEC, and the BEC inside the ESC).
Replacing your BEC
I often hear people say they “always replace the BEC” of an RTF plane for “cheap insurance.” Is that necessary? Like many things in RC, there is a dearth of statistically significant data. Everyone has a story or an opinion, but the plural of “anecdote” is not “data”. So all I can do is offer my own opinion. I see no evidence that, of all the errors a designer could make in designing an airplane, the choice of BEC has a higher error rate than anything else. Why not always replace the main wing spar? Surely that is also “cheap insurance”. What about the gear? They fail often.
The best you can do is monitor chat about a particular model, and react to what you read. If 10 people all broken the airframe in the same place, to me that is some evidence that a few carbon fiber rods should be added to that section. If 10 people report electrical problems, I might scrutinize the BEC. But if not, I see no reason that this component is more likely to be erroneously designed than another.
Another factor that might lead to replacing the BEC is if you know you are consuming more current than the stock design. If you add high throws, a gyro with high gain, maybe a few auxiliary functions that weren’t in the stock plane….all these add current. But in any case, you should carefully check the specs of what you are taking out and what you are putting in. If you don’t do this, your replacement might make the situation worse instead of better.
“The dumb thing”
One dumb thing you can do with a regulator merits special treatment, because it happens from time to time. If you have two batteries in series (for example, 8S configuration using two 4S batteries), then it is imperative that you connect the BEC to the lower voltage battery.
If you look at Figure 3 and Figure 4, you can see the smart way and the dumb way.
All of your components (ESC, receiver, BEC) will have a strong solid metal connection between all points that are ground. In fact, the PCB will often have one layer that is almost entirely ground, and every ground signal connects directly to this groundplane. Coloring all these “black” in the diagram to illustrate the groundplanes of each device, you can see that in Figure 4, you can trace all the way from point A to point B in the diagram touching only black. That means, you have an uninterrupted solid metal path between the + and – of one of Main Battery #1.
It might take you a moment to trace this path out in the diagram, but electrons are smarter and faster than you, because they will find this path INSTANTLY, and something along that path will be immediately and dramatically destroyed.
Receiver Battery vs. BEC
When should you use a BEC at all? The primary advantages of a BEC are : (1) weight savings; (2) convenience (not fiddling with multiple batteries for one plane). But a dedicated receiver battery is almost always much better in performance. Consider a tiny, lowly 2S 500mah 30C battery. That is about as humble a battery as anyone would use for a receiver. But 500mah x 30C is still 15Amps of capacity, which is a pretty high output. If you go up to 1000mah battery, that would be 30A. I’ve never seen a 30A BEC. So basically, a BEC is a lower-end solution than a receiver battery.
Another more subtle issue is that a BEC creates a ground-loop. You can trace a continuous circle of ground wire from the main battery to the ESC to the Rx to the BEC and back to the main battery. Magnetic fields passing through this loop induce a current. This is a long topic by itself, but is seldom a problem in practice. It may become a problem if you are putting together a system with multiple RF components (FPV, GPS, receiver).
Appendix : Why does current decrease with increasing input voltage?
For anyone curious, here is a very basic explanation of why current capacity of an LDO and switcher decrease with increasing Vin. The reason for each is different.
Figure 5 shows a loose conceptual diagram of an LDO. The main active element is a variable resistor. I have drawn it with the symbol for a potentiometer, but in reality it is a large transistor (called the “pass
FET” because it controls the passing of current to the load). A comparator inside the BEC adjusts this “resistor” such that V_out matches V_ref. The circuit is a simple voltage divider, such that
V_out = V_in x R_load / (R_load + R_pass).
If V_out is higher than V_ref, then the comparator will close off the pass FET a little, effectively increasing R_pass. If V_out is lower than V_ref, the comparator will open up the pass FET a little, effectively lowering R_pass. Equilibrium is reached when V_out = V_ref, and the comparator ceases adjusting the pass FET.
Of course in reality the load is not constant. Servos are turning on and off, etc. The comparator makes continuous adjustments to maintain V_out = V_ref.
The problem with overhead should be obvious from this diagram. The current I is constant throughout this chain – what comes in the top goes out the bottom. The power dissipated in R_pass is I^2 x R_pass. The power dissipated in the load is I^2 x R_load. So if you are dropping from 24v to 6v, then R_pass needs to be 3x bigger than R_load. That means 75% of the power will be dissipated (as heat) in the pass FET (R_pass), and only 25% of the power will go on to do useful work such as moving your servos. The current capacity is limited by the BEC’s ability to dissipate this heat, and the higher the overhead (V_in / V_out), the more power needs to be dissipated as heat.
A switcher operates on a different concept. The switch in Figure 6 rapidly opens and closes, at high frequency. When the switch is closed, V_in is directly supplying the load. When the switch is open, the V_in is disconnected, and the inductor and capacitor are supplying the load, with the diode completing the circuit.
The derivation is a bit complicated, but the simple result is that duty cycle of the switch determines V_in/V_out. If the switch is closed 50% of the time, then V_out is 50% of V_in. If the switch is closed 90% of the time, then V_out is 90% of V_in. And so on.
You can think of it as a “bucket brigade” that passes energy from the source (V_in) to the load (V_out) in little buckets of energy. These buckets of energy are dumped into the passive components (inductor and capacitor) when the switch is closed, and drained from them when the switch is open.
If for example V_out is 25% of V_in, then 25% of the time the source V_in will be connected. Let’s say the current being supplied to the load is I amps. In this example, V_in supplies I x V_in watts during the 25% of the time it is connected, and the regulator supplies I x V_out watts all of the time. These are equal. So the switcher efficiently transfers power from the input to the output.
In this case, the average current out drawn from V_in is only I/4 (because it supplies I amps but only 25% of the time). The current supplied to the load is a constant I (100% of the time). So the average output current is higher than the average input current by a factor of 4. Voltage has been reduced 4x, but current is increased 4x, and power remains the same.
So why does current capacity reduce with overhead? The reason is because the passive components are of a fixed size. If voltage is only dropping by 10%, then V_in is supplying the circuit for 90% of a switching cycle, and the passive components are only supplying it for 10% of a cycle. But if V_in is much higher than V_out, then the passive components are supplying current for the majority of a cycle. Basically, the “bucket” can only hold so much energy, and if you are drawing from the bucket 90% of the time, you can only draw from it slowly, otherwise the bucket will run out of energy before it is filled again.
Note that there are two ways to increase the flow of power out of your bucket – one way is to get a bigger bucket, but the other is to fill it more often. You can in double the current capacity (for the same size of passive components) simply by doubling the switching rate. The reason that your tiny Castle BEC can supply all your servos is not because the capacitors and inductors are huge, but because that switch is filling them many times per second. Typically switchers open/close the switch about 100K -2million times per second.
With a switcher, V_out will have a slight ripple on it at the switching frequency. This can affect your radio, FPV, or GPS. Also note that even if no current is going into the load, that switch is still spinning at the same rate. So a switcher will typically have a higher standby (no load) current.