Digital Servos, Gyros andLow Battery Voltages

by Rick last name withheld

Rick is an Electrical Engineer, R/C Helicopter pilot and factory representative for one of the more popular brands of heli.

The ideal battery or cell (for this discussion, the NiCd chemistry) will have a constant voltage for the entire discharge cycle. The rate of discharge will not affect the battery's ability to deliver the stated capacity, but the NiCd is not this ideal battery.  It is not always the stated capacity, has internal losses, and its ability to deliver current is affected by the load. Oh yeah, it is also affected by temperature.

Let's model this ideal cell as how it really works in our R/C environment.

 You have a voltage source and an internal resistance (due to materials and construction design) that are in series and are presented to the terminals of the cell. Cells are wired in series and terminated with a radio connector. As the cell is charged or discharged, there is a loss of voltage across this internal resistance. This is why a battery reads a higher voltage without any load, and drops with a load. The voltage that a cell produces is also affected by its state of charge and the load. A cell that is closer to full charge will have a greater output voltage at any given discharge current. The load is also factor in the cell's ability hold a steady output voltage. The smaller the load, the more efficient the chemical reactions are, the higher the output voltage will be, and the reverse is true. Higher loads will reduce the efficiency, thus the output voltage. Because of this dependency, battery manufacturers rate the capacity of a cell at a particular current. In the case of most NiCd makes, this current is C/5, where C is the milliamp hour rating. For example, the standard discharge rate for a 1200mah cell would be 240 milliamps. When you cycle a battery, you should use this current to get a rating under similar conditions that the manufacturer used to rate the cell (cutoff voltage being the same too).  Colder temperatures will reduce the capacity of a cell too. So, the amount of flying you can get out of a charge is based on temperature, state of charge, age of the cell, and how fast you discharge the cell (on two levels: just using the energy and the cells chemical efficiency at that rate).

 Back to internal resistance. Some cells are better suited to our R/C helicopter uses than others. A list of common cells that are used have these internal resistances (data directly from manufacturer):

• N-1300SCR 4 milliohms internal resistance (fast charge version)

• KR-1300SC 6 milliohms internal resistance

• KR-1400AE 10 milliohms internal resistance

• KR-1500AUL 16 milliohms internal resistance

• KR-1700AU 17 milliohms internal resistance

As can be seen,  the cell's design can alter the value of this resistance quite a bit. Note that we want the loss in the cell to be as small as possible. Cells that are father along in their life cycle will have greater values than these.

 General perception is that the higher capacity packs will work better in our application than a lower capacity one, but often the reverse is true due to the greater internal resistance. The cells with the lower internal resistances are quite good for our use but nobody wants the weight/size/lesser capacity penalty.

 To extend this to the 6 volt setup with a 4.8 volt regulator, the situation changes. We have the same cell and it's imperfect delivery of electrical energy, but add an active component. This changes everything. The regulator can compensate for the voltage drop of the batteries under differing loads and state of charge, giving us a much more stable voltage under flight conditions that a regular battery pack cannot.  Of course, there are down sides to this system: needs at least one more cell in the battery pack, the regulator and extra connectors can reduce system reliability, proper regulator design, and the rate of discharge is generally higher (see note below). This system can, however, speed up those servos that have been been starving for a solid receiver voltage.

 Note that an unregulated system with low output resistance cells or a regulator based system can (and will) run your batteries down quicker than you are used to. This is because, if the supply of voltage is more stable and or higher,  any load especially a servo, will draw more current. You will see better performance from the servo, but at a cost of greater current consumption, and since our batteries are rated in milliamp hours, this means less hours of flying.

Ron Cosby another EE, had this to say:

Ron, I just did a quick Google on battery discharge curves and found this link:

http://shdesigns.org/batts/battcyc.html 

I've seen other similar curves, this seems to be on track. I would expect there are curves out there for LiPoly for comparison.

The 4.7volt point is just below the %50 point of the curve and the author agrees this would be the "safe" point.
This supports the claim from the GEM people that you would still have %20 or more capacity to get safely back, assuming you did so when the GEM  went to steady on.

The problem with trying to calculate how long the pack will last is not knowing the current. For current peaks of 3amps, the voltage could dip below a safe level because of the V drop at the battery, plus the drop in the wiring. Even if there were only 0.5 OHM resistance in the wiring, that's a total Voltage drop of 4.7-3x.5= 3.2Volt, at the RX. Just like you said, you would need to run a test to know for sure, all helis would probably be different.

Based on what I experienced looking at the voltage readout on my GY502, I think a 3Amp spike could get you into trouble, if the wiring or batteries were not up to par. I saw mine dipping below 4.7V on the first flight and the GEM went intermittently steady at that point just as it should. I was able to get it into that state by slowly moving the controls "a little" while the heli was spooled up on the ground. It would NOT go steady when the heli was not spooled, like the rudder wasn't being loaded (I stay in heading hold).

Nicads have less internal resistance than NiMh, and those who use all digital servos would probably want to stick with NiCd, if for no other reason than the current spikes. I have no experience with Li or Duralites. Regardless of the batteries OR the wiring/switches you run, the GEM will tell you when you hit the 4.7V point, since its monitoring the voltage at the RX. I'm convinced its the safest bit of insurance you can get today.

 The GY401 is getting bashed pretty good on RR the last couple of weeks over this. I can't help but think that some of these people are just not paying attention and maybe some would IF they used a GEM or other monitor that would alert them to a problem.

Boil it down

Small packs with large capacity, such as the JR Extra packs aren't really ideal for use with all digital servos. These packs limit the current using a higher internal resistance. You do get longer flight times, but when the receiver calls for big current draws, they just can't keep up. Think of it as a bucket with a faucet on it. The SCR cells Rick talks about, with low resistance, have a big faucet. The current can really flow if it's needed. The AU and AUL cells have a small faucet. No matter how much you open it, only so much will flow.,,but it will take longer to empty the bucket.

4 cell NiMh packs are not really suitable for use with digital servos. Even if you are just running one on the tail, it can demand more current than the NiMh battery is designed to give. Before you start in on me, I know there are some of you out there who have been using NiMh and digital servos for 50 years without a problem. All I can say is that "Your very lucky".

Things to watch

• The age of your pack. Packs lose capacity as they get older.
• The length and gauge of all wires. Short and big is good.
• The switch needs to be high quality with big wires.
• No servos binding. This is critical.
• Temperature. Cold and hot temperatures effect charging and capacity as much as 30%.
• State of Charge. Check the voltage under a load. If I'm using all digital servos, I won't take off with less than 5 volts. That's a lot more than the requirement for standard servos. My limit there is 4.7 volts.

  ---

My experiment

I took a single Futaba 9253 servo and plugged it into a receiver. I powered the receiver with a variable power supply at various voltages from 4 - 6 volts. I measured the current being drawn as I moved the servo. With no binding and no load the current draw was about 350 ma. When I put the servo in a bind, it was drawing up to 1.7 amps while in a bind. This is an ongoing experiment. I plan to use a 4 cell NiCd battery next. I'll post the results here. I'll test other servos too. It should be revealing.

I took a standard JR4735 servo and put it in a bind. It drew a full 2.0 amps, which is more than the digital 9253 drew! I found the same results with 9202 and 9001 servos.

The difference was the servo arm on the 9253 didn't deflect hardly at all, while the standard servos had quite a bit of "give" in them. I felt like I was on the verge of stripping the gears on the 9253 when the current peaked.