A lower cost, simpler timer/controller

Purchasing one of the first Brodak electric Super Clowns and testing various props since the recommended (at that time) 9x4 prop wasn’t available, I was frustrated by the 5-minute minimum flight time. I knew that ESCs are designed to be connected to the throttle channel on a radio receiver, where they expect to receive a 5-volt pulse of duration about 0.001 sec (1.0 msec) to 0.002 sec (2.0 msec), corresponding to minimum (off) and maximum throttle, recurring about every 1/50 sec (i.e., a constant 50 Hz signal).This is also the “default” or “standard” signal to all servos. For example, attached is a photo of a sample output of the throttle channel of a Futuba receiver.

Looking at the signals sent by the Brodak timer/controller to its ESC, however, it was obvious that they weren’t of this standard type. So I determined to use another ESC and my own timer/controller. I’ve been using Microchip’s PIC line of microcontrollers to control servos for a decade or more, so the task didn’t seem too daunting. The little 8-pin microcontroller that I’m using includes a built-in clock and a digital voltmeter.

Initially I used an LED to indicate that the timer/controller was active (as does the Brodak), but then realized that every ESC signals a valid “throttle” signal with beeps of various types, so an LED really wasn’t necessary. I’ve ended up with three little potentiometers (which are variable resistors connected across the 5-volt supply) that tell the microcontroller (1) how long to sit idling (at 900 RPM or so, signaling the user to steel himself for the approaching prizewinning flight, without using a significant amount of energy from the battery), (2) the flight time, and (3) the fraction of maximum power desired.

The microcontroller reads the idle time potentiometer and converts it from a 0 to 255 number to 0 to 63 and adds this to the programmed minimum idle time of 2 seconds to determine the idling time as from 2 to 65 seconds, with the latter giving plenty of time to get out to the center of the circle. Similarly, the flight time potentiometer gives a number from 0 to 255 that is added (in seconds) to the programmed minimum time of 1¾ minutes to give a flight time of 1¾ to 6 minutes. (Both the 2 sec and 1¾ minute minimums are easily changed when the program is “burned” into the microcontroller.)

The 0 to 255 number read from the power potentiometer is used to obtain the maximum pulse width during the flight time; for non-governed ESCs, that is currently 0.018 sec to 0.020 second (about 75% to 100% full throttle), and the program does a soft start by taking a couple of seconds to reach flight power, and the “throttle” is advanced during the flight as necessary to keep the power nearly constant.

I haven’t yet been successful in getting a Castle Creations Phoenix 45 into its RPM-governed mode, but the Atlas ESC (sold by Hobby Lobby for the lower-cost Atlas outrunners) is easily put into the governed mode (with its programming card) and does an excellent job of increasing the current as necessary to compensate for the falling voltage and maintaining a constant RPM; for this ESC I simply give the ESC a number from 0 to 32 that corresponds to a power of about 65% to 100%, and let the ESC handle the voltage drop.

The Electrifly line from Tower Hobbies includes a wide range of outrunners, ESCs, batteries, and chargers, all nicely compatible without having to solder any connectors. I have found, for example, that the Rimfire 35-36-1000kV, along with the Electrifly 35A ESC, the 9x4.5E APC prop, and the Electrifly 4S, 2100 mAh LiPo battery (a compact 8.0 oz), make a sweet combination that powers my old Flite Streak (30 ounces, 398 sq in) to a very stuntable 5.0 sec laps on 56 eyelet-to-eyelet lines, with maximum flight times a little over 5 minutes. The cost of motor, ESC, battery, and timer/controller is just about the same as for the Brodak motor, ESC, battery, and timer/controller, at about $220. The charger + balancer also cost about the same.

However, the Electrifly ESC is unique in requiring a couple of seconds of “off” throttle, a couple of seconds of “full” throttle, and another couple of seconds of “off” throttle, beeping after each, before it is willing to accept throttle values, so the timer/controller I make for the Electrifly will not work with other ESCs. The only pre-programmed choice for this ESC is to put the brake or off when the throttle is moved to “off” at the end of the flight. However, my program seems to do a good job of keeping the power nearly constant throughout the flight. This ESC doesn’t come with a switch for-5 volt power to the timer/controller, but there is no possibility of the motor starting up again after the flight ends as long as the battery leads are pulled off after the flight (recommended for all ESCs as well).

Another unique program in the timer/controller is needed for other ESCs that are not operating in the governed/helio/constant RPM mode. They require a constant “off” throttle (1.0 ms pulse at 50 Hz) until the button is pushed. I have used my timer/controller with the Phoenix line, the Jeti line (recommended by Axi for its motors), and the Atlas ESC, with no problems. The program advances the pulse width during the flight to keep the power nearly constant.

A third unique choice is to program the timer/controller for constant RPM mode. So far I have only confirmed this for the Atlas ESC. It should work in the Castle Creations Phoenix line if I can figure out how to get them into that mode (the PC program says it has done it but so far it hasn’t worked out in practice). Also, I intend to check the Jeti SPIN controllers, which also have the constant RPM mode, when I get a chance.

This timer/controller weighs 0.4 oz and uses a printed circuit board that is about 1” x 1.5” in size. The potentiometers are best adjusted with a jeweler’s Philips screwdriver (although one of them could be fitted with a knob that sticks above the potentiometer about 3/8”, like on the throttle emulator thread that follows, if numerous changes between flights are anticipated). I don’t see this timer/controller as a replacement for the sophisticated $30 timer/controllers now available, but it is a simpler and very usable instrument that is especially suited to sport and trainer applications, as well as full stunt patterns. If you’re interested, I’ll send one ($16 ppd) or two ($30 ppd), with a money-back guarantee if you’re not satisfied with it. You’ll need to specify the application (one of the three above). (Will Hubin, 719 Cuyahoga St., Kent, OH 44240, whubin@kent.edu)
 

Will,


Great work you are doing.  It is a nice thing to have another person jump into the development end of electric control line.   Since I have zero talent in this realm of things I Thank you.

Linheart


 

I already asked it in another thread http://stunthanger.com/smf/index.php?topic=6734.0

... why so complicated? why you want set so many parameters?

look my timer: http://www.rcmodely.sk/Images/Clanky/2173.jpg

I'm surprised that the Phoenix governs at such a low pulse width -- I'll try that.

I've found it very nice to be able to adjust both idle time, flight time, and power level while experimenting with four planes and many different motors and ESCs. It is easy to make a change if you want to and easy to leave it as it is if you don't!

Hi Will,
There are two RPM ranges as options in the Castle governor mode. The low speed tops out at about 9000 RPM with a motor like Alan's, while the High speed range would go to about 4 times that. Of course, if you run much more than 1500 uS into that same motor, the governor loop looses it and the RPM is determined by the voltage available, not the governor input. For a 6" pitch prop, you will run around 1300 uSec for human lap times. The resolution for lap time adjustment will require much better than a 1-turn pot. I find a repeatable 4 uSec to be necessary. What's repeatable? ... 10 degrees? 5 degrees? We're looking at a 3 or 10-turn pot. Shame is that they weigh a ton. Yes, I am sure you'll figure a way around the problem.

Sounds good!

Dean P.

Hi Igor,
We are certainly not talking about any R-C timing: that would be drifty and just plain poor design.
Many of these 1-chip processors have a simple A to D converter. Often they are 10-bit, or 1024 levels. (The exposition is for the non electrical geeks in the crowd.) If the processor has at least 4 channels (many have 4 or then you measure the voltage at the top of the pot(s) and at the wiper(s). The ratio of wiper to entire pot voltage gets scaled appropriately and used for timing. Sure, the absolute accuracy of the pulses and times is then subject to a maybe 100 parts per million per centigrade degree maximum drift, but this is nothing compared to the RPM setting changes that weather changes require.

That was the subject when I wrote to Will. I find that 4 micro-seconds per "needle valve click" is small enough, and if you are twisting a knob or needle then you would like each click to be maybe ten degrees. My by-hand repeatability is easily 10 degrees. Then the difference between the morning cool and the afternoon heat, on the same day will be a bit less than 1/4 turn, and the difference between Winter and Summer may be  a bit more than 1/2 turn. A 1 turn pot will have tiny adjustments where it is easy to loose your place.

Sorry if I beat it to death,
Dean

     I have been able to implement the ability to adjust the RPM/power level during the first minute of “flight” time by reading the power potentiometer during that time. For a programmed RPM range from 9,000 to 11,000 (Phoenix timer), I could make adjustments of ±8 microseconds in pulse width, corresponding to about ±100 RPM, although adjustments would be even finer with a narrower RPM range.
     Also, William DeMauro, who is evaluating two of these timers, made the excellent suggestion to use the pushbutton to make available a stop after the idle time and flight time begin. So now I program the timer so that it goes into an “off” throttle pulse width in an infinite loop if it detects the pushbutton being held down at the end of the idle time, or at any time during the flight time. The battery should then be disconnected before another run. Even though the motor can always be stopped by pulling out the 3-pin lead to the timer or disconnecting the battery lead, this seems like a worthwhile addition.