MicroSizer / BitChar-G Hacks

R/C Cars Can fly!

R/C on a budget: In this installment of RC-CAM Labs madness, we use the radio system from a little R/C car to control a cheap electric airplane. Yes, just about everyone likes a good hack. Taking something and making it do new tricks is always a challenge. It is educational and fun too.

This article describes the construction of a rudder-only controlled model airplane that weighs 2.5 ounces. Total cost was about $35. For sure, I learned some interesting things along the way, all of which are shared here. The good news is that the electronic modifications are working well.

Perhaps there's something in here that you will find useful for YOUR own Micro-Sizer / BitChar-G airplane powered hack. In addition to the information shown here, there are similar projects described in an interesting thread on eZone: Bitcharger Conversion Discussion.

Attack of the Clones

During the early part of 2002 a quiet invasion started taking place across the globe. Little itty-bitty R/C car races started appearing in very odd places. Schoolyards, offices, boardrooms, and restaurant dining tables quickly became race tracks for these 2.5" long remote control cars. Its roots can be traced to the clever folks at Tomy Toys (Japan). Having been a smashing success in Japan, these hypnotic little cars are now found in every corner of the world. My local flea market sells them for only $10. Street corner vendors and discount stores sell them for about twice that. Despite their low price, they work well and are a blast to play with.

The cars are sold under various names. Most are from mysterious China based vendors. All appear to be unauthorized low cost clones of the original designs produced by Tomy (BitChar-G).Many of the cars and radios look alike, so it appears that even the clones have been cloned. The most popular models are sold in a package that looks like a snow globe. Here is a partial list of some of the brands.

Brand Name	Store Package Style	Comments
Tomy BitChar-G	Box	The original, ~$40.
Charge-N-Go	Snow Globe
Flying Arrow	Snow Globe
GoldLight Super Racer	Box	Requires simple assembly, two coil steering, ~$12
Huale Mini Racer	Box	Requires simple assembly, ~$12.
MicroQ	Box	Requires simple assembly.
MicroSizers (Hobbico)	Box	Tomy BitChar-G for USA Market, two coil steering, ~$30
Mini Fit Racer	Snow Globe	2 coil steering, ~$12.
Motor Works Mini Racer	Snow Globe	Avail at Walmart, $14
Thunderbolt Scamper	Snow Globe
Shen Qi Wie	Snow Globe
Super Racer	Snow Globe
ZipZaps	Box	Appear to have been designed by Tomy. www.radioshack.com & www.zipzaps.com, ~$20.

Using the toy involves installing two AA size Alkaline batteries in the transmitter. The car has a hidden rechargeable battery in it. It is charged through a clever connector on the transmitter case. It only takes 45 seconds to obtain five minutes of racing fun. The AA batteries seem to last forever.

Under the Hood

The cars are offered on six frequencies (27Mhz, 35Mhz, 40Mhz, 45Mhz, 49Mhz, and 57Mhz), but what is available in your area depends on what country you live in. Here in the USA the only legal frequencies that can be used are 27Mhz and 49Mhz. I have seen some street corner vendors selling 35Mhz and 40Mhz models too. But these are illegal to use in North America, so you might want to stay away from them if that is where you are located.

Also, many of the transmitters used with the cheap clones (Tomy knock-off's) are missing the expected FCC conformance labeling. So they are technically illegal to operate in the USA. RF power is a few scant milliwatts, so the radio equipment appears pretty harmless to me. Besides, we will be hacking the transmitter, so any regulatory approvals will be lost anyway.

The transmitted signal is RF and uses a serial encoded OOK AM transmission scheme. The receiver could easily be disturbed by a signal that is on a nearby frequency. For example, CB radios can affect the 27Mhz models and older cordless phones can interfere with the 49Mhz versions. You will recognize interference by the loss of motor power and steering. Due to the RF encoding method, typical R/C glitches will rarely be seen (if ever). Signal loss will cause the car's motor to turn off. The transmitters are crystal controlled, but that is of little consequence. The receivers are clever single chip Super-Regen designs with high sensitivity, but they are not at all selective.

It isn't possible to use more than one of these models at the same time if they are on the same RF band (i.e., two 27Mhz transmitters cannot operate together). Simply put, to race someone else you need to make sure the other car is on another RF band. Range is under 25 feet, so another race can occur if it is several yards away.

The radio system used with these little cars utilize the "TX2/RX2" IC chip set developed by RealTek's Actions Division. The TX2 is the transmitter encoder and the RX2 is the receiver/decoder. There is another chip set made by Shenzhen State Micro that is nearly an exact copy of RealTek's chips. They are marked as "SM6136" (encoder) and "SM6135" (receiver/decoder).

The TX2 transmitter/encoder and RX2 receiver/decoder chips have gone through three revisions. The latest Rev-C (TX2C/RC2C) offers a PWM based two-speed motor control, called the "Turbo" mode. When utilized, the forward motor speed can be reduced by half. As of Dec 2002, none of the clone cars seem to have this feature. However, Tomy is selling turbo button equipped cars as BitChar-G "Booster Machines."

There is a small chance that your clone car's radio has the Rev-C chip set in it. If so, then hacking into this feature would be easy. Permanently ground pin 9 on the RX2 chip. Now when you switch TX2's Forward pin to ground the motor speed will be slow (50% duty @ 50Hz). Switching the "Turbo" pin to ground will give full motor voltage. This is a nifty hack.

To further save costs, the receiver 's RX2 IC is often found in "COB" (chip on board) format. This will look like a blob of black epoxy rather than your typical IC. Some of the cars are using the standard 16 pin DIP style IC package, just like Grandpa used to make. Both are electrically the same, but the blob style COB packaging saves weight (good info for you indoor micro airplane e-flyers). The DIP package is best for special hacking.

The COB receiver weighs about 2.1 grams. The DIP version is about 2.8 grams. These weights vary with the different brands, so don't go taking these numbers to the bank.

One of the ingenious features of the latest Rev-C RX2 receiver IC is that it can operate on less than one volt. In fact, it is powered by a single 1.2V NiCD or NiMh 120mAH cell hidden inside the cars. Trust me, this is not a typical operating voltage for a high gain receiver and wouldn't normally work well at all. The trick is that the RX2 chip has a DC-DC boost (step up) controller built into it. Pin 8, the LX output, is used to run the voltage boost circuitry. Operating at a few KHz, it uses an external transistor, inductor, diode, and cap, to create about 3VDC. The resulting current is low, so the steering actuators and motor continue to be powered directly by the car's 1.2V cell.

Please note that the RealTek Rev A - B and the Shenzhen SM6135 do not have the built-in DC-DC controller feature. These chips can reliably operate on 2.5VDC to 5.0VDC, so external DC-DC trickery is needed with single cell operation. This is not a big deal, since boosting the voltage is an easy task. The solution involves a low cost external DC-DC controller chip such as the Microne Semi ME301.

Note: The RX2 receiver can use from one to four cells (1.2V to 4.8V) and not worry about problems with the existing DC-DC circuitry. If your power source is between 2.5V and 5.0V then you can remove the DC-DC components (driver transistor/controller, inductor, input electrolytic cap) to save weight. If you use more than four cells then the transistor and inductor must be removed. In their place, install a 5V low drop out voltage regulator and connect it to the anode of the schottky diode. The LM2931AZ-5.0 VReg is a good choice.

There are two different steering actuator configurations that can be found in the various cars. They all use magnetic steering techniques and look nearly alike. However, as a soon-to-be-microsizer-hacker, you must be able to recognize the subtle differences.

The most popular method uses a rare earth magnet and two actuator coils. There are four coil wires involved. Each coil is driven by a dedicated NPN transistor on the receiver board. The coils have one common connection that is connected to B+ (1.2VDC). The coils' other end goes to the NPN transistor, which is configured as a saturated switch. These bipolar transistors can sink about 400mA. However, they have a voltage drop across them that can become obscene if you attempt to switch this much current. The typical actuator coils are 25-40 ohms and the load current is usually under 40mA.

When steering straight, both transistors are turned off. Any turn will require that one of the coils is energized to draw the steering magnet towards it. This is a simple matter of turning on one of the transistors, which "sinks" current to the attached coil. Receiver IC logic prevents both coils from turning on at the same time.

The other steering method utilizes a H-Bridge amp to control a single coil actuator. Only two wires are needed. I like this approach since it reduces the weight of the actuator, something that can be important on a micro electric airplane. This single coil trickery requires twice as many transistors and has a slightly higher voltage drop. The clone cars do not usually offer this scheme since it adds about 4 cents to the product's costs. Don't laugh -- that is a huge impact to the toy's manufacturing cost.

The car's power plant is a variant of the motors used in pagers and cell phones for silent vibrate alert. The motor control circuitry is a typical H-Bridge amp too. As with the single coil actuator method, the motor bridge uses four bipolar transistors. They do not use MOSFET transistors because they are too costly. Typical motor current is about 100mA, which is about as much as you should push through the transistors to mimimize the RPM robbing voltage drop. But frankly, the efficiency of the implemented bipolar H-Bridge is unacceptable for powering a model airplane electric motor, even a small one. We will address that issue a bit later.

The BitChar-G receiver wiring is straight forward. Details used to found on the ausmicro.com site, but the link we have to the receiver wiring has gone dead.

Home on the Range

Using the electronics from a MicroSizer type car, in a model airplane, presents a few challenges. I will show you what I did so that you can see a typical installation. But before I built the model airplane there were some nagging issues that I needed to address.

Radio Range: First, there is the matter of radio range. Most of the systems barely eek out about twenty feet. I experimented with the transmitter by adding an RF amp stage, longer whip antenna, counterpoised elements (dipole antenna), and other madness. They all worked well, but as it turns out, were totally unnecessary. I ended up with two simple changes that even Tim the Toolman would appreciate; Add more power.

My range solution began by increasing the receiver's aerial length. The length was merely changed to about twenty five inches. The exact length is not critical -- just try to get it as long as you can. I used 40 AWG magnet wire, so the added weight is not a concern (it weighs less than the stock 6" antenna).

Once the aerial is lengthened and installed on the model airplane, the receiver's adjustable coil should be tweaked. Use a non-inductive screwdriver (hardwood dowel trimmed to a chisel point). Adjust the coil that is near the antenna feedpoint for maximum range. Keep your hands off the board and your body away from the antenna as you do so.

This coil impacts stability of the super-regen circuitry, so be sure that your adjustment provides reliable operation from all distances and operating temperatures. I should point out that it is not a traditional antenna matching inductor (like the big boy R/C rx's have). Poor adjustments will set the receiver into intermittent oscillation -- not a good thing to happen while you are flying.

RF Power Mod: XX Perhaps the most important aspect of my range modifications involved increasing my 27Mhz transmitter's RF power. It did this by operating the transmitter's RF amp from a 9V alkaline battery. In stock form the transmitter is powered by two AA cells (3V). Using a longer receiver antenna and 9V in the transmitter, range is extended from 25 feet to over 200 feet. That is plenty for a small indoor or backyard electric airplane.

But it is not simply a matter of cramming a 9V battery into the transmitter. If you do this the transmitter will quickly be turned into a fancy paperweight. The upgrade involves adding a voltage regulator to the transmitter's TX2 IC so that it is powered by 5V or less.

I used a LM78L05 IC VReg, but a cheap zener diode is fine too. I'll show you both ways to handle this. All that is needed is to cut IC TX2 pin 9 away from the adjacent circuitry and add a voltage regulator. The drawing on the right shows the required cut that is made to the IC. The drawings below show the new components. The parts are not critical and you can sub them with anything that is close.
Zener Version	LM7805 Version

On the Zener Version the values are: Z1 = 3.9V Zener (1N4730A), C1 = 10uF electrolytic cap, R1 = 330 ohms (1/4 watt).

On the LM7805 Version the values are: C1 = 0.1uF ceramic cap and C2 = 1uF electrolytic cap. The VReg IC can be either a LM7805, LM340T-5.0, or the smaller LM78L05ACZ, your choice. A heatsink on the VReg IC is NOT needed.

If you use a Dremel tool you can grind out the AA cell battery chamber. Only a little bit of plastic removal is required. The 9V Alkaline will fit in the battery chamber like a glove. While you are at it, disconnect the wires to the car charging jack. It would be insane to charge the car's battery with the 9V powered transmitter!

If your transmitter does not have an On/Off switch then you must remove the battery when it is not in-use. The idle current on my LM78L05 regulated transmitter is about 20mA. It increases to 50mA while transmitting (i.e., anytime a button is pressed). Zener diode regulation (as shown above) will result in slightly higher currents. A fresh 9V Alkaline will last several hours if you are careful.

By the way, for even more range you could double the length of the transmitter's whip antenna. For best results you should reduce the value of the inductor that is in series with the RF amp's output by about half. I suggest that you not worry about this until you have tried the 9V power mod first.

Lastly, some transmitters are now shipping with reduced output power. These can be easily recognized -- the car's range will be less than ten feet. The fix for this is found on the tinyrc.com web site: cripple cap and ausmicro.com site: cripple cap (sorry, dead link).

Before the 9V modification my 27Mhz transmitter's RF power was a wimpy 5mW. After the modification the measured RF power was nearly 60mW!  Some of the "big boy toy" R/C transmitters are only 100mW, so now you do not need to hide your head in shame. But obviously, the transmitter is no longer FCC approved -- assuming it was in the first place. Licensed radio amateurs (hams) can use the mod with impunity.

Warning: Do not use the car charger on the transmitter after the 9V modification. The higher voltage is not compatible with charging your BitChar-G cars.

Control Freak

My micro sized model airplane has rudder and motor control. As shocking as it may sound to you, it does not have elevator control. As a kid I few rudder-only R/C models, so having motor control too seems like a luxury!

On a model airplane it is a bit silly to use a motor control system that includes reverse. And the poor efficiency of the bipolar H-Bridge amp will really take its toll on motor performance. On the other hand, the motor control's bridge amp is ideal for operating a single coil actuator. So, that is what I did -- I used the motor control outputs to operate the rudder and the steering outputs to control the motor. No, I have not gone mad!

Rudder Actuator Control: Most of the micro cars use two actuator coils, as discussed earlier. I could have removed one of the coils from the car and installed it, but I was planning on using a four cell (4.8V) pack. My car's coils were only 30 ohms, so they would draw a more current on the higher voltage. Don't get me wrong, they would work, but I wanted to reduce current consumption to support a tricked out pulse encoder that I will describe later.

So, I decided to make my own coil. Thanks to a fellow named Buddly6, an excellent discussion on making an actuator coil can be found on eZone.

My target was 60 ohms. I wanted to use 42 AWG magnet wire because it has a high enough resistance per foot to allow for a decently small actuator coil. But, I did not have anything that fine. I ended up dismantling a junk box 12V relay that had a very high impedance coil. The wire looked to be about 40 AWG.

I wrapped about 500 turns around a crude paper bobbin form that I made. The bobbin consisted of a pencil with a layer of wax paper. The bobbin's end caps set the coil width (about 0.18") and were made of stiff paper. I started with a layer of PVA glue. On top of that I started wrapping. Every fifty wraps I smeared another thin layer of PVA. The wrapping action was compliments of a cordless electric drill on slow speed -- the pencil just spun around and I guided the wire. A very painless effort.

After about 400 wraps I carefully scratched the enamel and measured the ohms. Using a little math I decided I was about 100 wraps short. I repeated the wrap and scratch process until I had a 60 - 70 ohm coil. I applied a final thin layer of PVA on the finished coil and set it aside to dry. The end caps, wax paper,and pencil came off without a fuss. I now had a coil actuator that was ready for mounting on my aircraft's vertical stab.

Motor Control: The motor is controlled by the right steering output. I could have just connected one side of the motor to B+ and the other lead to the existing actuator drive transistor. But, my intended motor & prop would have been be too high a load for it. This is simple to cure with a Power MOSFET (or HEXFET) transistor.

I used a PVN012 Optoisolated SSR IC because I had some samples floating around in my desk drawer. It can handle up to 4.5 Amps. It is available from Digi-Key but is a bit expensive (~$5). There are other choices, like the cheap GF4410, GF2304, SI4410, ZXMN2A01FTA, and IRLML2502 MOSFETs. All are under $1. But keep in mind they will require a different connection scheme. While choosing your power MOSFET be sure to observe its Current Rating (ID), On-State Resistance (RDS), and Gate Turn-on Voltage (VGS[th]).

The drawing above and on the photo on the right both show how my PVN012 is connected. Using a drop of CA, I glued the chip "dead bug" style (upside down) onto the board and cut the leads short.

The resistor is 150 ohms, 1/8 watt. Connection B+ goes to Battery Pos (+4.8VDC), B- goes to Battery Neg (ground), and M+ goes to either one of the receiver's original actuator drive outputs. I used the right actuator drive transistor. My motor now turns on when I press the "right" steering button and turns off when it is released.

Transmitter PIC-me-Up

At this point you are probably thinking "man, all his transmitter buttons are mixed up." That is true, but they are easily rearranged by a few cut-n-jumps. Or, a center-return toggle switch could be installed for steering (now we're talking!). But, as usual, I have my way of doing things.

In this case, "my way" involves a microprocessor. Using a PIC12C508, a tiny 8-pin microcontroller, I created a pulse proportional encoder for rudder control.

For those that have not seen this ancient R/C control method, it is quite a treat. The control surface constantly flaps back and forth. It is the the average position of the control surface that matters. It is perfect match for magnetic actuators.

The circuitry is amazingly simple and cheap. Just one IC (the PIC), two caps, and a 10K pot. The PIC has two outputs that connect to the Fwd and Bkwd switch inputs on the TX2 chip.

The rudder joystick is a simple swinging arm made from a popsicle stick, mounted to a 10K pot. There are two standoffs that are used to limit the left and right stick travel. It is all reminiscent of the sticks on the crude proportional radios I flew as a kid.

Pulse duty limits are about 60-40% and 40-60%. Full stick swings give full left or right control (the pulsing stops), so I could still fly the old fashion way If I wanted to. I was disappointed to find that the pulse frame rate could not be pushed beyond about 3 PPS (pulses per second) due to the lengthy (and slow) encoding scheme used by the RX2/TX2 chip set. I wanted about 6 PPS, but on a slow park flyer the 3 PPS rate will suffice. There were also noticeable pulsing errors during operation because the PIC is not synchronized to the TX2's native encoded data.

But none of that seemed to be a show stopper, since it was working on a minimal level. The final straw occurred after I boosted the transmitter power (see 9V mod). Now RF was getting into the pot wiring and causing odd steering problems. If I extended the antenna the PIC would shut down.

I was expecting some troubles like this, but I was getting anxious, so I moved on to building the model. I decided to pop the PIC chip out and fly the old fashion way (rudder banging). I originally planned to return to the encoder problem after the model was airworthy. At this point the basic electronic hacking madness was done and I went looking for a model to put it into.

e-SkyRunner

Harbor Freight sells two electric free flight models. They are often on sale for only $7. I chose the SkyRunner Free Fly (#43678) model since it had a 23" wingspan. I must admit that it is hard to beat an ARF electric model deal that is under ten bucks. In retrospect, I should have bought two -- one to hack and one to fly as-is.

My biggest concern was with the provided motor. Mine was very "gritty" and drew about 3.4A at 1.1V (simulated single NiCD cell). With the stock prop I measured 0.50 ounces of static thrust. I didn't have a warm and fuzzy feeling about using the motor, so I decided to splurge and use a Mabuchi FF-N20PN surplus motor. These cost about one dollar and can be purchased from e-flight hobby suppliers as well as from surplus electronic stores such as All Electronics and B.G. Micro.

The SkyRunner props are decent but I mounted a GWS 3020 direct drive prop on the N20. Static thrust was measured at 0.60 ounces at 4.4V, 1.2A. The Skyrunner prop was very close at 0.52 ounces. I was surprised to find that the total power for the crummy stock motor was less than the N20 (~3 watts vs. ~4 watts). Despite its lower electrical efficiency, I installed the Mabuchi motor to help reduce weight (the stock motor is heavy).

I initially experienced range problems whenever the N20 motor was running. I had soldered a 0.047uF ceramic cap (with short leads) directly across the motor's two power terminals, but obviously more EMI/RFI suppression was needed. Adding a 0.1uF ceramic cap from each motor terminal, to the motor's case, eliminated the interference.

The stock Skyrunner weighs 1.9 ounces. As far as successful flight goes, the name of the game is weight. EVERYTHING matters, even the type of glue you use. I figured that in order to fly well with the given meager thrust, the model needed to weigh no more than about 2.2 ounces. Note: Mine came in at that weight with a 4-cell 150mAH NiMH pack. A 120mAH NiCD pack put it at 2.5 ounces.

Thankfully the model had all sorts of heavy parts to toss out. For one, the battery charging assemble is very heavy, so it was the first thing to go. I also used a lightweight carbon fiber rod in place of the heavy wood dowel boom. The twin vertical stabs were reduced to a single stab/rudder design and the plastic reinforcements were eliminated. Basically, I tossed all the sandbags off the ol' hot air balloon. The only glue I used was PVA (Elmers) and CA (No-fume, styro safe).

The Harbor Freight SkyRunner kit is simple to hack. I think the pictures speak for themselves.

1) The Free Fly SkyRunner is available at www.harborfreight.com.	2) Here is what is found in the box. Included is a few spare parts and PVA glue.
3) An X-Acto knife is used to cut out the battery assembly and motor. These two hefty parts weigh 23 grams.	4) The heavy wood boom was replaced with a 0.060" Carbon Fiber rod. It is stronger and the lower weight helps balance the heavier tail.
5) The custom wound actuator coil is mounted on the vertical stab. Hinges are 5mil mylar. The big rudder is from a styrofoam egg carton (sanded down). Total tail weight is 3.3 grams (Vert/Horz stabs, coil, and rudder).	6) The Rare Earth magnet is mounted to the rudder using a piece of 22 AWG copper wire. I used a 1/8" diameter Radio Shack magnet (#64-1895), but the one installed in the car would work fine too.
7) The rudder coil magnet wire follows the boom to the Rx. I used a connector at the Rx end to allow easy replacement.	8) The 25" antenna is made with 40 AWG wire. It is taped to the underside of the wing. The CG location is shown here too.
9) The receiver has plenty of room inside the fuselage. It gets shoved further back and is allowed to "float" inside.	10) The 4-cell 150mAH NiCD pack fits like a glove. The battery cover is made from styrofoam egg carton material.

Flights of Fancy

This project started in October 2002 and had stalled early on due to the poor motor thrust. With the GWS prop I can get airborne, but it is underpowered. My longest flight to date is under thirty seconds. With a good hand toss the SkyRunner starts off fine, but slowly loses altitude and calmly glides to a landing. Turns are wide and easy going -- no issues there. Radio range is decent enough for use in a small park.

I do realize that if I used a Li-Poly cell I could shave a few grams of weight and perhaps make this a much better e-flyer. Or, pylon mount a geared motor and swing a bigger prop. But my goal of hacking a micro-sized car to use its miniature electronics in a cheap airplane was fully realized. For me, successful electronic hacking was the goal -- a good flying model would have just been extra icing on the cake.

Hopefully these notes will help you achieve your micro flight dreams too. Good luck to you!

Update: March 20, 2003

I made some changes to the model and now it flies like a champ. To reduce weight I replaced the heavy battery pack with a single LiPo (Lithium Polymer) cell. These are nominally 3.7VDC and weigh only 12 grams.

I also replaced the heavy single coil actuator with two smaller ones that have more power. Using 42 AWG magnet wire, I custom wound them to 45 ohms each and cross-phase wired them {in parallel} to the receiver's H-bridge "motor" driver. I could have used the stock low ohm coils from the car, wired in series, but it was easy to wind new ones.

The cross-phased wiring is a nifty trick. At left or right rudder activation, one coil pushes the magnet while the other pulls it. Working together, they ensure reliable rudder throw. I also used the rare earth magnet from the car's steering system since it was a bit more powerful than the tiny ones from Radio Shack.

Top view of the new rudder Actuator	Side view of the new rudder Actuator

With these minor changes I was able to achieve 1.9 ounces AUW, which is exactly what the model weighed right out of the box! The model is a slow gentle fellow in the air and turns are wide and graceful. It can quickly gain altitude if it meets a thermal, so now my main fear is that I will lose it to one.

I can proudly say that the conversion was a success. I hope yours is too.

Small Print / Feedback

All information is provided as-is. I do not offer any warranty on its suitability. If you have technical questions or comments about this project then please post it on the rc-cam project forum.