Electric rideable kid's car soft start

[VelOBor]

Hi guys and gals! Need some advice on soft-starting a 12V DC motor, unfortunately don't have the exact specs.
Got my kid one of those power-wheels type cars, though this one is not the "brand-name", but a fairly decent copycat. The motors looked like regular brushed drill motors when I was putting the thing together, but didn't look closely.
The problem is this - when the thing starts moving, the motors kick into their full speed (they do have two settings selectable by a switch - slow and fast) and the car lurches forward (or backward if reversing). What would be the best way to give it a soft-start, say a ramp-up time of maybe 1 or 2 seconds until it reaches full speed. Preferably without using a microcontroller, although I have considered that option. Any advice is welcome! Thanks in advance.
Can provide the specs on everything if necessary. Photo of the toy attached, roll of e-tape for scale...

[Marco]

There are motor controllers on ebay with programmable soft start. The problem is how to wire it into the existing control system, which presumably already has it's own PWM full bridge.

[VelOBor]

I haven't cracked her open yet, but I presume it has no PWM. The car is either GO or NO GO
Has a remote control with buttons (forward-back, and left-right). Has a switch on the "door panel" for slow and fast. I'm guessing just switches between low and high resistance paths to limit motor speed. The remote control is so that the parent can drive the car with the kid in it, if the child hasn't quite figured out the controls. The gas pedal inside is also GO or NO GO. The steering is a little weird too, it doesn't seem to be a proper servo. A single press of the "right" button will cause the steering wheel to turn all the way right, a press "left" will return it to center and another press "left" will, you guessed it, turn all the way left. There is no way to turn a little. Either all the way to one side or dead-center. Rudimentary controls. I have considered buying a proper RC kit and upgrading the whole thing, and installing a proper throttle gas pedal with analog control, but I was thinking a quick and dirty "capacitor and resistor" type thing for the time being. Cheap, quick, easy.
And I'm not too keen on waiting for Ebay (sometimes deliveries to Russia take a month and a half).
Maybe I can get a similar ESC at my local Chip n Dip (Radioshack-type store in Russia)...

[Gyro]

You might just get away with a low value power resistor in series with the motor. Not enough to significantly affect its operating speed but just enough to reduce the motor voltage when starting from stall. The value would need some experimentation but possibly in the region of 50 - 200m ohms or so, depending on motor size and gearing. It would get very hot under continuous stall conditions though.

There is a lot of ex-Soviet military stuff on ebay from your part of the world, so you could hopefully pick up some decent sized wirewounds cheaply to experiment with.

[Fred27]

When I looked at one of these the slow setting connected the motors in series (so they see 6V each) and the fast setting connected them in parallel. Very cheap, but it works.

[Paul Price]

PWM controlling power to the motors in parallel with a suitable high current MOSFET on a large heatsink  would be best and a 555 timer and give you best battery operation with highest efficiency. You would take a string and spring to hook up the pot shaft to the gas pedal.
Depending on battery voltage. You might need a simple voltage boost circuit to drive the MOSFET gate so it is always turned fully on doing PWM so working most efficient doing PWM control.

Sounds like something fun to play with to put together a circuit to do this, but this may not so hard to do if you consider you can get help from people on this forum with any circuit ideas you can come up with. There is already a lot of circuits and info on the web about brushed motor control running on batteries.

A MCU would do a better job and provide the ability with cheap thermistors to measure motor temperature and some very simple tach circuit to detect motor stall conditions that could damage the motors by excessive current or high temperatures.

[bob225]

It sounds like it just uses a micro switch for the throttle a simple on off, what voltage is it ? 6 or 12V


A Pwm controller seems the way forward

Code: [Select]

[url]https://www.allaboutcircuits.com/projects/upgrade-your-power-wheels/[/url]

[VelOBor]

WOW! Awesome! I've been reading this forum for a while (and some serious binging on Dave's YouTube channel too), but this response rate is awesome! Thanks for the ideas, guys!
It has two 12V dc motors (generic kind, NSD B motor RS550S), each is wired through 2 relays (SARB-S-112D4U), I'm assuming for bi-directional control. The steering "servo" (although I haven't gotten that thing apart yet) consists of something similar, because the control board has two more smaller relays (SARM-S-112D4), which are wired in the direction of the front... I ASSUME (and we all know that when someone ASSUMEs they make an @$$ out of U and ME) that it is basically a DC motor with a couple of end-switches to tell the "brains" that the steering wheels are turned one way or other.
I've gotten into it superficially today, just to have a look-see and now I'm thinking a full-on tuning job on this thing. Like PIMP-MY-RIDE type thing. Proper remote control with ESC for the drive motors, servo for steering (I'm thinking one of those 1:5 scale types, with about 20kg/cm of torque should do fine), battery upgrade (higher capacity, not power), perhaps some protection/monitoring circuitry for the battery compartment and the motors. Now 2 questions remaining - do I build from scratch (Arduino-based, with something like an NRF24L01) and motor drivers and all that business OR should I cheat and just buy a RC kit (transmitter/receiver, ESC and blah-blah)...
I'm leaning towards the first option, considering that I can get help with all the circuitry.
This is going to be more or less a summer project for me, and I've been looking for something like this to do for a while now. Considering that I have a half dozen different Arduinos and bare-bones MCs laying around...
This is going to be fun.
Attached the photos of what I have discovered so far. Seems like this is pretty much the brains of the operation. I'll have to dedicate some time to actually figure out how and where all the switches come from and to, and all the LEDs on it and all the other stuff. There are two tactile buttons on the steering wheel, one honks, the other turns on and cycles through 4 tracks (one of which is Gangnam Style... go figure). There is a volume knob, an AUX IN 1/8" jack. A "gearshift" for forward and reverse (SPDT switch?), a lights on/off switch, manual/rc switch and slow/fast.
Oh, and a micro switch for gas pedal (motor on/off).
Doesn't seem too complex.
Wow, this is a lot of words. Sorry guys. Just excited about this project

[amyk]

The simplest "soft start" would probably be a suitably large inductor in series with the motors... although finding a "suitably large" one might be a bit difficult.

I'd recommend giving it an actual proportional accelerator pedal too, and agree with the others about a simple PWM using a 555. No MCUs, just do it all with good old fashioned analogue electronics

Is the steering "wheel" also a 3-position switch, or does it have an actual steering column and linkage to the wheels like a real car?

[Zero999]

I managed to find a basic data sheet for the motor. It seems like a fairly standard, inexpensive part.
https://www.robotshop.com/media/files/pdf/rbban64-banebots-DC-motor-specs.pdf

How about adding a resistor (a car headlight bulb may do but these also have a considerable inrush current) in series with the motor and bypassing it with a timer relay? The resistor doesn't need to have a massive power rating, because it won't be operated for long.

[ebastler]

I would forget about the remote control, and focus my energy on a PWM-based soft start. Remote-controlling your kid‘s powered ride just seems wrong to me. If he (or she) can drive the thing himself, he will hate to be controlled by you. If he can’t drive it yet, it is to early to give him the car.

[rstofer]

I had one of these RC cars back about 15 years ago and the controls were nearly identical to what is described here.  On-Off, hard-left hard-right, etc.

I never did solve the steering problem but I did change out the controls to an H-bridge on the drive motors.  I used an 8051 style uC but these days it would probably be an Arduino.

So, you need to measure the RC pulse width which should vary between 1000 ms and 2000 ms.  We consider 1500 ms as OFF, 1000 ms as full reverse and 2000 ms as full forward.

Now that we know which direction and how fast to spin the motor, all that remains is to generate the direction and PWM signals for the H-bridge.  This is the place where you add ramping.

There are a lot of similar projects on the Internet.

A brief overview:

http://blog.solutions-cubed.com/using-radio-control-transmitters-for-motor-control/

[VelOBor]

Lots of awesome ideas flowing in, guys!
The thing came "stock" with a RC unit (rstofer - we're definitely on the same page here), so that the kid can get used to sitting in the car first, then figuring out the controls later. My son is just over a year old (13 months right now), so steering the thing himself is a little early. I'm sort of leaning toward the idea of getting a NRF24 type unit pair, a cheap-o RC car control and hacking the whole deal to have proper control over the thing. Plus having an Arduino for the "breains" of the operation would open a bunch of doors to tuning the thing. Headlights, audio-system, proper accelerator pedal etc. ebastler - as of right now the car is either in full manual or full RC selectable by a switch inside the car, and my thoughts are to (perhaps) have parallel control just as a precaution, an additional safety level, so the boy doesn't run into someone's car or whatever.
PWM control for motor speed is the way to go, thanks guys, you have definitely convinced me of this. The only way to go if I want proper accelerator pedal. Question - PWM vs varistor (rheostat?) for this particular motor? Maybe PWM for RC mode and varistor for manual mode? Or if I use a micro (Arduino) - use a varistor for signal to uC and then the uC uses PWM for actual motor control. Now this begs the question - which motor controller will provide enough current without blowing up? I do have some heatsinks laying around and PLENTY of computer fans for cooling, but if the motor wants say 6A of current, it seems like no amount of cooling will prevent a L293 from releasing the magic smoke... Am I wrong on this one? What are the options? My plan is to next take the motor out and hook it up to my bench power supply (hope it can handle the current, the specs say 0-30V 0-3A) and see what kind of current it draws at start and operation. Put a little load on the output and see what it draws. Looked up datasheets on this motor and stall current is 85A. That seems a little nuts for this motor, but I guess makes sense. Although the car has a 15A fuse just after the batteries, so maybe the motors aren't that high. Hope that my power supply is protected from overcurrent. Would hate blowing it up...
Also I want to test the slow/fast speeds of the stock brains and how it works. Whether it's PWM or connecting motors in parallel/series for 12/6 volts. Idea is this - hook up a voltmeter across the motor terminals and put it in drive. Next step - hook up my scope to motors and drive them (with the wheels lifted up off the ground) in fast and slow mode. See if I get a clear PWM type signal or just straight up 6V or 12V DC.
What said earlier about the relays on the board I take back - the forward and back are controlled by a "gearshift" next to the steering wheel.
I will post updates when I get any further on the reverse engineering of this. Again - I'm leaning for a full-on rebuild, with microcontroller and full RC, rebuilding motor drive circuitry from scratch with PWM control, possibly headlights and tail lights, maybe a built-in MP3 player and all that.
Step-by-step though. First I plan on getting an NRF24 kit and two arduinos and see if I can get them to communicate. I've found some resources on that already (schematics, programming and all that). Should be fairly easy. Then see where I can go to next.

[rstofer]

The maximum inrush current will occur at 0 RPM so you can just measure the DC resistance of the motor and plug it into Ohm's Law.

There are a number of H-Bridges that could handle the current, some are for single motors, others handle two:

https://www.robotshop.com/en/medium-power-motor-controllers.html

ETA:  If you can't measure the low resistance, you can measure the current when applying full voltage with the armature locked.  Be quick, the motor is getting hot!

[VelOBor]

Awesome, thanks rstofer. I'll try that. Seems easier to just unplug and pull the motor rather than try to measure while it's running. Thanks for the link, I'll see what sort of chip I can come up with. I'm not a fan of ordering online due to the wait time, but Moscow is pretty decent at having pretty much anything you'd want. Except some really, like REALLY, specific stuff... But if you're willing to break into your piggy bank, you can get it too. But most electronic components are readily accessible. Found a TB6612FNG bare-bones chip at my favourite shop's site, says it can do 3A peak. First I have to see what the motors REALLY draw though, before going further. Maybe I'll have to shove my whole "don't want to wait" thing and buy online.
Thanks for the link!
EDIT:
And why do I need a driver anyways? Why can't I just build an H-bridge with suitable MOSFETs or relays?

[Gyro]

.... Found a TB6612FNG bare-bones chip at my favourite shop's site, says it can do 3A peak. First I have to see what the motors REALLY draw though, before going further....

I think you'll find that the motor current will be much more than that. 12V at 3A is only 36W, I doubt that will move anything. I would expect more like 15A , 180W, 1/4HP, with higher surge current at startup (the fuse will take more than 15A for short surges). As suggested in your edit, I think you are looking at discrete power MOSFETs if you want to do PWM.

Edit: Corrected figures

[VelOBor]

Thanks Gyro.
I'll get some measurements done and post the actual data here.
Need. Solid. Data. 
I'm going to try to get into the thing tomorrow if it's not in use by the youngun. He can't drive it himself, just sits in it and pushes the horn and music change button. The wifey is doing all the driving.

[CatalinaWOW]

These things usually have a large ratio step down gear chain following the motor.  While the electronics solutions will be more elegant, a purely mechanical solution may take enough of the "jerk" out of startup to do the job for you.  Add a flywheel (rotational inertia) on the motor end of the gear chain.  This is roughly the equivalent of adding a series inductance as mentioned above.   Depending on the exact configuration and your resources this could range from very simple to nearly impossible, but if it works it will be extremely reliable.

[rstofer]

But the flywheel works both ways.  It may take longer for the motor to accelerate but it will also take longer to decelerate.  Whether that is a problem is going to take a little experimenting.  The relay based controls that come with these toys doesn't have dynamic braking (IIRC).  The relay simply opens the circuit and the motor eventually costs to a stop.

These toys probably don't have brakes - at least I don't think mine did.  OTOH, mine wasn't a riding car.

[james_s]

I just went through this exercise with a cordless electric lawnmower. The original mechanical switch had failed so I replaced it a while back with a mosfet. Recently I replaced the SLA batteries with LiPo packs and the much lower ESR caused huge inrush that blew the mosfet. I built a simple soft start circuit using a ATTINY25 8 pin microcontroller diving a beefy mosfet. You can buy speed controllers ready to go if you wanted though, they make them for electric scooters and RC cars and such.

[VelOBor]

I wanna steer away from mechanical solutions. For one - I don't have any machining gear to make a proper flywheel and looking for one that will fit the motor axle is too much of a bother. I'm fairly certain that the motor has plenty of torque, so in order for this flywheel solution to be worth it - it would have to be a heavy flywheel, but the gearbox is fairly tight. I think in my case (considering all the criteria that I have put in place) an electronic solution will be best. Unfortunately I was unable to get into the guts of her today and measure the motor resistance, so I'm pretty much at a stand-still.
Will a MOSFET be able to handle PWM? So that I don't have to worry about a beefy enough motor driver/ESC. I don't need regenerative braking or anything of the sort. I'm thinking MOSFET-based H-bridge, an Arduino UNO should have plenty of pins to drive everything (again - considering all the options that I want, the headlights, temperature monitoring, RC, etc.) and still have a few to spare.

Okay, am I getting this right - if we number the MOSFETs in the H-bridge 1-4, like this: ¹₃H₄², then I just need 1 pin for each direction? Just wire 1 and 3 to one pin, 2 and 4 to the other. Then I only have 4 pins occupied for the two motors. Furthermore - if I wire the two drive motors in parallel onto the same bridge, I cut the pin number in half. So I only take up two pins for forward and reverse.
james_s - I've considered LiPo packs, but I don't trust myself enough to put something like that under my kid's butt. The lead/acid batteries provide enough current, and the specs say 2 hours of continuous run time. I'm thinking an hour is more realistic. I will be testing that also. One of the mods I have in mind is adding two more of the same batteries in parallel to increase run time, but that is optional for now. Maybe after a couple of years I will get more powerful motors and then upgrade the power supply also, but I do appreciate the idea though!

[james_s]

Well my mention of LiPos is only because that's what I used, the SLA batteries were shot and I already had a pile of LiPos for my RC planes, may as well put them to use. The speed control will be the same regardless of the power source.

It's a tad more complex than that for the H-bridge since the high side mosfets need to be driven differently than the low side. There are lots of H-bridge motor drive circuits out there though so that part isn't too hard to work out. The thing you have to watch out for with PWM control of largish motors is the spikes that you get each time the transistor turns off can easily kill it. You often need a snubber across the mosfet. This should give you a few ideas http://www.masinaelectrica.com/dc-motor-speed-controller-pwm-0-100-400hz-3khz-freq/

Also look for projects involving large robotics, many of them use brushed motors with H-bridges. Some of the larger RC car speed controllers might be suitable too.

[rstofer]

Just buy a commercial ESC, you aren't going to build a better one.  One thing you need to watch out for is 'punch through' where the top and bottom MOSFETs (on the same side of the H) are on at the same time.

The bottom MOSFETs will be N-channel and you might be able to drive them with logic levels but usually not (and get them into saturation).  The top MOSFETs could be N-channel, in which case you will have a complex issue driving them (you need to develop a gate voltage some 10V above the top rail) or P-channel which is easy to drive but more difficult to select.  That's why they invented H-Bridge driver chips.

You will drive the top MOSFETs with a solid direction signal and the bottom MOSFETs with PWM.  Some commercial controllers have logic on board to cut the number of signals down to 2 instead of 4 (per motor).

If you can get by with a 10A controller rated 30A peak (30 seconds), it will only cost $24.

The PWM frequency for MOSFETs is usually pretty low (like 1 kHz or so) and the reason is that the  current waveform is going to look like a trapaziod.  The problem is that at the leading and trailing edge, the MOSFET is working in a linear region and developing heat.  The width of this edge, for a given controller, is fixed and the fewer the better.  At a high frequency a much higher percentage of time is spent in the linear region.

MOSFETs can definitely control the motor but without knowing the inrush current everything is pretty much a guess.

Your idea of wiring MOSFETs 1 & 3 gates together is exactly how punch-through occurs.  During the transitions, both devices are on (slightly) shorting out the power supply.  When this occurs, the MOSFETS self-disassemble.

[VelOBor]

Thanks for the input guys!
Did a little measuring today (haven't tested for PWM with the scope yet, but will do shortly, although I STRONGLY doubt that it's speed controlled, just straight 6V or 12V depending on speed setting, you never know with the chinese though... also will test for ACTUAL current with multimeter) with a multimeter and got this:
the motor resistance is in the 0.6-0.7 ohms vicinity, which gives us about 10A, 60W @6V (slow speed setting) and 20A, 240W @12V (fast setting).

So I am going to need a commercially available ESC that is capable of handling that kind of power. If I want to be able to get the speed out of the car that it is designed for, I will have to get an ESC rated for 20A current, with higher peak performance, right? If I get one rated for 10A and 30A peak, and I'm running it at 20A (fast setting, throttle all the way down) I'm running the risk of burning it out. Or can I just provide it with ample cooling (heat sink and fan) and not worry about it?

[james_s]

Is that under load? With a kid in it going up a steep hill the current is likely going to be substantially higher.

My friend's daughter has a similar kiddie car that he retrofitted with a 4 cell LiPo, it will do a burnout now   Ok so the solid plastic tires provide very little traction so that helps too.