motor speed controller for kids' electric car

5 or 20

Yes, look up astable multivibrator 555, the datasheets have the calculations and there are probably many hobbyist websites that show the same calculations. The timing constant of .693 is actually the natural log of 2. For a trip down memory lane derive it from the equation for exponential decay in a RC circuit.

or 20

Hey Phil, what about the Schottky. Should OP get one with a bit more grunt?

"gearhead"

Hey Phil, what about the Schottky. Should OP get one with a bit more grunt?

** Nope.

The 16 amp rating is adequate cos, at most, the diode takes half the motor current.

..... Phil

se

..

15 or 20 d

Sure, Sorry just being lazy.

George H.

wn

en

ad

Your comment about adding series resistance got me thinking this would drop the power down a bit from the genuine 12V battery, and sure enough, it worked pretty well! Better than just one 6V battery, and less oomph than the 12V. My 5-yr-old daughter accelerated in no time. When my 7-yr-old (60 lbs) tried it, it stalled a bit trying to accelerate, but he just pushed off with his foot and then it took off. Will try upgrading the wiring to AWG 14 wire.

Thanks,

Michael

own

when

u

lead

You're better off using some rectifiers in series to drop the voltage. Series resistance has the effect of turning your voltage source into a current source, and the motor won't be able to respond to load changes.* That's why it wouldn't accelerate with the heavier child in it.

*Work out Ohm's law and what happens as a result of that series resistance when you put a load on the motor.

...

Looks like the batteries weren't sufficiently charged at the time. My

7-year-old is zooming around the house now, with no lag in acceleration.

A 20 amp car fuse looks like it's holding up well. A 15-amp fuse blew in a matter of seconds.

If I do go the PWM route later, is 100 hz to 1 kHz optimal? What are the trade-offs? Are there equations I can review relating PWM frequency to efficiency, etc.?

Thanks,

Michael

"Michael"

If I do go the PWM route later, is 100 hz to 1 kHz optimal?

** No.

What are the trade-offs?

** Lower frequencies cause extra heating in the motor that vanishes at higher ones.

The aim is to have near constant current in the motor at all settings.

Are there equations I can review relating PWM frequency to efficiency, etc.?

** No - it all depends on inductance of the motor and load current.

For small DC motors, use a frequency above 1 kHz for best efficiency.

For a 555 driving circuit, use a frequency below 10kHz for best switching efficiency.

...... Phil

I'm reviewing the circuit you posted - the original seemed to be from here:

Is the 47 ohm resistor from pin 3 to the mosfet gate necessary?

Thanks,

Michael

.

ng

Ok, I'm asking. I thought having a resistor at the mosfet gate would impede performance at high frequency, but maybe I was misled.

???

Michael

Yes, its necessary.

If you would like more reason as to why, just ask.

.

ng

I'm guessing it has to do with damping oscillations from the gate capacitance... but it seems that, from

V(t) =3D Vo exp( -t/RC ),

the smaller the R, the faster the gate capacitance will be discharged...?

Michael

The resistor at the gate is of a low value "47". It will not degrade performance out side of what the 555 can do..

At minimum, it will help reduce current peaks (inrush) as the output of the 555 cycles and help reduce over heating of the device.

The capacitance on the gate of the transistor will cause an inrush of current on the raise and fall of the 555's output. This can lead to heating of the 555, even though the 555 isn't really that fast on the output, hence the low value here which really is only helping slightly in this respect.

The main importance here is, you need to avoid parasitic ringing on the gate. With out some sort of device to lower the Q in the gate circuit between the driving signal and gate, you'll more than likely develop a ringing at the gate which in turn, could cause the attached device the Mosfet is driving to also get this effect. This will also put the Mosfet into a linear region where heating of the fet will most likely out sup past the spec's of the transistor and heat sink abilities. In the end, you have 2 things happening, a shorted MosFet and possibly the attached device damaged from the parasitic biasing the output out of control.

Hope that explained some of it.

The gate component of the FET becomes a little inductive. A high Q one at that.. You don't always get ringing, the circumstances have to be right for this to happen. Normally terminating the input with a R resolves this issue which places the gate out side this region.

t . .

cy.

hing

put

h of

n

ibly

How about reducing the resistor value to, say, 4.7 ohms? Or 0.47 ohms? Basically, of all the low value resistors to choose from, I'm wondering why a value of 47 ohms was chosen.

Thanks for the explanation,

Michael

at

gs.

nt.

ency.

tching

m

ld

e

utput

ush of

on

o

s.

ssibly

f

If you look at the datasheet for the BUZ11, it has a gate capacitance of 2000pF, equal to 2nF or 2 * 10^-9. Now you can calculate the timing constant defined by the gate resistor. To make the arithmetic easy, call it 50 ohms. Multiply R times C and you get 10^-7, or 100 microseconds. In that 100us, the gate charges to about two thirds the supply voltage (actually 1-(1/e), or 63.2% IIRC). By convention, a cap is considered fully charged after 5 or 6 tau (time constants), but let's get generous and call it ten. Then you are past the point you truly won't be able to see difference between the cap charge and the supply voltage. So less than a microsecond is all it takes to complete charging the mosfet gate using that 47 ohm resistor. Running at a few kHz, a microsecond is irrelevant. If you needed a faster gate drive (for example, if you had a circuit running at hundreds of kHz), you could indeed use a 4.7 ohm resistor instead of 47 ohms. But the idea is to use a big enough resistor to make sure the gate doesn't ring. You have to balance the two requirements -- big enough resistor to prevent ringing, but small enough to drive the mosfet at your frequency. Fortunately your frequency is low, so you are not squeezed between competing requirements. If you want to get into it deeper and investigate the design requirements for more demanding circuits than yours, you could start by looking up the Miller effect. But you don't have to worry about such stuff in the circuit you're contemplating building today.

I haven't looked at the data sheet on that part how ever, that value sounds common to me for this low frequency application you're taking on..

It's always good to have as much resistance as you can get, driving the gate with out compromising input drive..

If you were to calculate the maximum Slew and skew time (time coming on and time turning off) That you could get away with, taking into account for the Gate capacitance and max spec's of the device, you may be able to use a lower value how ever, 47 to me sounds like a good value with out actually doing the math..

You may be able to get away with values down to 10 ohms, but I would scope the gate looking for ringing before putting any time on it.

Also, make sure the R's are non inductive types.

es at

ings.

rent.

ciency.

witching

rom

ould

ade

output

nrush of

to

y

ng on

lso

t

ies.

possibly

of

Thanks a bunch! Now it is starting to make sense for me...

Michael

es at

ings.

rent.

ciency.

witching

rom

ould

ade

output

nrush of

to

y

ng on

lso

t

ies.

possibly

of

0.1 microseconds, you mean?

Ok. My spreadsheet shows after 5 time constants we're at 99.3% of supply voltage, and that's good enough for me, but hey, sure, 99.995% of supply voltage is good too, to be safe...

Ok. So if I put the 4 BUZ11s in parallel, that's 8 nF, and RC =3D 0.4 us. Ten RCs get me 4 us. At 10 kHz, that's 10,000 cycles per second, or one cycle in 0.0001 second. 4 us would be 0.000 000 4 sec, or 0.4% of the cycle time. Got it!

Thanks a bunch, again.

Michael