A Mobility Scooter Motor Controller issue(s)!!

O.K. Guys/Gals...

I have a mobility scooter that is acting up pretty darn good these days.
NOTE: No longer under warranty and Medicare will not pay for it.

The problem was this:

A) Intermittent(...no, don't say that!)
B) When it does not work:
When pressing the forward/reverse accelerator pot the motor
tries to go forward, but may only creep or not go at all.

I have since took a lot of variables out of the equation:
1) The "forward/reverse accelerator pot"
2) The "max speed pot"
3) The circuit board with the pot connectors
4) The cable wiring from the pots to the motor controller
5) The wires from the motor controller to the batteries
6) The wires from the motor controller to the Motor

I've pretty much narrowed it down to the motor controller

The catch: A new replacement motor controller for this
scooter runs in the neighborhood of $500 - $600 US.

That is a big ouch!

I even when into the motor controller and replaced the 3 24V
relays and ended up with the same symptoms.

Thus I am here in hopes for some suggestions and or ideas on
my next step.

My inclination is to build a new controller.

Meanwhile I thought of just building a very basic controller that will
enable forward movement only.

Haven't really done any testing in the motor itself other than running it from
a 24V bench power supply.

...more to come



Notes on this circuit:
1) D4 is not needed as the MOSFET has an internal diode
2) I need to verify the appropriateness of R3 (2K pot)
Changed from 2K to 500K(V Gate increased from 7.94V to 10.54V at full on pot)
3) I need to verify the pot is configured in the circuit correctly
Pot was incorrectly configured, fixed this.
4) I need to record the Gate voltage and make sure it is bringing
the MOSFET into saturation.
Not sure what nomenclature designates this in the datasheet.

MOSFET Datasheet

Problems with it thusfar...

With 24V bench supply or 24V Battery, the wheels spin and the speed varries while on a block but do not when on ground.
Not sure why that is.

...After making changes (In Dark Red, above) The wheels spin faster, but when on the ground, the scooter does not
move. What would I check from a motor perspective?
Still not sure the MOSFET is operating correctly. Should not the 555 pulses to the gate always be at full on Voltage, like 10V? It should not vary from 0V - 10.5V!?

If the motor is a brushed type, have you inspected the brushes and the commutator?

beenthere said:

If the motor is a brushed type, have you inspected the brushes and the commutator?

Click to expand...

Been meaning to do that...

o I's say about 3/4 - 7/8 on both brushes, both smooth and not grooved.
o Commutator also looks clean, like polished brass, and smooth.

iONic said:

Not sure how to check the commutator, but I can check the brushes...

Click to expand...

look at it. if it's all black and/or gnarled and/or concave then that's a problem. you can clean it with fine grit sandpaper.

EDIT: the commutator is what the brush rides on in the center of the motor. there may be a 2nd plate so you can access it, or you may have to remove the brush holding assembly to access it.

I would omit D4 (the MOSFET has a built in one, anyway) and replace D5 with a high power fast Schottky diode, in the neighbourhood of 5-10 amps. Make sure to give the MOSFET good heatsinking. If you put both on the same heatsink, make sure to keep them isolated using a mica insulator or similar, or you will see sparks. A fuse in line with the battery for the motor would not be a bad idea.

With a 12V supply, do not worry too much about saturation: the MOSFET will usually be saturated above 10Vgs, anyway (the 555 will drop 1.5V on the output usually, so you're clear.) To be extra sure, you could use a 7815 instead of a 7812.

And, because discharging a lead acid battery too far can damage it, do you have some kind of under-voltage lock out?

@ Ionic - Quote "Notes on this circuit:
1) D4 is not needed as the MOSFET has an internal diode"

You will/do need the diode D4. Even though the mosfet shows that there is a diode in the mosfet, it is called 'an intrinsic diode'. It is just part of the internal structure of the mosfet, it does NOT take the place of the diode shown in the circuit.

@Ionic- your trying for a PWM circuit to control the motor,right? The 555 by it's self doesn't make a very good one to drive a mosfet/motor of much amperage. with a couple extra components you can make it much better

This circuit from Bill Marsden; http://forum.allaboutcircuits.com/blog.php?bt=678 Fig. 4.3 works and is a true PWM circuit. It gives an adjustable duty cycle square wave out put that the mosfet gate needs.

tom66 said:

I would omit D4 (the MOSFET has a built in one, anyway) and replace D5 with a high power fast Schottky diode, in the neighbourhood of 5-10 amps. Make sure to give the MOSFET good heatsinking. If you put both on the same heatsink, make sure to keep them isolated using a mica insulator or similar, or you will see sparks. A fuse in line with the battery for the motor would not be a bad idea.

Click to expand...

As per the next user post I will keep the Diode D4 in place. I do already have a fuse (not shown in schematic) in line with the battery (25A)

tom66 said:

With a 12V supply, do not worry too much about saturation: the MOSFET will usually be saturated above 10Vgs, anyway (the 555 will drop 1.5V on the output usually, so you're clear.) To be extra sure, you could use a 7815 instead of a 7812.

Click to expand...

Yes, I thought of the 7815.
The problem is that the voltage to the MOSFET seems to vary with the pot
position and do not know why this is.

tom66 said:

And, because discharging a lead acid battery too far can damage it, do you have some kind of under-voltage lock out?

Click to expand...

Since I live where I have to go uphill regardless of direction from home the scooter gets sluggish, letting me know that I had better get home.
I charge the batteries after every use and do not go more than 2 miles/use, generally. This will be considered in the new motor controller, with digital readout for individual battery voltage and current.

strantor said:

look at it. if it's all black and/or gnarled and/or concave then that's a problem. you can clean it with fine grit sandpaper.

EDIT: the commutator is what the brush rides on in the center of the motor. there may be a 2nd plate so you can access it, or you may have to remove the brush holding assembly to access it.

Click to expand...

They check out O.K.

shortbus said:

@Ionic- your trying for a PWM circuit to control the motor,right? The 555 by it's self doesn't make a very good one to drive a mosfet/motor of much amperage. with a couple extra components you can make it much better

This circuit from Bill Marsden; http://forum.allaboutcircuits.com/blog.php?bt=678 Fig. 4.3 works and is a true PWM circuit. It gives an adjustable duty cycle square wave out put that the mosfet gate needs.

Click to expand...

I will check this out ASAP.

shortbus said:

@Ionic- your trying for a PWM circuit to control the motor,right? The 555 by it's self doesn't make a very good one to drive a mosfet/motor of much amperage. with a couple extra components you can make it much better

This circuit from Bill Marsden; http://forum.allaboutcircuits.com/blog.php?bt=678 Fig. 4.3 works and is a true PWM circuit. It gives an adjustable duty cycle square wave out put that the mosfet gate needs.

Click to expand...

What would you suggest for the specs of the BJT's?



Is this circuit follow your suggestion??

24v on the gate of the FET will blow it up. I really don't like your gate drive circuit. I would use a regulated 12 for the push-pull pair, and still add a 13v zener diode from the FET gate to ground.

1N4005 for D3 is TOTALLY unsuitable, at medium speeds (50:50 duty) D3 will be conducting for as much time as the FET and will likely have a higher forward voltage drop and needs to be a large power diode, with better heatsinking than the FET.

At low walk type speeds D3 may be on for 2/3 of the duty cycle!

A 40 amp power schottky in TO220 or better still TOP3 pack would be a good choice. You can put D3 and FET on the same large heatsink (although isolated).

Where I said diode D5 in my last post, I meant D3 - the diode across the motor.

I would recommend you examine wiring CONNECTIONS very closely. When a previously working piece of equipment becomes intermittant, the cause is usually something that has worked itself loose. Screws, nuts and bolts, push on connections(either the tab or the wire crimp) and finally bad solder joints that have broken after being in use for a time.

If you are SURE that you have eliminated all those variables then go on with the modifications as others have suggested.

THE_RB said:

24v on the gate of the FET will blow it up. I really don't like your gate drive circuit. I would use a regulated 12 for the push-pull pair, and still add a 13v zener diode from the FET gate to ground.

1N4005 for D3 is TOTALLY unsuitable, at medium speeds (50:50 duty) D3 will be conducting for as much time as the FET and will likely have a higher forward voltage drop and needs to be a large power diode, with better heatsinking than the FET.

At low walk type speeds D3 may be on for 2/3 of the duty cycle!

A 40 amp power schottky in TO220 or better still TOP3 pack would be a good choice. You can put D3 and FET on the same large heatsink (although isolated).

Click to expand...

I have a couple of larger Diodes on hand:

1) BY459-1500
Vr - 1500V, If(peek) - 12A @48kHz, If(peek) - 10A @82kHz, Ifsm - 100A

2)HVR-1X3
Ifsm - 30A, 50ns recovery

Not sure if any of these will work.

Below: A revised schematic:



Not sure about the 13V zener though...
Any suggested Part for the "push-pull pair", the BJT's??

Here is a circuit that will work for your project; http://www.rmcybernetics.com/projects/DIY_Devices/homemade_pulse_controller.htm

I maybe wrong but, think that what you have been doing is changing the frequency to control the speed of your motor. You want to choose a frequency and leave it there. The frequency needs to be fast enough that the individual pluses aren't noticeable to the motor but not too fast. This is the P(pulse) in PWM.

Then after setting the frequency you adjust the voltage pot to the comparator part of the circuit. That is the W(width) in PWM. That controls the out put to the motor and the speed. While still allowing full torque in the motor.

A good read on what your trying to do (although the circuits are a little harder to follow) The circuits are ones that he was building to sell he explains how PWM works;
http://www.4qdtec.com/pwm-01.html
http://en.wikipedia.org/wiki/Pulse-width_modulation

My main objective is to get the scooter running or at least limping while I build a motor controller. Thus.... just getting a working PWM circuit to drive the MOSFET and motor correctly is of most importance. The sooner the better, my new dog is driving ME crazy and I need to get her out for walks to calm her down.

I have a couple of MIC4451 MOSFET Driver's I could use in the circuit also.

Note: the 555 PWM circuit I am using seems to vary the duty-cycle from aprox. 5% - 95%, but the voltage offset changes as well.

If it only drops to 5%, won't the motor keep going when it is in the stopped position?

tom66 said:

If it only drops to 5%, won't the motor keep going when it is in the stopped position?

Click to expand...

Possibly, I am not certain on the % Duty Cycle range. I think it is theoretically 1% - 99%.

iONic said:

I have a couple of larger Diodes on hand:

1) BY459-1500
Vr - 1500V, If(peek) - 12A @48kHz, If(peek) - 10A @82kHz, Ifsm - 100A

2)HVR-1X3
Ifsm - 30A, 50ns recovery

Not sure if any of these will work.
...

Click to expand...

How many amps will this thing draw? From my (limited) knowlege of electric wheelchairs and ride on scooters etc I thought it was common to get currents in the 20A range? Maybe 40A peaks for a second or so when you need to get up a kerb ramp?

I would be using a large FET and a large diode...

This is a circuit I have been working on for a while. It's based around a PWM oscillator by 4QDtec, but it integrates a variable frequency drive system - which can make some DC motors perform better. Basically, at low speeds as well as a low PWM duty cycle the motor is also run at a low frequency and at a high PWM duty cycle it is run at a higher frequency.

See attachment for image. Or here: http://forum.allaboutcircuits.com/attachment.php?attachmentid=30449&stc=1&d=1305116873

VFD is provided through the marked resistors R1 and R2. I made another, more complex version which had a JFET, and had better linearity. This one has fairly poor linearity issues at lower frequencies but levels out at higher frequencies. Both frequency and duty cycle are controlled using a single variable input ranging from 0V to 4V, and duty cycle is variable from 0% to 100% with a small dead band at either end.

Increasing R1 will reduce the frequency range. Decreasing it will increase it, but may affect stability. (It may stop oscillating.) R2 should be in the range of 0 to 22 ohms, but dropping it below 12 ohms causes negative spikes on the ramp waveform (about 500mV), which are an invalid input for the comparator and likely to cause the output to incorrectly change states; it also reduces the dead band. Increasing it can cause stability issues and it may stop oscillating.

Both op-amp symbols should be comparators, the lower comparator may need a pull-up if it is driving some kind of logic for the MOSFET. You can use an LM393 dual comparator. Oh, and the MOSFET driver isn't shown.

Here is the source for http://www.falstad.com/circuit/:

A 24V, 10A Power Supply does indeed move my scooter, but is not strong enough to do any hill at all. Thus I would be needing upwards of 15A to get me up the hills in my neighborhood and still not blow the 25A fuse.

Conclusion: although the PWM circuit I have works as designed for the 1.5A motor I have, it probably only offering less than 10A, maybe even less than 5A to my motor which is rated at 24V, 15A @3500RPM.

This assures me that the 25A MOSFET should work, but we are somehow not driving it into saturation. Not sure why as the scope trace shows 10V at the gate.

NOTE: 2K pot is now 500K.

My question is now, will the circuit below solve the problem??
And, can the 555 arrangement be left as is in the first circuit and the junction of R4/R5 be tied to pin 6 of the 393.