DC Motor Control Using MCU

Hi,

gauravkothari23 said:

but still the problem continues.

Click to expand...

What exactly is the problem?
And what exactly is the question?

***
To me the whole circuit makes no sense.
* Why all these recitfiers?
* Why no capacitor at all?
* why the paralleling of the MOSFET
* and the gate drive circuit makes no sense at all. I wonder if it can work at all.
(Why the diodes, why the 4k7, why the 47R?)
What´s the idea behind it?

Did you use a scope and do some measurements?

Klaus

KlausST said:

Hi,


What exactly is the problem?
And what exactly is the question?

***
To me the whole circuit makes no sense.
* Why all these recitfiers?
* Why no capacitor at all?
* why the paralleling of the MOSFET
* and the gate drive circuit makes no sense at all. I wonder if it can work at all.
(Why the diodes, why the 4k7, why the 47R?)
What´s the idea behind it?

Did you use a scope and do some measurements?

Klaus

Click to expand...

The problem is that the motor RPM drops when brake is applied and it takes around 60 to 90 seconds to reach its set speed.
suppose the motor RPM is set t 2000 RPM, after the brake is applied for 500mSec, the RPM drop to approx 1850 and then it takes around 60 to 90 seconds to reach back to 2000 RPM. it increases quite slowly say around 2 to 3 RPM per second.

The rectifiers are KBPC610 6A diode 3 in parallel to convert 220AC to DC to drive the motor.
Initially i was using single mosfet to drive the motor, but after a specific run, the mosfets used to get SHORT. so i added 5 Mosfet in parallel, where i am facing no issue of mosfet getting SHORT.
47R is gate resistor.
4K7 is Mosfet gate Pull down resistor. to creates a known level on the gate in the absence of an input signal which i have not installed.
1N4148 diode parallel to Gate Resistor 47R is to turn OFF the mosfet faster. which is again not installed.

Hi,

So when I drive my car .. hit the brake, then it slows down. And when I release the brake it takes a lot of seconds to get to full speed. Nothing special about it.
1) Why do you activate the brake?
2) Isn´t that all expectable?

***
I still don´t see the need for 3 rectifiers. Why 3?
I always used only one.

***
The diode never becomes activated.
A 47Ohms resistor to turn off faster? How does this work?

Can you do calculations about the expectable gate voltage?
Or use a voltmeter? Or a scope? Or a simulation tool?

In another thread you wrote you don´t have a scope. But I see you need one. Urgently. Try to lend one, buy an old one, build yourself one with the use of a microcontroller_ADC and some resistors, use your PC´s sound card....

Klaus

If your motor is driving an infinite mass, it will take an infinite amount of time for it to get up to speed...

Is there a control loop, or is this running open loop?

KlausST said:

Hi,

So when I drive my car .. hit the brake, then it slows down. And when I release the brake it takes a lot of seconds to get to full speed. Nothing special about it.
1) Why do you activate the brake?
2) Isn´t that all expectable?

***
I still don´t see the need for 3 rectifiers. Why 3?
I always used only one.

***
The diode never becomes activated.
A 47Ohms resistor to turn off faster? How does this work?

Can you do calculations about the expectable gate voltage?
Or use a voltmeter? Or a scope? Or a simulation tool?

In another thread you wrote you don´t have a scope. But I see you need one. Urgently. Try to lend one, buy an old one, build yourself one with the use of a microcontroller_ADC and some resistors, use your PC´s sound card....

Klaus

Click to expand...

when i drive my car, hit the brake (even when acceleration pedal is pressed), definitely the car will slow down, but as soon as i release the brake the car takes an instant pickup because acceleration pedal was not released anytime.

applying brake is the part of our machine which is applied automatically by some mechanical force.

the inrush current or startup current of motor is around 10 to 12 Amps, and single diode is of 6 Amps, so for that reason i am using 3 Bridge.

The diode 1N4148 is not used.
47R is just gate resistor.

barry said:

If your motor is driving an infinite mass, it will take an infinite amount of time for it to get up to speed...

Is there a control loop, or is this running open loop?

Click to expand...

its in a open loop

When I drive my car and I step on the brakes AND the accelerator, my brake pads overheat and my rotors warp and then i have to spend a lot of money for a brake job.

As mentioned by Klaus, you need some kind of test equipment. Maybe your MOSFETS are overheating. Maybe your brake is overheating. There’s no way to tell.

barry said:

When I drive my car and I step on the brakes AND the accelerator, my brake pads overheat and my rotors warp and then i have to spend a lot of money for a brake job.

As mentioned by Klaus, you need some kind of test equipment. Maybe your MOSFETS are overheating. Maybe your brake is overheating. There’s no way to tell.

Click to expand...

but the brake is applied only for 500mSec after every 8 to 10 seconds, so as i have checked, mosfets are not overheating, temperature is hardly upto 60 Degree. and if brakes are overheating, i am not concern because its not an electronically controlled part, its an external mechanical paart.

what kind of testing i have to do.

Hi,

paralleling power diodes doesn´t work very well because of thermal runaway.
So it´s far more reliable to use one 5A rectifier instead of 3 x 2A ones.

And a rectifier is rather robust, it can withstand some overcurrent for a limited time.
Additionally you may apply a soft start via PWM.

****
What benfit do you see in using the 47 Ohms.
I see no need for paralleling MOSFETs. One single MOSFET with a small heatsink should be sufficient.
I don´t think that overcurrent cause the MOSFET fail.
Two other reasons are more likely
* too high power dissipation because of wrong GATE drive
* voltage spikes

High power dissipation would be obvious because of the temperature.

Voltage spikes are more likley to cause a long term fail. And honestly I see no tiny piece of overvoltage protection in your circuit. Not even the tinyest capacitor to suppress spikes.

My opinion: A circuit without a capacitor is no (reliable) circuit. (with a twinkle in the eye)

Klaus

Hi

gauravkothari23 said:

what kind of testing i have to do.

Click to expand...

This is rather straight fowrard. A MOSFET usually becomes defective because fo these reasons:
* Exterme overcurrent (killing the internal bonding wires) for your 20N60 much more than 100A
* overtemperature (too high power dissipation, to small heatsink)
* overvoltage on D-S (even spikes >1us)
* overvoltage on (G-S), even spikes.

Now you exclude overtemperaure.
We may also exclude currents beyond 100A
Thus the remaning reason is: overvoltage.

--> So you need to measure V_DS as well as V_GS with a scope.

And when I say V_DS (or V_GS) .. please take this seriously: One connection as close as possible to the MOSFET_DRAIN and the other as close as possible at the MOSFET_SOURCE.

Klaus

The speed control obviously lacks authority, responsiveness or both. This is where you pick up hardware debug skills, if you want a career at this. Because strangers at the far end of a wire can only go by what you give them and have no sense of smell, touch, hearing, sight.

Open loop speed control will sag under load unless you add a feed forward current sense to the demand signal.

Do you know for sure that you aren't running too low a chop frequency, since maybe you are expecting the armature to be the inductor in the DC reconstruction? But have you a driver lineup that can go faster (and are edges sensibly fast enough now?)? Should you add an inductor that doesn't want to have electrical response to shaft speed & torque?

Do you know for sure that your setpoint is not bound up against some "ain't no more" limit on signal, supply, slew? What happens if, while stuck in the slow rise interval, you bump the speed knob? What does the outcome tell you about loop behavior?

And just for curiosity's sake, do you have any certainty that the schematic you show and the rest that you don't, have worked together before? Word is, not everything you find laying about on the 'net is trustworthy or in context.