| Author |
 Topic  |
|
boron
Mad Scientist
Canada
223 Posts |
 Posted - Apr 19 2009 : 7:07:40 PM
|
Hello all,
I have been trying to wire a 350 watt 24 volt dc electric motor with PWM in order to control the speed of a scooter. The PWM works very nicely on a smaller scale when the signal is fed into a MOSFET. When I then scale it up to the full test, with the scooter suspended meaning no load, the acceleration PWM and motor all work beautifully.
However, when you try to ride it we hit some trouble. The rider gets on and manually gets the scooter moving, then engages the motor. At minimum speed the motor is not turning fast enough to move the rider. So he accelerates with the PWM. But then at a certain point in the acceleration, as the duty cycle increases and the motor starts to get to the point where it starts pulling the rider(not very far into it) the mosfet blows, shorting and causing the motor to attain max speed until we turn it off.
My question is why does this keep occuring? I have tried this with two different mosfets: IRFZ46 and IRF540N both blow in the same way at the same time.
I know that as the load increases so does the current flowing into the motor and mosfet. Does ohms law give the maximum possible current if I use v=24 and r= motor winding resistance? If not how can I approximate the maximum current that can occur?
Also, I have been using about 5 volts to activate the gate of the mosfet. Is that enough and if not would that be causing this problem?
Any tips or help would be very greatly appreciated. I think its just something I'm accidentally overlooking.
Thanks in advance. |
|
|
audioguru
Nobel Prize Winner
Canada
4218 Posts |
Posted - Apr 19 2009 : 11:26:51 PM
|
The datasheets for the Mosfets you blew up say that the gate voltage must be 10V for them to turn on completely. Your 5V is probably cooking them to death. Inside they can melt before their outside gets warm.
You forgot to post your schematic for us to see if the high capacitance of the gate of the Mosfet is given enough current for the Mosfet to turn on quickly so it does not cook.
You also forgot to say what is the resistance and current ratings of the motor. |
 |
|
|
boron
Mad Scientist
Canada
223 Posts |
Posted - Apr 20 2009 : 07:58:11 AM
|
Ah okay, so above ten volts would be best then.
Oops, I'm sorry I forgot to post that.
The motor label reads as follows: 350 W, 24 volt, 19.2 amp
And when the resistance of the motor winding was measured with a multimeter (subtracting the resistance of the multimeter leads) it is about 0.9 ohms.
The PWM circuit is as follows: http://www.cpemma.co.uk/pwm.html (the first one on the page, using the lm324)I realize now that the diagram calls for 12 volts... and I was using 5-6 ish... Should the gate of the mosfet also have a resistor connecting it to ground?
Thanks again, appreciate it. |
Edited by - boron on Apr 20 2009 5:29:40 PM |
|
|
|
audioguru
Nobel Prize Winner
Canada
4218 Posts |
Posted - Apr 20 2009 : 08:38:04 AM
|
The circuit you built is made to drive a little fan. With a 12V supply, the opamp provides about 10.5V to the gate of the Mosfet to turn it on fully.
But the lousy old LM324 is too slow to switch a powerful Mosfet.
The circuit is missing an important diode in parallel with the motor.
Your motor draws 24V/0.9 ohms= 26.7A when it starts and when it stalls. The IRF540 has a max allowed continuous current of 28A when will be extremely hot at its max allowed temerature even if it has a huge heasink. The IRFZ46 is slightly better.
The resistor feeding the gate should be very close to the gate pin. A 10k resistor to ground will protect the gate from static electricity when the power is turned off. |
|
|
|
boron
Mad Scientist
Canada
223 Posts |
Posted - Apr 20 2009 : 12:38:05 PM
|
Ah okay I see.
The mosfet number is actually IRF540N and has a max drain current of 33 amps, but that's still going to cause it to heat up a lot right? So if possible I should get a mosfet that can handle much more current otherwise would it be possible to wire 2 or more of these mosfets in parallel to distribute the current and cause each to heat up less?
I had added the diode already across the motor to dissipate the voltage spike. I take it that although the lm324 is not an ideal way to drive the mosfet, that it will still work because it worked nicely in unloaded tests? and what do you mean by its too slow.
Thanks for all your help. |
|
|
|
audioguru
Nobel Prize Winner
Canada
4218 Posts |
Posted - Apr 20 2009 : 1:13:40 PM
|
The IRF540 has a max allowed drain current of 33A when its case temperature is 25 degrees C. Then its chip temperature will be at its max allowed temperature of 175 degrees C.
But its case will be much hotter than that, depending on the quality of the heatsink and the airflow.
Then its max allowed current is determined by the quality of the heatsink and airflow.
The LM324 has a very poor high frequency response so the switching is actually ramps. The time that the LM324 is ramping is the time that the Mosfet is heating.
If you add another Mosfet in parallel then the capacitance that the LM324 is driving is doubled which also causes ramping because the output current of the opamp is low. Try it. |
|
|
|
Gus B
Apprentice
USA
149 Posts |
|
|
boron
Mad Scientist
Canada
223 Posts |
Posted - Apr 21 2009 : 2:27:22 PM
|
Ohh, okay. I understand the ramping now and that the current in the mosfet will cause it to heat up more and more as we approach the maximum current and that this value is dependent on the temperature of the mosfet which depends upon quality of heatsink and airflow, etc.
So I will try it with two wired in parallel and if that does not work I'll try a mosfet with a higher current rating.
I'll let you know how it works out.
Thanks a lot for your help, appreciate it. |
|
|
|
Gus B
Apprentice
USA
149 Posts |
Posted - Apr 22 2009 : 07:19:10 AM
|
| Calculating the power dissipation in the MOSFET isn’t quite that simple. You have to take the I²*R loss and multiply it by the duty cycle. Also, the Rds on of the FET will change with temperature, so you have to figure that in. You never said at what frequency you will be switching… that will matter (gate current and how it relates to switching losses vs. I²R losses). Total gate charge, not input capacitance (which is part of total gate charge) is what’s important when switching the FET. |
|
|
|
boron
Mad Scientist
Canada
223 Posts |
Posted - Apr 22 2009 : 5:42:18 PM
|
| The frequency its switching at is currently about 1.3 kHz. Yup, I understand how an increase in temperature causes increased resistance and power losses and lowers the maximum current that can flow before it blows. But I shall still try wiring two in parallel to split the load while attempting to keep them as cool as possible and see what happens. |
|
|
|
boron
Mad Scientist
Canada
223 Posts |
Posted - Apr 23 2009 : 8:17:07 PM
|
I have another question. On the diagram there is a resistor between the output of the lm324 and the MOSFET gate pin. I haven't been putting this resistor in place during the tests. Could that be a cause of this problem as well?
I performed another test with two IRF540N wired in parallel and this time the same thing occured. You give the scooter a kick start and engage the motor, upon throttling up it blows and kicks to full power. This time however the mosfets were warm afterwards and smelt.
So is that resistor between the output and the gate pin crucial and potentially also causing this to fail? If not, what can be done to stop this from happening? Will using the IRFZ46's and better cooling stop this from occuring?
Any suggestions would be appreciated. |
Edited by - boron on Apr 23 2009 8:35:28 PM |
|
|
|
audioguru
Nobel Prize Winner
Canada
4218 Posts |
Posted - Apr 23 2009 : 10:16:45 PM
|
Mosfets oscillate at a VHF frequency when the very important resistor is missing that is in series with the gate and mounted at the gate pin. Mosfets blow up when they oscillate at such a high frequency. Try 10 ohms.
Also the very important supply bypass capacitors are missing that should be a 0.1uF ceramic disc and a 220uF electrolytic capacitor mounted very close to the circuit. |
|
|
|
Gus B
Apprentice
USA
149 Posts |
|
|
boron
Mad Scientist
Canada
223 Posts |
Posted - Apr 25 2009 : 09:49:45 AM
|
Ah... okay I'll make sure to include that resistor this time round. (However on the circuits page it said it wasn't a necessity... I guess since I modified the circuit it now is...)
As in the diagram I have two capacitors, 0.1uF and 220uF going accross the 12 volt supply into the PWM circuit. There is nothing accross the 24 volts powering the motor. (Its set up so that two 12 volt batteries are powering the whole thing. one of the batteries does the PWM while both do the 24 volt motor, sharing the ground) Are these the caps you were reffering to?
Also, is there a minimum amount of current that must be fed to the gate pin of the mosfet? If so, is the lm324 capable of delivering that much current?
That's a pretty wicked pdf. I don't know if I can read all 37 pages, but its really interesting anyway. Thanks. |
Edited by - boron on Apr 25 2009 10:10:10 AM |
|
|
|
audioguru
Nobel Prize Winner
Canada
4218 Posts |
Posted - Apr 25 2009 : 10:16:55 AM
|
The lousy old LM324 is too slow to drive the gates of Mosfets. It takes all day and half the night for its output to switch. Its slow output is a ramping voltage where the Mosfet is half conducting and its chip is heating a lot.
The LM324 also does not have enough output current to quickly charge and discharge the high capacitance of the gate of a Mosfet.
If you use "normal" opamps that switch fast then two transistors can be used as complimentary (NPN and PNP) emitter-followers to drive high currents into the gates. |
|
|
|
boron
Mad Scientist
Canada
223 Posts |
Posted - Apr 25 2009 : 10:38:05 AM
|
alright. so what would be a suggestion for a "normal" op amp? I can get a TL084 which has the same pin output. Would that be better? Also in that case, what would be good transistors for the job?
Thanks. |
Edited by - boron on Apr 25 2009 1:11:55 PM |
|
|
|
Topic |
|
Electric Motor and PWM
