Paul & Sabrina's cheap DIY 144v motor controller

[MPaulHolmes]

The diode would point the other way. Otherwise it would just skip the motor altogether and when the mosfet was on, it would short straight to ground, destroying the mosfet instantly probably. ya!

[princeton]

I don't guess the revision had time to upload before your reply. Anyway, the drawing now shows the correct direction of the diode.

[princeton]

I think theory, the need for the diode can be overcome by large-enough capacitors in parallel with the motor. How large is needed to overcome the inductance is anybody's guess, though.

I think the best approach is a combination of a capacitor and diode in parallel with the motor along with a high-enough controller frequency. I am a bit concerned about having a diode in parallel with the motor without a resistor (in series with the diode). The currents through the diode may reach extremely high levels without a resistor. The problem with using a resistor is that it would limit the voltage dampening effect of the diode. Has anybody had a problem with the diodes burning?

[ElectricZX2]

I made myself very depressed tonight. I have an escort zx2 that I had a 72 volt controller in but it just wasn't working for my commute (couldn't keep up on the roads). I was trying to hold out for the1000 amp version but I couldn't do it and was able to find a deal on a lightly used Curtis 1231c that I really didn't have the $ for. Well tonight I think I may have blown it. I have 6 decent no name marine batteries and 5 crappy ones. I have 2 extra contractors outside of the ignition contactor. They are wired such that when the car ignition is on one or the other is closed depending on the position of a switch. One contactor connects the negative of the good batteries to the controller for 72 volts and the other connects the negative of the last bad battery to the controller adding them all together for 132 volts. I tested both positions after installing in my driveway and it worked but haven't had to drop out the crappers since, until tonight. I was in a hurray so I took the highway at 80 MPH which really drained the crappers fast. When I got onto the surface street I decided it'd be a good time to try the switch. I was going 40, turned of the ignition, flipped the switch, ignition and the volts were into the 70s. Had to shift for more torque. Could only get about 100 amps in 5th gear and speed dropping to 30 so I said screw the crappers I need more speed. So ignition off, flip switch, ignition on 130 volts, throttle does nothing. Back to 72 and still nothing. With the ignition off and the precharge resistor on I now have 0 volts, the controller doesn't charge. I didn't think the jump from 72 to 132 would be to hard on the capacitors just once. Could there have been some back emf from the motor or something? I had to do a direct bypass to the motor to get home...that was fun. Oh yea, I have a 9 inch GE. I was hoping to sell the Curtis to pay for the 1000 amp parts...now I have a piece of junk and no funds and no controller. Am I overreacting or am I screwed? Any thoughts?

[princeton]

I am sorry for your loss (said with head hanging low).

I doubt it will cost too much for a few mosfets to replace the blown ones. How are your soldering skills?

[princeton]

I thought these o-scope images might be interesting for some of you (especially those whose mosfets keep burning up).

The first image (img_5162#2.jpg) shows my PWM circuit running at 20KHZ running an IRF240 N-channel mosfet through an inductive load (coil of wire wrapped around an iron donut). I also have a resistive load of 4 ohms in series with the inductive coil. Power supply is a 12 volt car battery. The O-scope lead is placed on the drain of the mosfet. The peak voltage of the square wave is 12 volts but the peak voltage caused by inductive feedback and oscillation is close to 90 volts. The Mosfet is rated at 100 volts, so it's not blown.

For the next image (img_5263#2.jpg), I took a capacitor and 10 ohm resistor and connected them in series. Then I took this combination and placed them in parallel with the inductive load (from drain to V+). I tried simply using a capacitor alone in parallel with the inductive load but this caused too much oscillation. Notice how the peak voltage of the inductive feedback at the drain is blunted but significant oscillations remain.

For the next image (img_5264#2.jpg), I took a diode (reverse polarity) used as a "flyback diode" and placed it in parallel with the inductive load (drain to V+). Notice a mild reduction in the peak voltage from the inductive feedback. Also notice the mild reduction in the duration of oscillation.

For the last image (img_5265#2.jpg), I used both the flyback diode and the capacitor/resistor placed in parallel to the inductive load (drain to V+). Notice how almost all of the oscillations have been resolved using the combination of both.

If there's someone that could test this combination on their PWM controller pushing an actual motor (not a simple inductive load), it might help us to find a combination that allows us to use higher voltage batteries while preventing mosfet failure.

[flores]

Isn't be best combination the one used in de current schema of the open revolt controller? A diode and a capacitor? I seems like you haven't tried that?

My guess is that it would be difficult to find a resistor that can handle the current of a real motor..

[princeton]

I assume you are referring to this schematic at http://www.dancook.name/cougar_contr...r_power_2d.pdf

The schematic on the wiki shows the capacitors are placed in parallel with the power supply (between V+ and GND). I also tried this configuration but it had no noticeable effect on the output waveform.

The schematic on the wiki does show the "flyback diode" in parallel with the inductive load (motor). I tested this configuration and showed the O-scope output above.

I tested capacitor and diode in parallel with the load (from Drain to V+). See the attached circuit. The resistor I used was in series with the capacitor only, not in series with the motor. You are correct that a resistor in series with the motor would be a very bad idea.

[flores]

You're right..

I didn't notice that before. I always thought they were there to level out the voltage of the mostfet PWM output.
If the mosfet is closed, this is the same as being parallel with the motor, right? So what happens when the mosfet is open?

I don't really get it either..

[princeton]

You hit on my exact point. The high voltage inductive feedback hits the drain when the mosfets open (at which time the capacitors on the wiki would be effectively disconnected from the drain circuit due to opening of the mosfet switch). Curiously, the wiki circuit doesn't show any way of smoothing out this high voltage feedback on the mosfet drain except for the flyback diode. Without any dampening circuitry on the drain, the voltages can reach many times the supply voltage (as shown on the O-scope waveforms above) and easily blow any mosfet. If we're using 200V mosfets and 120V battery supply, the inductive feedback may reach 800 Volts or more, blowing the 200V mosfets. As the O-scope waveforms show, the flyback diode has a very limited ability to dampen the inductive feedback hitting the drain after the mosfets open. We need more than just the diode if we're going to prevent mosfet failure. Also, the diode tries (only partially successfully according to the O-scope waveforms) to shunt the high voltage drain voltage to V+. When this energy is directly shunted to V+, it wastes this energy and reduces the efficiency of the system. If we use capacitors to store this inductive feedback voltage between cycles, we INCREASE the efficiency. As I learned from testing, these capacitors cannot be placed in parallel with the inductive load without a series resistor. Without a resistor, there is nothing to dissipate the energy from oscillation from feedback between capacitor and inductor and sets up a vicious oscillatory cycle.

Please note: I edited this post because I had accidentally reversed the words open and closed, causing confusion).