I have a very large inductive load (a 20 kW DC motor) running at 200A and 100V at nominal load. I definitely need a flyback diode for the switched control circuit I am applying to it. However I cannot seem to find a part number. I presume that a diode for an application like this is somewhat rare? I wouldn't think so but I can't find much doing google searches.
Can anyone recommend a diode part number that would work in this case? I don't know the exact numbers but I would think that the forward voltage would have to be 1000+ and the amperage would have to be 400A or greater. Does that make sense? If so, where can I find a diode for sale like that? Thanks.
I have a very large inductive load (a 20 kW DC motor) running at 200A and 100V at nominal load. I definitely need a flyback diode for the switched control circuit I am applying to it.
Maybe not. What device are you using to switch 200 amps at 100 volts?
A high power MOSFET the size of a box of raisins. Switching frequency isn't a constraint so I can use whatever is convenient. It will be generated by the AVR and run through a boost circuit to feed the high power FET due to its large capacitance. I assume the need for large voltage ratings because of the laws of inductive loads whereby they generate huge voltages in order to maintain their current flow when the supply is interrupted.
Gahhhrrrlic:
I assume the need for large voltage ratings because of the laws of inductive loads whereby they generate huge voltages in order to maintain their current flow when the supply is interrupted.
The diode is there to short circuit that high voltage back-emf spike.
Many youtube videos about the subject.
I think to answer your question one would need to know a lot more about the motor, the driver and the configuration of the diode you are asking about. I'm assuming you plan to put the diode in parallel with the motor.
Are you planning to use only one diode? I've seen induction heater drivers with many Mosfets and diodes driving one coil.
The stud mount Vishay suggested is only rated for 50/60 Hz operation. Attempting to use such a diode at high switching speeds will cause large current spikes thorough the diode and Mosfet(s) due to the poor reverse recovery time. (the time it takes to sweep the majority carriers from the junction).
Thinking more on your question, lets look at what the diode is trying to suppress:
Motor coil inductance. Usually dissipated in a short fraction of the duty cycle. Not a lot of energy ( J = 1/2 L I^2 )
Motor back EMF. This will likely last for the full off time. If you are really drawing 200A, the diode will have to dissipate 200A x 1.6V = something over 300Watts during the off time. If this is the case, I would suggest you consider synchronous switching (using a MOSFET in place of the diode, driving the gate to reduce the losses).
The motor does indeed draw the full 200A @100V (real power) + field energy, whatever that amounts to. I would prefer to keep the circuit extremely simple to reduce failure modes. I can use a massive heat sink with active cooling or Peltier heat pumps or something if I have to but an over-designed diode that I can shunt in parallel and forget about would be a lot of sweat off my back if I could find one. The motor will run for no more than 15 seconds and more than likely it would be closer to 6-10 seconds before ample coold-down so heat dissipation is limited anyway. Just need something that won't blow up in 15 seconds of operation. The FET is rated at 400A... I will try to grab the data sheet.
Figures... I find a decent part and I can't buy it. Maybe it's an obsolete part number but I can't find it for sale.
To answer your questions:
I'll be PWM'ing the motor at close to 100% but it is throttl'able (it's a supercharger in a car) so I would say for much of that interval it will be at high duty.
I hadn't planned on keeping it on during spin-down but I can do it if it will help keep components from burning up, at least if it's just to cover a transient of a second or 2.
The motor is connected to a gear-train to multiply the speed of the output shaft so the inherent loss of efficiency would tend to slow the motor down quickly I would think. The field energy should sink quickly into that loss I would think.
Gahhhrrrlic:
I assume the need for large voltage ratings because of the laws of inductive loads whereby they generate huge voltages in order to maintain their current flow when the supply is interrupted.
You assume wrong. The entire point of a flyback diode is to tame the voltage spikes by providing a path for the inductive current to flow through (that's what free-wheeling is). The reverse voltage only needs to be able to withstand the maximum supply voltage.
Where did you try searching? Digikey's got several dozen diodes with more than 100A average forward current. Something like the APT100S20 is probably the minimum I'd use for something this big. 7.71 USD + shipping in single quantities.
I'll be PWM'ing the motor at close to 100% but it is throttl'able (it's a supercharger in a car) so I would say for much of that interval it will be at high duty.
High duty cycle for the motor means a lower duty cycle for the diode, since the diode only conducts when the driving transistor is switched off.
Call Powerex and ask. They were pretty easy to work with the last time I worked with them (a very specialized SCR)
The reason I asked about the PWM and duty cycle is the back EMF is created by two phenomenon.
a) A short voltage spike occurring at the very beginning, immediately after being switched off: This is the energy of the motor inductance trying to keep current flowing. If left unchecked this inductively generated voltage could reach very high voltage levels. The faster you switch, the higher they could be. See the attached waveform for a visual. Note the attached waveform is for a very small motor and the values (voltage, current, times) will not be even close to yours.
b) The generated back emf of the motor: This will be approximately the same voltage as the drive voltage. This voltage does not need to be suppressed (but will be by the diode).
If suppressed by diode, the motor will slow down more slowly due to the continued current aided by the diode.
If NOT suppressed, the motor will slow down more quickly due to the reduced current in the motor.
You might consider a 120v MOV (see Littlefuse) It would suppress the spike and let the motor back emf attain its "normal" voltage.
Note: This magnitude of drive current is a little out of my experience range but the concepts all apply. I usually work with motors with the currents in the 10 to 20 Amp range.
BTW How do you plan on evaluating what you decide? An oscilloscope? Watching for parts to blow up?
Due to the shear magnitude of voltages and currents involved and the inherent safety risk of getting close to the rotating assembly with any sort of sensitive equipment, I think scoping it would make me a bit uncomfortable. More than likely I will simply do some sort of crude break-in test whereby I cycle the motor back and forth between 0 and progressively higher duties. If it works at maximum switching levels a dozen or more times successively then maybe it'll be fine. Would a capacitor help at all in controlling the voltage spikes? I don't know a better way to measure several hundred volts in a fraction of a second except by decoupled sensing like hall effect sensors or amp clamps or something like that, which I don't have at hand.
First question: Is it better for the system for the motor to slow down quickly or more slowly?
If a faster slow down is desirable then try the MOV. It would be best to scope this solution.
I understand you concern regarding the scope. However these are not high impedance circuits. I would:
Divide the voltage in half by 2 x 10k resistors.
Connect them to a long twisted wire. Tie the wires down well (so they break before pulling on the scope.
Connect the other end to a 10:1 scope probe.
If slow is better then the Diode. You can check it is working by the heat on the diode after running it for a short period of time.
Large inductances can generate whatever voltage it takes to sustain the 100A or whatever. This means
a 10k:10k voltage divider is ludicrously inadequate!
Proper high voltage probe is required, and a sound understanding of electrical safety.
However these are not high impedance circuits
Au contraire! On a timescale of microseconds a large motor's winding inductance is an extremely high impedance.
I'm not sure I can agree with your assessment. Keep in mind what the goal is here, the verification a MOV is functioning.
And while on the whiteboard the voltages can be very high, in actuality winding self resonance and the likely hood of the MOSFET(s) breaking down (if the MOV is not quenching it properly) limits the voltage.
My suggestion was for safety. Finding and using a high voltage probe, near rotating machinery is not the approach I would take. Staying away from it with a 10 to 20 foot wire is a lot safer. IMHO
Au contraire! On a timescale of microseconds a large motor's winding inductance is an extremely high impedance.
How do you figure the impedance is extremely high? Are you concerned the 20K load would load the inductive spike to the point any readings would be useless?
JohnRob:
First question: Is it better for the system for the motor to slow down quickly or more slowly?
If a faster slow down is desirable then try the MOV. It would be best to scope this solution.
I understand you concern regarding the scope. However these are not high impedance circuits. I would:
Divide the voltage in half by 2 x 10k resistors.
Connect them to a long twisted wire. Tie the wires down well (so they break before pulling on the scope.
Connect the other end to a 10:1 scope probe.
If slow is better then the Diode. You can check it is working by the heat on the diode after running it for a short period of time.
Hope this helps
John
The way it would normally work in the field is, it would provide boost under load until the voltage is cut as a result of the acceptance criteria not being met (drop in throttle position, drop in manifold pressure, AFR too lean, etc). Once power is cut the blower will freewheel however the boost has to be purged through a blow-off valve and some of it will backdrive through the impeller. Also the 2 pulley gear train will provide drag on the motor. I believe the deceleration will be rapid but rapid on the order of seconds, not mS. The rotating mass has a lot of inertia, which limits responsiveness. I think the diode's heat dissipation would cause a noticeable rise in temperature that could be measured by IR or a thermocouple.
