I'm chasing alternator whine due to ground-loops in audio circuits in an aircraft. Since it is dangerous (and noisy) to work around a spinning prop, I'd like to be able to connect an "alternator simulator" in place of the real engine-driven alternator and have it produce about 14.2V D.C at about 30A (to run the normal aircraft loads), with an 1 to 2 kHz full-wave-rectified ripple component riding on the D.C., just like the real alternator would have.
An aircraft or automotive alternator is basically a three phase sinosoidal current source (high impedance output) feeding into a six diode full-wave rectifier. The rectification creates six overlapping current pulses, so the ripple is well defined and determined from the three phases.
After the battery is recharged, the average D.C. current from the alternator supplies the loads; but almost all the ripple current flows into and out of the battery. The minimum system bus voltage is determined solely by the battery.
Since the airframe is used as a ground return for everything in an airplane (or car), the ripple current can couple into any audio circuits that connect to the airframe (car body) in multiple places as a common-mode noise. The solution is to identify and break those ground loops. I would be using the alternator simulator box to create the ripple currents as would be flowing if the engine and alternator was turning.
I'm thinking I could do this with an electronically-regulated 30A D.C. power supply which starts about 1V higher than the battery voltage (nominally 14.25V with the engine running) in series with a giant NFET connected as a source follower, whose gate is driven with the synthetic ripple waveform.
The source follower can only source current (not sink it), but that is the way the real alternator works, i.e. if the instantaneous alternator voltage is less than the battery voltage, current flows out of the battery to the loads. During the part of the cycle when the alternator output is higher than the battery voltage, current flows from the alternator to the loads and to the battery to put back the charge that was "borrowed" during the previous low point. The battery acts as a giant filter capacitor, but has a finite series resitance.
The obvious problem would be the power dissipation in the NFET. It acts as a Class A amplifier, with an average current of 30A modulated by the ripple. The average drain to source voltage would be about 1V, so the dissipation would be about 30W. Should work as long as it is mounted on a large heatsink.
Can anybody suggest a better way to emulate what the alternator does?
An aircraft or automotive alternator is basically a three phase sinosoidal current source (high impedance output) feeding into a six diode full-wave rectifier. The rectification creates six overlapping current pulses, so the ripple is well defined and determined from the three phases.
After the battery is recharged, the average D.C. current from the alternator supplies the loads; but almost all the ripple current flows into and out of the battery. The minimum system bus voltage is determined solely by the battery.
Since the airframe is used as a ground return for everything in an airplane (or car), the ripple current can couple into any audio circuits that connect to the airframe (car body) in multiple places as a common-mode noise. The solution is to identify and break those ground loops. I would be using the alternator simulator box to create the ripple currents as would be flowing if the engine and alternator was turning.
I'm thinking I could do this with an electronically-regulated 30A D.C. power supply which starts about 1V higher than the battery voltage (nominally 14.25V with the engine running) in series with a giant NFET connected as a source follower, whose gate is driven with the synthetic ripple waveform.
The source follower can only source current (not sink it), but that is the way the real alternator works, i.e. if the instantaneous alternator voltage is less than the battery voltage, current flows out of the battery to the loads. During the part of the cycle when the alternator output is higher than the battery voltage, current flows from the alternator to the loads and to the battery to put back the charge that was "borrowed" during the previous low point. The battery acts as a giant filter capacitor, but has a finite series resitance.
The obvious problem would be the power dissipation in the NFET. It acts as a Class A amplifier, with an average current of 30A modulated by the ripple. The average drain to source voltage would be about 1V, so the dissipation would be about 30W. Should work as long as it is mounted on a large heatsink.
Can anybody suggest a better way to emulate what the alternator does?