Running an alternator and SMPS in parallel

Question

Can one connect a car alternator and a switched mode power supply in parallel on a network and run them at the same time without damage to either component?

My understanding given their implementation is as follows:

• car alternator is a 3-phase stator rectified by diode. Excitation voltage controlled by a voltage regulator circuit which stops excitation if voltage on the alternator output is higher than a threshold.
• looking into an SMPS, there are the smoothing transistors preceded by the rectifier diodes which rectify the transformer voltage. Voltage is controlled by PWM with the help of voltage sensing at output.

In consequence:

1. It should be ok to connect both in parallel as both have internal diodes which avoid short circuits.
2. If one sets the SMPS at a higher target voltage than the alternator one has a "priority" configuration in which the SMPS supplies (virtually all) the power to the network when it is switched on.
3. When it is switched off, the alternator takes over power supply seamlessly.
4. The smoothing capacitors in the SMPS will help to smooth the voltage in the network produced by the alternator (3-phase ripple) even if the SMPS is off.

Can anybody confirm this reasoning before I go and produce a cloud of smoke?

Answer 1

With a synchronous SMPS, you must add your own "OR-ing diodes" from the SMPS output to the power bus. It doesn't matter that most SMPS have diodes on the input. Most synchronous SMPS, when it sees its output voltage is "too high", will pull current from the output and store it on the "input" capacitor. If a person pulls the output voltage too high for too long, that "input" capacitor will eventually fail from over-voltage.

If a power supply -- such as practically all alternators -- have diodes on the output, then those diodes can do double-duty as the "OR-ing diodes".

With "isolation diodes", sometimes called "OR-ing diodes" (either the internal built-in ones of an alternator, or the external ones you'll probably need to add to a switching regulator, or both), you can connect two power supplies to a system and it will work just as you expect -- whichever power supply is turned on and generating the higher voltage will supply almost all the power to the load.

You may want smoothing capacitors at the load. When the SMPS is turned off, no current flows from the power bus "backwards" through the output OR-ing diodes into the smoothing capacitors of the SMPS, so those capacitors are effectively isolated and do nothing to smooth the voltage produced by the alternator.

"Can power supplies be connected in parallel for redundant operation?"

"Considerations for parallel power supply outputs"

"Using Isolation Diodes for Parallel-Redundant Operation"

"Connecting Switching power supplies in parallel is possible if certain precautions are observed."

Answer 2

The devil is in the details :-)

Your reasoning is largely correct and will work in general.
One always needs to be ware of boundary conditions and special cases. eg

The following are "makeups" - I'm asking myself what Murphy may attempt in such a case. It SOUNDS very reasonable, but long experience tells that Murphy is very adept when you do things that the manufacturer did not intend.

---

Is the alternator always going to be happy if its output voltage is high when it keeps backing off the excitation with no apparent, as far as it can see, effect. There seems no reason that an intelligent alternator controller designer would not make this so. Pray that you got an intelligent one. MAY it be better to put the whole alternator behind a final diode or near zero loss MOSFET recifier so it regulates happily at rated voltage when the smps Vout is higher. Rather than regulating at low voltage when the smps is running. If the alternator is backed right off by the smps and the smps drops out, will the alternator trying to pick up tyhe existing load from low level cause a surge? Will it cause a load dropout? Would this have occurred if alternator was running at about nominal voltage. This may be the biggest question of worth. Or not :-).

Does smps or alternator object to step changes on their output. Is smps stable as the two are almot equal and may tend to hunt to and fro. This is almost certainly not a problem - but it would probably be wise to perturb the system to and fro across the switchover boundary to see what it does.

If yu short an alternator or smps diode (rare but possible) does it destroy the other device and does it need to.

SMPS may nave N outputs and alternator usually has one, Is the SMPS output the main power one and does this matter.