LTspice three phase rectifier simulation

I am interested in creating from scratch a regulator rectifier for motorcycle. I struggle to comprehend the ground placement in the simulator. I cannot find instructions about tree phase simulation. The difference between the two is the additional ground that there is on the schematic on the right. I probed different angle phase on each so you can see clearly. In the second case each phase does not go under 0V when in the first case on the left the sine is as expected. Could you tell me how to simulate this properly as eventually when a battery, load and regulator come into play it will be a common ground with the diodes. I guess my error is in the phase representation. I hope my question is clear.

• You can't have the ground point as shown in the circuit to the right, the diode will conduct right across the source during one half cycle. Tie the three sources together in a "Y" configuration without connecting to ground, and then you can put the ground as shown in your circuit on the right. If the simulator complains, put a 10Meg resistor from the common point of the sources to ground. Jul 16 '20 at 20:24
• Is the generator a delta or Y configuration? Jul 16 '20 at 20:48
• @JohnD thanks but I get similar results with this solution Jul 16 '20 at 22:15
• @VoltageSpike It is Y configuration. Jul 16 '20 at 22:17

It's important to note that a rectifier's positive and negative outputs are the highest and lowest voltages from the input respectively. That mean that the negative output isn't ground. By grounding the negative output in the right circuit, you short any input that has a negative voltage.

You can still use the circuit on the left but you need to set the negative output as some sort of virtual ground. Connecting the two grounds will cause a short circuit, and the two grounds will oscillate with respect to each other. The easiest thing to do is ignore the ground from the generator, so you treat it as a delta source rather than wye.

• I suspected this would be the case, it is just when the rest of the circuit is being added it becomes a mess as the measurements the simulator gives are confusing, I wish I had an actual alternator or a guinea pig motorcycle and all this would not had been necessary. Jul 17 '20 at 7:27
• I recommend setting the negative output of the rectifier as the actual ground. Tie the negative terminals of the three sine sources together but don't connect it to ground. That should make the rest of the circuit a lot simpler. Jul 17 '20 at 18:01

The stator phases aren't grounded. They should be tied in the middle together in a wye configuration, with each phase feeding the rectifier. So replace the three grounds on your voltage sources with a single net, but don't ground it.

You effectively do that on the right-hand schematic so long as you don't also ground the (-) output of the rectifier. The left-hand one fails because you're shorting out the stator wye common to ground.

See this diagram:

Try a Falstad sim here: 3-phase regulator sim

• Just to be on the safe side, simulation-wise, the center of the Y connection could be tied to ground through a high enough resistor, 1meg...1g or so. Jul 17 '20 at 6:32
• @hacktastical Thanks for the Falstad sim, this is what I need as I cannot do that with LTspice, that was exactly my problem, I know that the alternator is not connected to the ground but LTspice needs to do that, it does not have AC generators or I haven't found them, this was my question anyway: if it is possible to simulate the alternator like Falstad does. If I do the schematic in LTspice like a real Y alternator (without ground) it does not simulate properly. Jul 17 '20 at 7:32
• The AC sources are two-terminal, so you should be able to tie them. No need to provide a return path. Jul 17 '20 at 23:00