Filtering capacitor is increasing the voltage, ideas needed

[Neil E.]

My last project (motorcycle LED tailights) was a great success thanks to ideas posted on this website. Now to the current project:
I am adding a computer case fan to one of my KTM radiators to get some cooling when I'm in a tight spot with no forward movement. Others have done this with good results, but they are using 12vdc battery power.
These machines have a split system with a common ground point part way through the windings. Most of the windings above the common point output AC that goes to the regulator/rectifier and exits as regulated AC for lighting loads. The other windings output AC to the regulator/rectifier that exits as DC for battery charging.
This DC is only about half an amp, but is sufficient for the intended purpose. My goal is to power the 12vdc fan from the main regulated AC power so as to not sacrifice battery charging. The fan would run anytime the engine is running which is fine by me. If I can't solve this, I'll install a thermostatic switch and use DC like everyone else.
There is 10.5 volts AC (RMS I believe on my fluke 73 multimeter set for ac) at the headlight with the engine idling at 1800rpm (redline is 10,000). With a 4 amp bridge rectifier connected here, the output is 8.6 VDC. This does run the fan, but at too low a speed. I thought I needed a filtering capacitor on the DC side, so I hooked up a 35V 10,000uF electrolytic. I was very surprised to see the DC voltage jump to 17.5 volts.
This of course is too high (great airflow though). Can someone please explain what is happening, and possibly suggest a better capacitor value?

 

[Papabravo]

What is happening is that you are not measuring what you think you are measuring. Let's take it one step at a time.

10.5 Vrms * 1.414 = 14.84 V-Peak
This is the peak voltage out of the rectifier, with the capacitor in place.
The capacitor will only take out the ripple.

You don't want to try to measure DC voltage at the output of the rectifier without the capacitor it is still AC and it swings between "GND" and 14.84 Volts

What we don't know is where Ground is with respect to either side of the headlight signal or the minus side of the rectifier. This may contribute to the 8.6 and the 17.5 volt measurement, neither one of them makes any kind of sense although it is hard to tell without a schematic.

 

[Neil E.]

I hear you on the schematic issue. Vehicle manufacturers never provide any breakdown of the stator coil connections. I understand what you mean about the ground. One of the headlight wires is actual ground. I have not tied the minus of the rectifier to ground. Should this be done? It might create a proper reference point for voltage measuring. I am concerned that doing this could affect the original equipment regulator/rectifier.

 

[Papabravo]

I know regualtors and power supplies. I'm definitely weak when it comes to vehicles and their wiring diagrams. I tried to help my brother with his BMW bike. It was in the vicinity of the headlight and it was a disaster.

Generally speaking you would not normally connect one side of a transformer and the rectifier output to GROUND. The only case where this might happen is with a center tapped transformer where the center tap is connected to GROUND.

 

[Russlk]

Some part of the alternator is grounded, so yu cannot connect the negative of the bridge rectifier to ground, it will be a short circuit. The signal from the alternator is probably not sinusoidal, so the RMS to Peak calculation does not work. The capacitor is holding the peak voltage. Connecting a resistor between the rectifier output and the capacitor will reduce the voltage. Measure the fan current at 12 volts in order to calculate the value of the resistor.

 

[Neil E.]

My understanding is that this sytem is single phase AC as opposed to many that are three phase AC. An aftermarket 3ph stator is available for these machines for the people who want maximum output. The voltage regulator/rectifier is replaced with a unit that outputs DC only.
I did hook up a scope to the headlight connections to look at the waveform. It was definitely sinusoidal. I couldn't stay connected long enough to verify the peak voltage due to the very fast temperature rise on this engine. I need to use a big external fan to allow more time for measuring. The computer case fan motor shows a rating of 0.22 amps at 12vdc. Assuming the same current would flow through a series resistor that has to achieve a voltage drop of 5.5 volts, I expect the resistor would need to dissipate about 12 watts. I don't see any point to wasting this energy, so running the fan from battery power is probably the best solution.
Thanks for looking at this, I'm still amazed at how great the internet is for sharing ideas.

 

[Hero999]

Don't use a resistor, use a voltage regulator like the LM7812, add it to the output of the capacitor. A resistor is no good since the voltage across the fan will vary depending on the speed of your engine and loading on the fan itself (which will alter the current draw).