dualing rectifiers

[kitestrings]

One last (?) cold snap here in VT, so there's a lull in our maple sugaring season.  Hopefully there'll be more time on this project after that.  I've been giving quite a bit of thought to the control scheme for the new windmill.  Below is a rough (very rough some will say) schematic that I was hoping to float out here for comments.  The basic strategy is unchanged - boosting in low-wind conditions - the windmill would be always connected to the batteries via one of two rectifier paths.  Background information is here:

http://www.fieldlines.com/story/2007/1/31/23935/3343

In low winds 7-14 mph (or ~60-140 rpm) the boost converter would be connected through a bridge rectifier to the batteries.  The converter would be current limited using a frequency to voltage source and re-scaling module to match, as closely as practical, the alternator characteristics.  The converter would be connected via a 3-pole, normally-closed relay.  This relay would bi-pass the converter above a preset cut-out/threshold.  The converter would be further protected by means of a high voltage limit (58-60 volts), as well as over-current protection on the DC input (~35 amps).  This rectifier would presumably utilize `low drop', lower power diodes.

Above the boost-mode, whenever the voltage exceeds the battery bank, current would flow directly via the second, higher current rectifier.  I am assuming that this path would never be broken.  That is, if the converter tripped out, or failed for whatever reason, we'd still be connected and loading the alternator; all-be-it limited to production at higher winds.

I have several questions, besides the obvious, "How feasible is this?" assuming it is:

1.     Is there any advantage to switching on the 3-phase side vs. switching the DC output from the rectifier?  I assumed the current would be lower, and perhaps less arching as shown.
   
2.     What should we be looking for to specify "low-drop" diodes?
   
3.     Can the two electrically isolated rectifiers share the same heat-sink, or is this risky?   My existing generator utilizes heat sinks as the side panels to the control box.  It seems pretty tidy.
   
   

As always, comments are welcome and appreciated.

-kitestrings

[Flux]

The idea will work if you can sort the details. I assume that for some reason your commercial converter requires the relay. I don't use anything to break the boost converter circuit, just stop the pwm pulse. If you can switch the relay after you are on the main bridge then I don't see a problem but keep on the ac side as shown.

If you are timing it to operate before the converter has forced the change then you may have fun.

For a 48v system I wouldn't worry about low drop diodes. At 12v it would be critical, I didn't bother at 24v and at 48v I definitely wouldn't bother. The only low drop devices are schottky and their voltage ratings will be an issue at 48v. Even if you find extra high voltage ones you will find that their drop is higher than their low voltage relatives.

There is no reason why the 3 negative diodes can't be common to both bridges. I used 3 heat sinks, one with 3 negative diodes and 2 with positive ones. If you want to use isolated bridges then One sink can be common to both.

Even when enlarged I can't read all of your diagram, but if I could it wouldn't help much as I know nothing about your converter.

Flux

[electronbaby]

If your worried about voltage drop,... Flux is right, I wouldnt worry about it in a 48v system. Here is something that might be of interest. Full wave bridge using FET's. Very low voltage drop. Maybe you could adapt it to a 3 phase bridge design.

http://www.thetaeng.com/FETBridge.htm

[SamoaPower]

Unfortunately, the FET bridge, as shown, has some issues that would not make it suitable for high current battery systems. They point out the caveat for capacitor loads which would also apply to a battery load.

There is also an issue with switching the FETs with a sine wave. This would result in higher switching losses that to me, wouldn't be acceptable.

There is, however, another way using a comparator to detect polarity and switch the FETs quickly. There is an integrated example of this in the NIS6111. Unfortunately, it's only suitable for 12V systems. Using other components, the same design can be applied to higher voltages.

I've designed a 150A 3-phase bridge using this approach that's currently under construction (am doing the PC boards now) that should dissipate only about 10 watts. It uses a total of 30 FETs, yet only occupies a volume of 3" x 6" x 1" including heat sinks and at a cost of about $50 in parts. I'll post it when done and tested.

Even for 48V systems, diode voltage drop isn't the only problem associated with the rectifier. At 100A, a conventional diode bridge will lose (and dissipate) about 200 watts requiring large heat sinks. I, for one, don't want to throw away this much hard-earned power.

[kitestrings]

Folks,

Thanks for the comments.  The primary concern on the diodes was based on an an apparent requirement for the converter that the input volts needed to be 2V below output.  In an earlier post the low-drop diodes were recommended due to this.  I have sine clarified with the manufacturer that limitation can be reduced to "an IR drop."  Does this requirement change anything?

I like the idea of a common negative diode(s) bridge.  I'm assuming that there is no harm in leaving the normal (non-boost) bridge connected even when the voltage is low and the converter is needed?

I'm sorry about the quality of the pic.  I tried to scan it, but could only seem to get a pdf or tif file (which I couldn't seem to convert).  I resorted to a screen-shot/jpg.  Here's a slightly better shot of an admittedly ugly schematic:

Samoa- Your FET bridge sounds promising.  You could nearly cook an egg on my existing (12V) heatsinks in prolonged high winds.  Lot's of room for improvement.  I look forward to seeing your progress with this.

Kind regards,

-kitestrings