Rectifier losses

[GerryS]

 I read that the rectifier losses are about 0.7V per diode and I also read that rectifier losses are 1.4 volts in a 3-phase system using 2 bridge rectifiers. Are  three-phase rectifier losses 2.1V since three diodes are involved (2 in one rectifier and one in the other)?  Or is it 1.4V lost?  If it is 1.4, how come the third diode isn't included? Upon testing, each phase of my 3-phase system is producing about 3.2 volts spun at 170RPM. I know the total output of the 3-phase system is 3.2x1.73=5.5V.  Does this mean that the open circuit DC volts coming from the rectifiers is going to be 5.5V-1.4V=4.1V?  

[asheets]

Your rectification losses are going to be based on the diode/bridge actually used.  For instance, if you use a National Semiconductor unit, the spec sheet will tell you exactly how much you will lose.

[ghurd]

It's only 1.4V.

Either 3 diodes at 0.7V are conducting, but 1 is in series with 2 in parallel, for 1.4V,

Or only 2 in series are conducting for 1.4V.

Is that 3.2V raw AC or rectified DC?  You probably missed a step or are counting it twice.

G-

[alancorey]

I agree with ghurd, it's 1.4 volts, since you mention the 1.73 factor I assume you're wiring in star.  If you were using Jerry-phased with all 3 phases rectified and then wired in series you'd drop 1.4 * 3 volts, but you'd have 3 times the voltage to start with.  At least I think that's what Jerry-phased is.

I was looking for a place to steal an image of a bridge rectifier schematic symbol and stumbled across this discussion: http://www.faqs.org/docs/electric/Semi/SEMI_3.html

About halfway down they show some 3-phase stuff.

The thing I was going to point out is that you lose roughly 0.7 volts per diode, but a bridge rectifier has 4 diodes in it.  How much loss depends on how you connect it.  The link above discusses it better than I can.

  Alan

[finnsawyer]

The three phase star actually uses six diodes two of which will be forward biased when current is flowing into the battery.  That gives about 1.2 - 1.4 volts.  The rectifier loses are actually the sum of the diode voltages times the instantaneous current flowing into the battery.  The diode voltages will change slightly with current, but assuming they are constant works O.K.  Here again is another argument for getting an oscilloscope.  You might try analyzing the circuit to see when diodes are conducting and when they are not.

[GerryS]

I have looked at the bridge rectifiers and gone through their electricity flow.  It seemed to me that 1.4 V was the proper approximation, just wanted to make sure with the experts.  Yeah, I think I'll run down to Sears at lunch and buy an Oscilloscope!!! Jokin.   I have access to plenty of them at work.  Actually, I was thinking about setting up a data acquisition system using NI Labview.  I have access to some pretty high frequency analog input boards and could easily write a program that would have the same function as an Oscilloscope...I think.  Is there a big difference between using a high frequency data acquisition system -vs-using an oscilloscope?  I like the concept because with Labview I could build in some pretty good data analysis and presentation.  Could also present history data well and overlay graphs for comparisons etc.

[finnsawyer]

Well, there are a few things you need to do the same things as an O'scope.  An imput impedance of one megohm, adjustable hi-gain input amplifiers, an adjustable stable time base, and an X-Y mode.  There's no point making something that will almost do what a scope does.

[SamoaPower]

GeoM,

"The diode voltages will change slightly with current, but assuming they are constant works O.K."

Not really. Power diodes will almost double their forward voltage drop at rated current. Not trivial when condidering power loss and heat sink calculations in a high power system. Check the data sheets.

Diode type also makes a substantial difference. Schottky will have about half the voltage drop as garden variety silicon. For lowest loss, FETs used as synchronous rectifiers are the best.

[GerryS]

I was just thinking of it for use in voltages and currents for the most part.  Literally I guess, it would just be a data acquisition system for volts and current produced.  I guess, in the end, it would just be a really fancy multimeter.  Actually, I don't even need to write any programs.  I have a Keithley multifunction multimeter running a program already.  The oscilliscope is better for what it actually does.  Your point is correct about not reinventing the wheel.  But it's pretty cool to be able to watch constant graphs of the alternator output and be able to play with them on the fly, zoom in or out, etc.  The thing I think is really cool is being able to have 3 or 4 charts displayed at the same time...power, volts, current, RPM, ...etc.  I like the thought of having analyzed data/worked-up data fields also charted on the fly.  But alas, it would not serve the same function as an oscilloscope.  It would just be a pretty cool data acquisition system.  

[GerryS]

Hopefully most people are not using diodes near their rated current.  I think it's good to engineer designs to use rectifiers with twice the rated current than what you think will be peak current, which makes the voltage drop from current changes a little less of an issue.  Is the voltage drop in diodes from current increase a linear or exponential relationship?

[finnsawyer]

For the kind of stuff you want to study you need both the waveform shapes and amplitudes and an accurate time base or way to measure time.  I have concluded that you would be well served by enrolling in the local college and taking courses in circuit theory, electronics, and electromagnetic theory.  You seem to have the math background, but need a good understanding of the physical models that have been developed for the various phenomena.  Oh, I would also suggest a course in advanced calculus for engineers.

[GerryS]

Accurate time base is what deterministic programming/data control/acquisition systems are all about. I would indeed be well served taking circuit theory, electronics and electromagnetic theory.  In chemistry grad school I concentrated in an area between electrochem and inorganic molecular orbital (electronic) theory (the physics side of chemistry!).  So while I have studied the fundamental phenomena that commonly underlie boths fields, I have not studied the physical models that directly apply to alternators.  You conclude correctly. At first I felt your comment was rude, thinking there's no way you could possibly know what education I've already had or courses I've taken.  But then I thought about it and decided that you were correct and there's point in being offended.  I agree fully about the accurate time base being important.  In my experiences it's really important for kinetics and thermodynamics studies of chemical reaction mechanisms that take place in fractions of seconds and must be monitored.  NMR is another area where it's been really important.  All about the rate a molecule spins and precesses at in a very strong magnetic field.  We use the low temp to slow down the reaction process and we can see reactions taking place on a molecular time scale...accuruate timing is very important.    

[finnsawyer]

One can build up an idea of a person's education from how they deal with issues over a number of postings.  That's why I made the recommendation as I did.  I was not being rude.  In fact, now I see that with your background you are quite capable of getting the knowledge that would help you from books, such as those of Circuit Theory, Electronics, Physical Electronics, Electromagnetic Theory, Advanced Calculus, and Vector Calculus.  I have degrees in Electrical Engineering, and before posting here I read a book on Aerodynamics by one Dan Smith that turned out to be quite a gem, as at the time my interest was in blade design, not alternators.  Maybe I should have stayed away from the alternators.  In any case I wish you luck in your endeavors.    

[GerryS]

ha ha ha.  You got a good chuckle out of me for "Maybe I should have stayed away from alternators." At this point in my life time has become the biggest issue.  I'm getting married in a month or so and ,God-willing, starting a family shortly thereafter.  My company would readily pay for any education I want in electrical engineering.  But time, oh time.  Not just time, I think advancement is a big issue.  Where I am now, furthering my education in science will not get me much further.  So if I were to go back to school and spend that much time doing it, it would be difficult to justify it for a "hobby."  Whereas something like an MBA degree might be justifiable as it would open up good career advancement opportunities, which makes the time commitment more justifiable to my soon-to-be wife and eventual kids.

[finnsawyer]

Well, I wish you the best of luck.  Just keep in mind that when you have a fight with the wife you can always curl up with a good technical book while she glares at you.  Naw, of course that won't happen to you.  Me?  Been there, done that.  Three beautiful daughters, five grandchildren, one divorce.  All my ex's live in Texas.  Yes, she does.

[GerryS]

Oh, you're one of those guys that only makes baby girls!  (bet you've heard that a million times!)   Thanks for the good luck wishes!  I've got my technical books all ready for those good times!