The amount of electronic and electrical equipmentin today’s cars has increased exponentially over the past decadeand we can look forward to even more in the near future. The demandsfor electrical power are increasing and, consequently, greateroutput currents are needed from the alternator. In this article,we will examine what determines the output current of an alternatorand will look at a high output alternator originated from stockparts, as well as a "built from scratch" unit.
In the early 19th century, Michael Faradaydiscovered that if you move a conductor within a magnetic field,a current will be caused to move within the conductor. It is onthis principle that the modern alternator operates. The rotorcoil generates a magnetic field that is intercepted by the statorcore that passes this "changing" magnetic field intothe stator coil, in turn producing an electrical current. Thealternator is a transducer that converts mechanical energy intomagnetic energy, then magnetic energy into electrical energy.
It’s not possible to cover every detail ofalternator design in the confines of one article. However, byknowing the basics of unit operation and testing methods, onecan understand how to generate more output current from a givenalternator.
In modern electronics most electronic componentshave a defined "mathematical model" developed througha process called finite element analysis. Information is generatedwhich shows how a bunch of electronic components will behave whenconnected together in various ways. Models have been developedfor magnetic devices such as transformers and solenoids. But tomy knowledge, no one has published a program for designing analternator.
If such a program were available, you couldkey in the basic parameters such as speed range, output voltageand current, expected temperature rise, expected current at idlespeed, expected stator wave form, maximum speed required, etc.A click of the mouse button and your computer could then tellyou what stator core size to use, recommend thickness of laminations,determine how many turns and what stator wire gauge to use, statethe rotor pole geometry and cross-sectional size, recommendedgauge rotor coil wire and how many turns should be employed.
Furthermore, your computer could determinewhat size rotor shaft is required plus the bearing size, whatpitch fan to use and the minimum size of the shell required tohouse the unit. Unfortunately, until such a program is available,the design of an alternator is almost as much art as it is science.
By applying basic electrical principles toalternator design, though, we can learn a lot. And, there aremore than 30 years of operating alternators in the field. As theold saying in the engineering field goes, you don’t have to reinventthe wheel to improve upon existing designs.
The stator core
If you disassembled an alternator, you wouldnotice that there are no wires attached to the stator core fromany kind of input energy source. The only energy that can be passedinto the stator core is the magnetic energy from the rotor. Astator core does not care if its coils are generating 1,000 voltsat one ampere or one volt at 1,000 amperes. The key specificationof a new alternator design is, therefore, "output power."And output power is the product of volts times amperes.
If you take off one of the end caps from analternator and view the relationship of the rotor pole piecesto the stator core, you should realize that the magnetic forceof one of the rotor poles is trying to get to one of the adjacentpoles that is mounted on the opposite side of the referenced polepiece. To get there, the magnetic flux travels outward into thestator core, goes both left and right, then down to the oppositepole piece.
The output power of the stator core is thendependent on how much magnetic energy the core can handle. Thereis only a limited number of materials suitable for core construction.Non-grain oriented soft steel is the most common material. Theamount of magnetic flux this material can handle is almost constantdue to the natural permeability of this type of steel.
The key then is to apply a magnetizing forceto the steel. Using a gauss meter (a device for measuring theamount of magnetic flux), find the point where further increasesin magnetizing force does not increase the amount of flux densityof the core material. When increases in magnetizing force do notincrease flux density, the core is said to be saturated.
The only way to get more output current froma core is to increase the size, or more precisely, the cross-sectionalarea of the core. For simple closed-loop cores, output power equals0.5184 times the frequency (in Hz) times the cross-sectional area(in inches) squared. This equation is only a ballpark approximationof how much power you can get from a given cross-sectional area.
An alternator stator is a complex device inthat the stator core is composed of segmented sections for eachpole. A six-pole stator (one with six fingers in each rotor polepiece) can be thought of as six sub-cores where each core onlyhas one-sixth of the total stator winding. In effect, each coresection is in parallel so that the total cross-sectional areaof the core need be only one-sixth of that computed by the coresize equation.
The stator receives its magnetic energy fromthe rotor. As the rotor must rotate, an air gap is required; foreach pole there are two air gaps. An air gap has high reluctance,therefore, it consumes a great deal of magnetic force, loweringthe magnetic force across the stator core cross-section. Withlower magnetic force across the core, the flux density decreasesrequiring a larger core.
Following this line of thought, the core sizeis decreased by adding more poles to the alternator, but increaseddue to the air gap. The force across the air gap increases bythe square of the distance. For example, if the original air gapwas 5 mils and you reamed out 2.5 mils from the center of thestator core, the gap distance would double, dropping four timesthe magnetic force. The size of the air gap is extremely criticalin alternator performance.
From the core size equation note that the poweroutput is (within limits) proportional to the frequency, but itincreases by the square of the cross-sectional area. If you havea 1,000 watt stator that you want to make into a 2,000 watt stator,you would have to increase the cross sectional area by the squareroot of two or 1.414ý. While this is not exact, it willbe close for testing.
The ideal cross-sectional shape of the coreis a square in terms of the stator winding. A square has the smallestperimeter for the area. Rectangular shapes use a lot more statorwire for the same cross-sectional area. The exact wire gauge sizeto use depends on the efficiency of your alternator. Larger wire(smaller numeric gauge size) runs cooler, but you may not havespace for the larger size.
Magnet wire manufacturers have charts tellingwhat gauge wire to use for a certain temperature rise. If yourwire has a high temperature insulation you can sacrifice efficiencyand just let the stator run hotter. The energy consumed by thewire is called copper loss and is equal to the TRMS (True RootMean Square) value of the winding current squared times the windingresistance.
You have a choice of connecting the three windingsin either a delta or wye circuit. The delta lets you use a smallerdiameter wire, but requires more turns as each coil must developthe full output voltage. The wye needs a larger diameter wire,but requires fewer turns as two coils are in series.
You don’t get twice the voltage as the twowindings are 120° out of phase; you will get the square rootof three times a single winding voltage. As a delta configurationforms a closed loop with the three windings, any imbalance inthe turns ratio or wire length results in a large loop current.The idea of an alternator is to provide an output current; deltaloop currents take away from the output current and just causeexcess stator winding heat.
Delta coils should be tested for uniformity.Just place a TRMS-AC ammeter in the delta loop (while not drawingany output current) and run your test alternator at full speed.A couple of amperes is okay for loop current as you will neverget a perfectly balanced stator. But more than this requires adjustmentsto the coil winding equipment.
The stator core has losses called hysteresislosses due to the continuous remagnetizing of the core material.Eddy current losses also exist due to the fact that steel is aconductor that sees the same changing magnetic field producedby the rotor. To reduce hysteresis losses the core steel mustbe physically soft.
A permanent magnet is hardened steel and isdifficult to demagnetize; soft steel does not retain magnetism.The laminations should be annealed to soften the material. Annealingalso oxidizes the laminations to provide lamination-to-laminationresistance. An odd thing about oxidation is that most base conductorssuch as copper, silver, steel and aluminum are good conductorsuntil oxidized. Most metal oxides are good electrical insulators.
Laminations should be treated like fine china.Steel is hardened by forging, a pounding process; tossing a statorcore into a bin does the same thing. Hardened steel requires moreenergy to constantly magnetize it, resulting in more core heat.Electricity intended for higher output current is used to heatthe core.
To reduce eddy current losses, the core islaminated to form thin sheets. The reduced cross-sectional areaof each lamination offers a greater resistance to current. Superthin laminations would seem to be ideal to minimize eddy currentsand such is true in terms of just eddy currents. But the oxidecoating on each side of the lamination is not magnetic, therefore,non-magnetic regions are formed that somewhat defeat the purposeof the core.
If the laminations are made too thin, a parametercalled "stacking factor" enters the equation requiringthat more laminations are added to provide the minimum cross-sectionalarea resulting in a thicker core. A thicker core requires longerwires to form the stator coils, increasing the stator resistanceresulting in additional copper losses.
In terms of copper and core losses, the idealstator assembly will be designed in such a way that the copperlosses are equal to the core losses. This power loss equalityresults in the most efficient stator; and an efficient statoroutputs more current. Use a TRMS ammeter in series with each statorcoil, and multiply this current times each stator coil’s resistanceas measured by an ohmmeter. Adding the three coil powers togetherprovides the copper loss values.
Measuring core loss is not as simple. Coreloss varies considerably with frequency. Typically, most alternatorswill not output more current when the speed approaches 6,000 rpm.This is due mostly to core losses. At these higher frequenciesthe core cannot be remagnetized fast enough and the eddy currentsincrease drastically.
Increasing the speed results in a flattenedcurrent curve. The speed where the current flattens out is a goodpoint to check core losses. Another frequency of interest is thenormal idling speed of the alternator; concern about output currentis important at this speed.
In power transformer design, determinationof core losses is easy as the input, output and copper loss powercan be measured. Simply subtract the sum of output power and copperloss power from the total input power and you have core loss poweras the difference.
Such is not the case in an alternator as theinput power is mechanical rather than electrical. To measure mechanicalinput power a device called a rotary strain gauge transducer ismounted between the drive source and rotor with some type of tachometer.The strain gauge transducer provides input torque and the tachometerprovides speed so that input horsepower can be calculated.
An easier way to determine core loss, say at6,000 rpm, is to divide this speed by 60 RPS and multiply thisresult by the number of pole pieces. For six poles this givesus a stator frequency of 600 Hz. With data provided by your corelamination supplier, this frequency will provide a certain numberof watts per pound of core loss. Simply weigh your coil-less coreand multiply it by the watts per pound value to learn the totalcore value.
With core and copper losses known, you canadjust each by changing core geometry, lamination thickness, laminationtype and wire gauge size until you find the optimum value. Soundlike work? It is! Even though you may never design a stator, youshould understand what goes into its development and gain an appreciationfor how critical stator parameters are.
Determining the number of poles is the nextitem in alternator design. Generally, more poles result in a greaterstator frequency that would cause a stator to output more currentat idle speed. If this is your goal, use more poles. But addingmore poles results in thinner pole sections resulting in fingerfluctuations at high speeds which require a larger rotor-statorgap.
Core losses also increase with more poles,limiting the high-speed output of the alternator. When designingan alternator, the phrase "trade-offs" enter the picture.When you improve one parameter other parameters suffer. A goodalternator design depends on optimizing each parameter to getan overall balance in the design. It also takes skill and patience.
There is a choice to be made of either rotatingboth the pole pieces and the coil, or just the pole pieces arounda fixed coil. The rotating coil method requires the familiar slip-ringsand brushes to get field current into the field "electromagnet."The fixed coil with rotating fingers option is far more reliableas the troublesome brushes are eliminated. But it does add weightand cost to the design. The fixed coil also adds another pairof air gaps in the magnetic path requiring both a larger statorcore and pole piece. Again, more trade-offs!
The purpose of the rotor is to provide thestator with a changing magnetic field at the correct amplitude.Let’s talk about the changing magnetic field first. A voltageis only induced into the stator if the magnetic field is constantlychanging; how much voltage is a function of the "rate ofchange."
If a rotor were made of parallel bars equallyspace around the circumference, with half the bars tied to thefront of the rotor and the other half to the rear side, the statorwould see alternately no metal and full metal at any one pointwith rotor rotation.
You’d think this would generate an ideal squarewave to minimize diode dissipation, but the only change that occursis from air to metal and vise-versa. This stator wave form wouldbe a series of sharp positive and negative spikes. The changein magnetic field strength, and hence, flux, will only occur whenany point changes from metal to no metal to metal again.
During the rotation where no metal or fullmetal occurs, there is no change, thus no induced voltage. Fromthis analogy, you should get the idea that the shape of the polepieces and the rate of how much metal between the rotor and statorcoils changes determines how much output current an alternatorcan produce.
Pole piece design can inhibit alternator performanceas much as it improves it. To induce maximum power into the statorcore, the best wave form is a sine wave. But a sine wave is notthe best wave form for the main diodes. A sine wave has a highpeak voltage that causes all the diode current to flow just duringthis peak.
To get an average current of 100 amperes, thepeak current can be as high as 900 amperes! This results in ahigh Vsat and lots of wasted diode power that not only reducesthe output current but burns up diodes as well. The high peakcurrents also drastically increase stator copper loss. You mustuse TRMS voltage and current values to learn true power dissipation,not average DC values. DC values are much lower than TRMS andare inaccurate for determining the actual power consumed.
The worst wave form is where the air spacebetween the adjacent rotor pole pieces is too large. This resultsin short, high-positive peaks that greatly increase the TRMS powerin both the stator coil and main diodes. One of the best waveforms was originated by the Delco CS-series where a concave surfacein each pole finger inverts the positive peak. With slow gentlecurves in the pole fingers, the rate-of-change parameter is maintainedwhile providing an almost square wave-like output to the diodesfor a lower peak-to-average current ratio. This lets the diodeconduct longer and drastically reduces the TRMS power dissipatedboth by the diodes and stator coils.
If the design of an alternator depends on aparticular stator wave form and you change it by using a pooraftermarket reproduction, you are asking for problems for youand your customers. If an alternator, however, is designed tosurvive with high positive peaks and you add an improved waveform such as the CS-type, you will be providing a superior productwith little increase in cost.
Also, be aware that the concave surfaces generatea higher frequency component that translates into greater coreloss. If a lower quality core is used, expect more heat and shorterdiode life.
Rotor pole pieces must be of very soft ironconstruction. You cannot do any machining to a rotor without generatinghard spots. These hard spots act as permanent magnets increasingthe minimum output current, i.e., your alternator will alwaysovercharge.
Tossing a rotor into a rotor bin is beggingfor a return; like the stator core, rotors should be treated likefine china. The soft rotor pole pieces are another trade-off ashardened steel should be used to withstand the centrifugal forcein a rotor that is rotated above 20,000 rpm.
On one hand we want a lightweight, hardenedsteel rotor to survive high speeds. On the other, we want a heavypole piece, soft iron rotor for both high output range and current.We can’t have both so the best compromise must be found. Rotatingharmonics also enter the picture as harmonic oscillations cancause a perfectly balanced rotor to fly apart. Most stock rotorshave an ample amount of pole metal due to the mechanical requirementsof the rotor.
Here is a simple trick to learn how much reservea stator-rotor combination has. With ample size rectifier diodesinstalled, feed the output of your test alternator directly intoan ammeter shunt that is a direct short. Run your alternator at6,000 rpm while increasing field voltage from a lab-type powersupply. The current you read on the output should be in exactproportion to the externally applied field voltage.
As long as the output current increases linearlywith field voltage, you are not saturating either the rotor orthe stator core. You may find that in some conservative alternators,you can increase the field voltage of a 14.5 volt field well over20 volts and still see a change. Other alternators will flattenout at 15-16 volts.
It is the ampere-turns of the field coil thatdetermines the magnetic force of the rotor. Remember that we mustkeep the air gaps constant when determining the field strength.If you can’t increase the output current by increasing field voltage,you need more metal in either the field, stator or both. The poweroutput increases by the square of the stator cross-sectional area,but it is proportional to the pole piece area.
The final stages in alternator stator-rotordesign are testing and tolerance determination. Testing does notinclude running your design on a test bench for a few secondsto see if you are getting your output current. It does, however,include operating your design constantly under full-speed andoutput range, and measuring the cores and wires for temperaturerise noting any hot spots.
Temperature rise can be measured at room temperature,however, the rise is added to the maximum ambient temperature.Most applications require reliable operation at ambient temperaturesranging from -40° to 130° C. This temperature must belower than the breakdown temperatures of your components.
The weak components are the insulation varnishon the magnet wire, the junction temperature of the main diodes,and the melting temperature of the bearing grease. You can obtainthese values from your parts vendor. You must also check the expansionof your materials from the coldest to hottest temperature to besure there is no binding.
Once your prototype unit is completed, yournext step is to establish production tolerances using a worsecase analysis. Designs that cannot be manufactured without minorvariations that always seem to occur are useless. Other itemssuch as shaft and bearing size, slip rings and brushes, rectifierdiodes and heat sinks are left to the discretion of the designer.Practically all alternators depend on some type of fan cooling.The thermal resistance of heat sink material can be decreasedby over a factor of 10 by blasting air on it. Voltage regulatordesign is too complex to even cover in a series of articles. Isuggest you contact a reputable manufacturer that may alreadyhave a design that could be adapted to your needs.
Whether you plan on designing a new alternator,modifying an existing one, or just plan on standard unit remanufacturing,the material presented in this article should give you a goodstart. The photographs included with this article show what othercompanies are doing to accommodate special needs. Read the captionsfor further details.