lets talk alternator design

[Frank S]

 Fab I know where a couple of Jakes are in Europe but they are little more than rusted hulks
When we return to the States we will probably have to make Southbend IN our first stop because of the wife's family and to transfer her bank accounts . It's only about a 4 or 5 hour drive to Chris's place I believe, Plus I have about 2 dozen folks I need to stop and see who are into antique trucks scattered between Sb and WI. Before we can turn south I'm supposed to have a 61 Kenworth sitting in Fargo ND  but not sure of its condition LOL I have a 49 Chevy pickup a 85 freightliner cabover and a 70 Ford F700 boom truck in North TX The only thing I know for sure that runs is my chevy pickup it was my daily driver before we left, the guy taking care of it keeps it in his garage and drives it once in a while My youngest daughter has 27 of my rifles & shotguns in her gun vault that I built for her husband
 I just found out the other day that one of my old customers has a retired Ford 750 backhoe with attachments including a 3000lb Hydra hammer, that he is willing to either loan me on a permanent loan or sell to me I own equipment and JUNK scattered in 12 States and 4 countries But I see no reason to make a special visit to Spain to buy the 2 old Jakes at this time LOL
 I told the wife that if we don't make out first landing in TX that she needs to be prepared to spend a couple months just to drive south
 Sorry for the Hijack

[bob g]

well it has taken me a while to dig back into stacks of books i have collected, in regard to the effects of peripheral velocity effects (rim speed) on EMF.

while digging i came across two things

one was two formula's used, one relating to to the use of diameter and  peripheral speed, along with the more commonly used formula  you guys favor.

i also came across this, which i thought was pretty interesting, considering its age (a bit over 100 years)  some attention to peripheral speed/diameter is discussed before any other parameters are set in a design.

http://share.iit.edu/bitstream/handle/10560/701/designofpolyphas00gilm.pdf?sequence=1

anyway here is the formula as best as i can state it on my keyboard.

from
standard handbook for electrical engineers
mcgraw hill
1949

section 7-7

instantaneous volts per inch of conductor = 2.0 * 10^9 lines per sq/inch flux density in the airgap * ft/min peripheral speed

or...

rms volts per inch of conductor = 1.42 * 10^9 lines per sq/in air gap density at peak of fundamental component of flux density * ft/min peripheral speed.

section 7-8
conventional emf formula's

L= length of armature core
V=peripheral velocity in feet per second
B=max gap density, in maxwell's per sq/in
e=rms volts per inch of active conductor
N=series connected active conductors per phase
kb and kp = differential factors (distribution and pitch factors)

e= (12/sqrt2) V*B*10^8 = 8.5*V*B*10^8

the text also lists the variety you guys like to use, which is factors flux, turns and frequency, and it is stated as being an averaged view
not taking into account other more or less relevant factors such as distribution, pitch, skew  and phase differential factors. (which might not have much issue with aircores, i don't know.

so you guys will have to forgive me for taking exception to the assertion that diameter and peripheral speed has no effect on voltage generated.

you have to understand that i started my education many years ago using texts much older than the one cited. i had no idea that folks started to accept an easier to conceptualize formula to explain their machines operation.

while some of you dismiss out of hand the idea that peripheral speed has an effect on machine output, some of us (at least me) are not so quick to do so.

anyway bottom line is this,  with the information related here, is it so hard to see where i am coming from? and why i question things?

fwiw
bob g

[Flux]

You have to determine a size of machine normally to do the intended job. Too small and the losses are too high or it will overheat. Make it too big and it uses too much material and becomes expensive.

These factors are determined by the type of machine, but in general there is a minimum volume of material to do the job. If it is an alternator of defined frequency ( as they always were in those days) you end up with long thin high speed machines and short large diameter low speed ones. To some extent you can trade length for diameter but there are limits on peripheral speed for mechanical reasons.

Once you have settled the basic physical size then you start the more basic electrical design. Methods have changed over the years, but having settled a frequency and a suitable diameter you can use peripheral speed and it did fit in with the basic emf equation that we were all taught.

This is not really the issue we were discussing before where we were looking a  given set of magnets and playing with the diameter.

Using this argument you would make a bigger machine by making it larger and the field and windings would be scaled up to keep similar proportions, in which case larger diameter does increase the output.

If you try playing with the diameter to change the peripheral speed you will have to change other things such as flux per pole to get it into the new diameter.

Another way to look at this is that having chosen your design frequency you have fixed your rotational speed  for a given number of poles so the output is determined by diameter and length. This won't work just by using the original set pf poles for one diameter and fitting them at a larger diameter, you have to make them bigger ( back to more or bigger magnets).

The thesis you linked to is interesting, at present I could only read about half of it but I will try loading it again sometime. You are quite right that by about 1900 virtually all the machine design was in place.

If you are a fan of 2 axis theory that came later but I can't follow it.

Keep up the good work.

Flux

[bob g]

Flux

thanks for the comments, and particularly the clarification

i guess originally i missed the point on having a fixed set of magnetic's and coils wherin an increase in diameter isn't going to effect much if any change in voltage, unless other changes are made.

my only point was peripheral velocity can have an effect, i guess i just assumed changes to the magnetic structure and pole design, both of which would be necessary to take advantage of any change in peripheral velocity... once we do all that we have a different machine in every way but in theory.

in other words we can't take a machine that works with a set of magnets and coils and simply increase the diameter and expect an improvement,  at least not much improvement, and maybe a reduction at some point of increase, using the same magnet and pole coils.

makes better sense to me now, we were talking apples to oranges, both of course being fruits, but very different in reality.

interesting

bob g

[ChrisOlson]

i guess originally i missed the point on having a fixed set of magnetic's and coils wherin an increase in diameter isn't going to effect much if any change in voltage, unless other changes are made.

Bob, the reason this is so, and it doesn't seem to make sense to you, is because you have to look at it from a different viewpoint of the area under the sine wave to the zero baseline.

The other thing is that the rectifier diodes only conduct during the time that the voltage is higher than the DC load voltage.  So a short peaky high sine wave doesn't conduct thru the diodes for very long because it's shaped more like a steep mountain than a gentle hill.

Over time experience has shown me that compacting the thing down as small as possible yields best performance, and making it bigger only adds resistance to the winding, or makes it necessary to use bigger wire to get the same resistance you can get in a smaller diameter.

Voltage is not the issue - voltage is easy.  You just spin it faster to get whatever voltage you want.  Resistance is the biggie.
--
Chris

[bob g]

Chris

i realize resistance is the biggie, and it is much bigger with low voltage
load (12volts) wherein any voltage drop is a much larger proportion of the generated voltage than it would be had the load been much higher.

from what i see now is a failure on my part perhaps in communication, in that i would not presume to build a bigger machine without either having more of the same size poles of a smaller machine or larger pole appropriate to the larger diameter of the machine.

my only point from the beginning was we can't out of hand dismiss diameter as having no functional relationship to the voltage generated, unless we don't change the size of the poles, number of the poles or both.

as for not understanding the area under the sinewave and all that, i do appreciate this, in that i alluded to the dead zone between poles should we only increase diameter but not increase pole count or size.

what i am still not convinced of is this

what if we increase the pole count of the stator, such as the following
(the following two examples are for illustrative and discussion purposes only)

alternator A is a 6 pole rotor machine of 10inches, and has proven to be successful for windgeneration, it is a 3 phase machine having 6 stator poles, it produces 50volts at 500rpm

alternator B has a diameter of 20 inches, having 6 poles on the rotor and 12 stator poles of the same coil count as in example A, it is configured as a 6phase machine.

what voltage will it produce at 500rpm?  it will be higher than simply the result of twice the stator poles? because of the higher peripheral velocity?  the area under the sine wave is the same in both A and B examples, or rather there is no dead space between sine waves as there would be had we used the same size and number stator poles in both examples A&B?

this would be much easier to communicate with a blackboard and all of us in the same room.

so i guess we will work it out as best we can?

bob g

[bob g]

lets move forward a bit, and take a step back as well

this discussion has caused me to dig deep in my library collection and i finally found what  appears to be an engineering text for generator and alternator design.

from what i can gather from the text and the formula used it would appear the common basic formula used by some here is only useful to explain the operation of well designed DC generators and not as useful as a tool to design AC alternators.

why there is a difference is not all that clear to me as of yet, but the word "average" is italicized and i have many times read right over it before i caught it and found an explanation.

we can use factors such as rpm, lines of flux cut, number of conductors cut to use a rather simple equation to determine voltage of a coil or coil group of a well engineered machine as an "average" voltage number. and we can only really use this for a generator and not an alternator, where another factor enters the equation at a minimum to arrive at another "average" voltage number.  that factor is a constant of either 4.44 or 2.22

bottom line the formulas assume one thing for sure, that is the machine that is being explained is assumed to have its stator filled with poles without large gaps. it is assumed that for a given diameter there will be a proper number of poles to fill the periphery.

it would also appear that increase in diameter (if the rpm remains constant) has a "square" relationship. in that if we double the diameter, we indeed double the peripheral velocity, and will necessarily have to double the pole count. doubling the velocity roughly doubles the voltage, and doubling the pole count also doubles the voltage, so doubling the diameter will square the power output of the machine.
this assumes the use of same flux density, same coils (just twice both) and the same rpm.

of course there are other factors to be considered but it is fair to state that increasing diameter allows for a much higher output at the same rpm or a significantly higher output at a lower rpm than the smaller machine.

the trick is, how to build a bigger machine without gaining an unacceptable amount of weight, and where to come up with the components to build the damn thing while keeping with a similar cost per unit of output or reducing that cost.

theoretically i think it might be arguable that the larger machine out to produce more power per unit cost than a smaller machine, and it might also be the gain in weight is not necessarily linear either.

anyway, i have an excellent book on AC alternator design, that has all the formula needed to engineer/design and build a successful unit of any practical size.  i am going to see if i can scan the pertinent pages and post them if there is interest, as i don't relish the thought of having to type them.

while i am not sure how applicable this all is to the axial air core machine, it might be useful for a reference for other considering other designs. axial air cores have issues and factors that i am not sure how best to apply to formula. things like skew, pitch, etc are often much different than is common of radial or iron core machines.

i am still investigating the revisit use of the wound field as i think it might have an advantage in a higher voltage machine.

my thought is we have no guarantee of continued availability of neo mags at prices that are acceptable, and having control of the machines output so that it can be better matched to both the prime mover and the load seems like a plus to me.

continuing the process

bob g

[electrondady1]

if my bench was clear, (some day ,some day !) would build a 24" dual rotor radial .
using a composite material for the body and just a ribbon of steel for the backing.


 

[Flux]

The easiest way to reduce weight for a given output is to increase rotational speed. Low speed machines are always big and heavy. The only things you can do to control the weight is to keep the non active parts as light as possible. The magnetic circuit including the field and windings you are stuck with but the other parts can be reduced in weight as long as they meet the mechanical demands. In the old days everything was cast iron whereas now we use die cast aluminium alloy for the body parts.  Composites may have some application as well .

There really is little difference between an axial or radial design but in larger sizes it is probably possible to save weight on a radial as the forces are in directions more easily dealt with.

There is a difference between a slotted iron cored design and an air gap one. With slotted cores the flux leaps from tooth to tooth and cuts a bunch of wires in the slot as though they were a single turn. The distribution is needed to sum these emfs into something approaching a sine wave ( in applications where this matters).

The air gap design is inherently distributed as not all the turns of a coil can occupy the same place and be cut at the same time by the flux.  The down side to this method is that instead of linking turns in a slot as in the iron cored machine, the air gap machine actually drags the flux across the conductors and unless they are very thin you get eddy currents within the conductor and lrger conductors have to be made like Litz wire with many insulated strands in parallel ( equivalent to laminating a core to stop eddy currents).

There is no basic problem with wound field machines and for machines of conventional design running at near full load there is no problem, the excitation is only a small % of the total output and for that matter the iron loss is also not that important and it needs to be proportioned to the copper loss for best results.

With wind we ideally need zero loss in low winds where output is tiny and we can tolerate more loss in high winds where there is plenty of energy.

the only machine that can reasonably approximate to this condition is an air gap permanent magnet alternator with no field loss and no iron loss. You can make a very good wound field machine but it won't be able to produce power in winds down to 5 or 6 mph. No problem if you can live with this.

Flux

The 4.44 factor you refer to is the form factor needed to convert from the peak voltage to the rms. In a dc machine the terminal volts is basically peak, but the difference between peak and mean is very small as the ripple is very tiny with many segments on the commutator.  When loades in a conventional way this really doesn't matter with an alternator but if rectified and fed to a battery it increases losses as the conduction period is reduced. I suspect this factor is inherent in a dc machine as well but we don't normally think about it.

[electrondady1]

i built an dual rotor axial flux with two phase stator last winter.

it had 32 coils and 16 poles.
 results were encouraging  but the tapered coils caused difficulties
so i want to try a radial for the simpler overlap this winter.







 

[ChrisOlson]

i built an dual rotor axial flux with two phase stator last winter.

it had 32 coils and 16 poles.
 results were encouraging  but the tapered coils caused difficulties

What difficulties did you have with the tapered coils?  I like the two phase because it runs as smooth as three-phase, can get better voltage out of it with weak magnets because of more coils in series (as opposed to a "flat" three phase axial), can wire it parallel for high amps, or wire it series for 1.414x single phase voltage and get higher voltage with less amps.

Overlapping coils in a two phase has been a problem for me in getting the stator thin enough.  But since ferrite magnets are cheap I've had good luck putting the two phases in separate stators, skewing them 90° electrically, and using a three rotor design with magnets on both sides of the center rotor.

Operating at high enough frequency, they pretty easily match a three-phase for DC "ripple" when rectified for battery charging.  And my experience has also shown that they run quieter with less harmonics in the winding than a three-phase
--
Chris

[electrondady1]

like your self , i like trying stuff "they" say won't work.
it was a tight fit to get everybody  in there without adding to the thickness.
there's a trick to it.

the little vertical mills i do don't spin fast enough  for three phase.
single phase allowed me to use all the mags in a series.
but watching the magnet pairs rotate through the  centre hole and realizing i was only getting what i wanted half the time
 

[opo]

Here is an idea, a little bit crazy, but still I think it may work to eliminate some weight. Here it goes: In the axial flux we have the magnets and coils mounted at a 90 deg angle with respect to the rotation axis. Basically the idea  consists in rotating magnets and coils 90 deg wit respect to the rotation axis (see attached image). To do so the magnets are mounted by pairs in individual C shaped steel/iron "modules" with north and south poles facing each other. This individual modules can be used to safely store the power full magnets, recall how the HD magnets are mounted with almost no magnetic leak.

Now suppose that we have 4,8,etc. such individual modules. We can mount them on a disc (made of non magnetic material maybe(?)) equally spaced. As in the figure

Now for the stator, we need also to rotate the coils 90 deg to make a cylinder shaped stator.

This design could save some weight IF for example the disc in which the modules are mounted is made of some light material (al, wood?)



Ok enough... my brain hurts.


Cheers,

Octavio

[gizmo23337]

Has anyone suggested C-core alternator design to reduce weight? I did a search on "lightweight flux alternator" and found some interesting thoughts.

http://www.commodityintelligence.com/images/2009/nov/30thNov/EWEC2008fullpaper.pdf

You can find more info if you follow the footnotes to patents and papers etc. I'd be interested in thoughts, if weight to reliability argument might hold with this type of design?

[opo]

Exelent find gizmo! It works then.

I was thinking on building a small 4 poles 3 coils using an a/c motor eclosure, its rotor and using the plastic fan to support the Cshaped mounts. I might go with a 3-ply disc instead of the fan. Will definately try it now... if I can fit 6 coils in this enclosure I might try the 8 pole 6 coil config.

Cheers!

[gizmo23337]

I didn't say it works! Actually I was curious if this has been tried on a smaller scale, and I am sceptical for the weight to reliability. I have many questions, and read the evolution thread at least three time (the whole thing). I have a lot more reading to do.

I'd like to politely ask where a guy would ask questions for mixed wind/solar, "net dollar zero" questions for a home owner? I'm a do-it yourself engineer, and have much to add at some point. My trade is digital circuits, dc/dc converters and low power design.

[stratford4528]

With reference to chain drive Over 30years ago I made my first wind turbine 4 blades 10ft diameter a 12 volt dynamo geared 9to1 with a chain. Amazingly I saw it produce 80 amps but I didn't know the wind speed. Never used the power but enjoyed making it. I have still got the dynamo. Unfortunately didn't get back into wind power till 5 years ago. I am just about to build my first alternator similar to the 10 ft on this site to replace the PMA I bought off Ebay which has just burnt out for the second time