Generator with ferrite magnets

[ChrisOlson]

then i came upon the GE digital residential meterheads on ebay, brand new for something like 15 bucks each.

Bob, you mean those meters the utility company uses to measure kWh for grid power?  I got an old meter that the utility company gave me for free, with the socket.  I got it installed on the output of my inverters to measure the power into the Main panel.  It's an old one with a wheel in it and the dials.  It says on it that it pulls like 3 watts at 240 volt AC, 60 Hz.  But it certainly doesn't do all the stuff you mentioned.  It's just a kWh meter.

because of just what you are going through Chris, i went with analog meters and precision shunts

I've never had real good luck buying any quality stuff off eBay.  I guess you can, but I never know what to think about stuff on there, so I don't trust it.  I called the electric motor shop today where I get my winding wire and they can order me three of the same Bristol ammeters like I have on my bus, except 0-150 amp scale, with the precision shunt, for $294 each.  That's a pretty stiff price and it doesn't seem I paid quite that much for the 500 amp one I got on my bus.

I think I might just get some cheaper ones so I can tell at a glance what the turbines are doing (I've never liked those Doc Wattson meters for that anyway) and use my Sun AVR when I test a generator or turbine.
--
Chris

[ChrisOlson]

If you mean "Cat5" cable (commonly used for ethernet) then it's just plain copper wire and will have no practical bearing on your readings whatsoever.

Yes, that's what I meant.  It says in the instructions to use a twisted pair wire.  I knew ethernet had twisted pairs in it.  So I tore the outer jacket off a ether cable and got the wire out of it.  That's why I called it ethernet wire.  If I would've gotten a twisted pair out of an old phone cable then I would've called it telephone wire.   
--
Chris

[bob g]

not the old type with the spinning wheel, but rather the new digital type
which has no wheel and no batch of little clock needles either.

bob g

[ChrisOlson]

Bob, I didn't realize that that type of meter would work on a turbine where you have widely varying voltage and frequency.  I thought they were designed only to measure stable AC power.  And usually split phase for residential power (in the US and Canada).
--
Chris

[vawtwindy]

Chris

Sorry to interrupt in between,

Have you ever seen the difference in "required torque" between Neo's and ferrites? which one needed more torque/

[keithturtle]

I just received a box of these ferrites

http://www.magnet4sale.com/Ceramic-Magnets-C8-Hard-Ferrite-Magnets.html

Shipping was three times the product price.

C 8;    I intend to use in a drum type alternator for waterturbine

Turtle, still at it

[scoraigwind]

Have you ever seen the difference in "required torque" between Neo's and ferrites? which one needed more torque/

Torque is an element in mechanical power.  Power is equal to speed times torque.  Just like electrical power is equal to voltage times current.

In an alternator, more speed gives more voltage (open circuit voltage anyway), and more current output will result in a higher torque input requirement.

If you use more powerful magnets then you will get a higher voltage for a given coil design, and speed.  But unless there is current there will not be any torque involved (except to cover other losses such as friction and stray magnetic losses).

In the sort of machines we are discussing, most of the loss at high power tends to be due to the heating of the copper by the current, so it's fairly easy to predict the losses, given a knowledge of current and resistance.  The input (mechanical or shaft) power needs to be equal to the losses plus the output.  That's because of the law of 'no free lunch'.

So you can calculate torque from this, and be aware that it has no direct relation to the type of magnet used.  But neo magnets do tend to allow relatively higher voltages from coils with lower resistance, so they can produce high currents in a short circuit which means excellent braking torque from small alternators.  Using neo magnets (or a speed-up drive on ferrites, which has almost the same effect) you can improve the efficiency of the alternator and get more electrical power out for less mechanical input power.  But you will still need enough input torque to produce the output power, and if the rpm doesn't vary much then this torque will stall your blades.

As a rule with a simple alternator of the axial type with no cores in the coils, the more efficient the alternator (lower resistance) the more constant the speed.  It doesn't need more much more 'open circuit' voltage to overcome the resistance of the windings so it can kick out more current at a speed not much higher than the cut-in required to reach battery voltage.  And this constant speed is not suitable for the blades because they need to run faster in higher winds.

Less efficient alternators need to produce higher 'open circuit' voltage for the same output current (so as to overcome the resistance) which means running faster.  (The output voltage isn't higher, because some of this 'open circuit' voltage gets used up in the internal resistance once the current flows.)  So they don't need more torque whereas the do need more speed.  This is a happy situation for the blades because they will be better at providing more speed as the wind increases.  But it doesn't alter the fact that the electrical efficiency has to drop in order for this speed increase to happen. 

So it's a trade-off.  For a direct coupled alternator that is connected to a battery, you can only work well over a narrow range of windspeeds.  When the wind is much stronger then you have to either have an inefficient alternator (high resistance) or inefficient blades (stalling).

[oztules]

Chris,
If you used your very considerable fabricating  talents to build a torus like in this thread..... http://fieldlines.com/board/index.php/topic,145994.msg996385/topicseen.html#msg996385

I think you would have a machine to die for. The design seems to tick all the boxes.... if I had cheap ferrites over here (instead of boxes of neo) and access to silicon steel band, I'm not sure I could help myself.  I'd have to do it.

Hugh, what chance of a cook book on that kind. I don't see any other design taking advantage of cogless, tight air gap, possible home fabrication (easier than stamping stators), and excellent use of the poor flux..... just seems the beez kneez for me...... I'm dribbling I think....... The brake drum style is good, but not for higher power as the air gap seems too big for ferrites.



...............oztules

[ChrisOlson]

Have you ever seen the difference in "required torque" between Neo's and ferrites? which one needed more torque/

They're the same for the same speed and power.
--
Chris

[Flux]

The problem for home build is the core. If you can find a source of decent core material at a realistic price then it is a good approach.

To be any use the core has to be high quality magnetic material. Because the flux enters in one direction and travels in another the only thing that will work properly is a tape wound torus similar to that of a toroidal mains transformer or current transformer.

There is nothing made that way that gets scrapped so you are likely to be buying new. Core materials in small quantities don't come cheap.

With the low flux of ferrite a wire wound core may be good enough but there a lot more flux breaks. Even then decent wire is not readily available, possibly soft iron wire used by florists may be ok. Mig wire and other stuff will be poor.

I did get a core made by a transformer manufacturer but the cost was way more than a set of neo magnets for the same size machine.

If you can overcome that problem there are a few tricky but not insurmountable issues to deal with. It is a very good design.

Flux

[ChrisOlson]

The problem for home build is the core. If you can find a source of decent core material at a realistic price then it is a good approach.

I briefly toyed with the idea of building a iron core with ferrites.  But the cost of putting magnets on the other side of the stator, vs a core of some sort, is less, with better performance.
--
Chris

[electrondady1]

turtle ,
that's a lot of mag for .60 cents.

[ChrisOlson]

vawtwindy-

Since I found our my meters I've been using are worth about as much as the few cents worth of plastic and silicon they use to make them, and all of the ferrite generators I've built to-date are on my turbines, I'm going to build another one for testing.  I'm going to set it up on a regular prony brake and make a chart of the rpm and torque required to drive it vs power output measured with my Sun AVR thru a 150 foot wire run to a battery bank.

That won't tell you the actual efficiency of the generator because there's losses in the rectifier and wire run.  But it will give an idea of what sort of performance these dual stator units are capable of on a real turbine in a real installation charging a real battery bank.
--
Chris

[bob golding]

The problem for home build is the core. If you can find a source of decent core material at a realistic price then it is a good approach.

To be any use the core has to be high quality magnetic material. Because the flux enters in one direction and travels in another the only thing that will work properly is a tape wound torus similar to that of a toroidal mains transformer or current transformer.

There is nothing made that way that gets scrapped so you are likely to be buying new. Core materials in small quantities don't come cheap.

With the low flux of ferrite a wire wound core may be good enough but there a lot more flux breaks. Even then decent wire is not readily available, possibly soft iron wire used by florists may be ok. Mig wire and other stuff will be poor.

I did get a core made by a transformer manufacturer but the cost was way more than a set of neo magnets for the same size machine.

hi flux, does the core have to be one continuos strip, or can it be done in sections? reason i ask is i sometimes see scrapped variacs with strip cores and thought if you cut one up into strips about say 2 inches wide you could join the strips together to  make a siutable core. a lot of work i know "just cut though 3 inches of high grade silicon steel" but it could be done. given the price of neos you could pay someone to do it and still save money

If you can overcome that problem there are a few tricky but not insurmountable issues to deal with. It is a very good design.

Flux

[ChrisOlson]

In an alternator, more speed gives more voltage (open circuit voltage anyway), and more current output will result in a higher torque input requirement.

Or higher speed input.  You get more current with more speed too, with constant torque input.

I know for a fact that the reason these ferrite generators are making such good power is because of the speed.  They have to run faster than the neo, which also requires the blades to run faster.  When the blades run faster they can capture more of the wind - up to a certain point where the blades are running too fast for the design of the airfoil and the pitch they're running at.

My geared generators require very low torque input.  There is very high torque input at the transmission input shaft.  But that high torque gets converted to low torque at the PTO shaft at much higher speed.  I like the high speed, low torque generator because it puts less strain on the stator casting and mount, and that's how I can get away with a fairly light duty mounting (five 3/8" diameter studs) on the stators.

When I build my test generator I'm going to drive it direct with no transmission using a gear type pump/motor instead of a roller stator motor to eliminate the losses in the transmission from the picture, and to get a better idea of what the generator itself is doing.

I wish I could come up with a way to test the generator on a purely resistive load like Bob mentioned, without going into a battery bank.  But I don't know how to set up a purely resistive load that would hold the voltage at the constant level you get with a battery bank for accurate results on what it would do in actual service charging batteries.  If anybody has a good idea on how to do that I'm open to suggestions.  But for now, I think the only real accurate way I have to use to my batteries for the load, and unfortunately, that requires a 150 long wire to get the power to them because I'm not set up in my utility/battery room to test run generators there.
--
Chris

[Flux]

Took me a long time to find the question buried in there Bob.

Variac cores are very high grade material and you could do that. Every cut introduces extra reluctance but it would still probably work perfectly well. Certainly be better than a wound iron wire core. Not a commercial proposition but for one off it could be worth the effort.

Flux

[bob golding]

sorry about that. i was expecting the quoted bit to come up highlighted. just found a core manufacturer here in the uk  that does strip cores. just waiting for the catalogue to download. i was thinking around 2 inches x 2 inches as a starting size? i dimly remember the formulas for core saturation in transformers. are they the same in this application?

[Flux]

"I wish I could come up with a way to test the generator on a purely resistive load like Bob mentioned, without going into a battery bank.  But I don't know how to set up a purely resistive load that would hold the voltage at the constant level you get with a battery bank for accurate results on what it would do in actual service charging batteries.  If anybody has a good idea on how to do that I'm open to suggestions.  But for now, I think the only real accurate way I have to use to my batteries for the load, and unfortunately, that requires a 150 long wire to get the power to them because I'm not set up in my utility/battery room to test run generators there."

I think we are at cross purposes here, a resistive load will have a voltage that changes with current. This is a normal load for most things. For a battery charger it is no use. The best you could ask for is a load simulator that maintained constant voltage irrespective of current. This could be a battery bank with a dump regulator or an electronic load bank without a battery.

The only disadvantage with a battery with no controller is that the volts will change a bit with current so you need to keep checking volts as well as amps.  Well discharged large battery banks stay fairly constant. Otherwise you need to add some sort of drain load as the current increases.

The alternator behaviour will be totally different into a constant voltage load compared with a resistive one.

Flux

[ChrisOlson]

The alternator behaviour will be totally different into a constant voltage load compared with a resistive one.

Well, I guess I'm confused as to why Bob suggested it should be tested into a resistive load to get accurate output figures.  I've tried that before with my Sun AVR and the carbon pile gets hot and it will only take it for a little bit.  Plus the knob has to be adjusted constantly to hold it at a particular voltage.  That's why I went to testing them with a regular battery bank, which seems to be a lot more stable.
--
Chris

[ChrisOlson]

It dawned on me what Bob was probably talking about - using an AC resistive load to measure generator efficiency and eliminate the rectifier loss component from the calculations.
--
Chris

[oztules]

I can't think why you would use a resistive load to test a battery charging alternator.... makes no sense to me.

Unless you wanted to see it drive a resistor instead of a battery??? but for what?





.................oztules

[scoraigwind]

In an alternator, more speed gives more voltage (open circuit voltage anyway), and more current output will result in a higher torque input requirement.

Or higher speed input.  You get more current with more speed too, with constant torque input.

If we are talking about a given battery load with given cables etc and everything constant, then you are not going to be able to increase the speed without increasing the torque.  Once you chuck in a bit more torque, the speed will rise, voltage will rise and so will current (due to the extra voltage).  When the current has risen enough to counter the new torque the speed will stabiliise again.

Real life is always complicated (especially true recently on this thread) but the theory behind all this is relatively simple. 
1. Higher speed will give a higher open circuit voltage.  It's the rate at which the flux cuts the wires that determines the voltage. 
2. And more current will produce a higher torque.  Its the reaction of the current to the flux in the airgap that determines that torque. 
I find the basic physics very useful.  But it's not always easy to express it clearly.

We are very much in agreement that the ferrite magnets producing lower torque is helpful in allowing the blades to run in their optimum power band.  I don't seem to be managing to explain that this only occurs because of increases in the electrical losses due to higher resistance in the system.  Of course there are other ways to allow the blades to speed up but without recourse to electronics, converters, inductors etc you are stuck with the dilemma of lower electrical efficiency as a trade for better blade performance.

[ChrisOlson]

I can't think why you would use a resistive load to test a battery charging alternator.... makes no sense to me.

I replaced the 5.0 Briggs on my hydraulic pump with a 9.0 Honda yesterday for more power (I'd like to try to burn one of these up on the bench to see what they'll take).  I set up the prototype generator that I showed in this thread on it to make sure everything works.  That prototype isn't the best thing because it has different turns in the stators.  I am going to build two new 48 turns 12 AWG stators for it before I test it.

I ran it a bit to make sure everything is working on my test setup.  I found out the generator is actually running around 42 volts with the battery bank at about 29-30.  I have an old silo unloader cord hooked up to the rectifier - about 100 feet of 8/4 wire.  And two #2 welding cables about 150 feet long from the rectifier to the battery bank.

I have to figure out why I'm getting that much voltage drop with that big wire.
--
Chris

[ChrisOlson]

If we are talking about a given battery load with given cables etc and everything constant, then you are not going to be able to increase the speed without increasing the torque.

Hugh - no, that is wrong.

At the same load, comparing a direct drive to a geared drive, for instance; the geared drive drastically reduces input torque and increases speed for the same output voltage and current.  It takes two different design stators (a much more efficient one for the geared drive) for the different approach to driving the generator.

I was just pointing out that you can build a very high speed generator that has a low input torque requirement that is much more efficient than its direct drive counterpart.

With a given battery, resistance, generator combination where everything remains constant except for speed, then the torque requirement will increase with speed.  And I intend to show folks that relationship when I get this test setup running and tuned so I can test it with known accurate meters.
--
Chris

[hvirtane]

Chris said:

"you can build a very high speed generator"

What about making the magnet rotors and the stator with really big diameters?
You would get the magnets moving with high speeds relative to the stator coils.

Has anybody tried that?

I think that by building both the stator and the magnet rotors on spoked wheels you could make them quite light.

-hv 

[ChrisOlson]

That doesn't work because it's the frequency that does the trick with a given number of generator poles, not the actual speed of the pole over the coil.

It's really hard to explain, and Hugh and I are sort of talking about two different things when it comes to speed vs torque.  I'm talking about designing generators that specifically require lots of speed and low torque input to make current.  Hugh is talking about a constant, no change to components setup, where you increase current by increasing speed and torque input.

I try to come up with designs that require the speed increase out of the blades to proportionately increase current output, to get a better "match".  Again, I hope to show folks that relationship here soon when I can get my test setup working the way I want.  I already got a problem - the high speed gear motor I was going to use to drive my test gen doesn't have enough power, even with the 9.0 Honda on it.  So I might have to use the geared drive anyway to test the thing.
--
Chris

[hvirtane]

You wrote:

" it's the frequency that does the trick with a given number of generator poles, not the actual speed of the pole over the coil."

But nothing prevents us of putting lots of small magnets on big disks.

-hv

[Flux]

It is possible to build direct drive alternators by using larger diameters, the thing to remember is that increasing diameter alone doesn't help. You have to use the increased diameter to add more magnets or use bigger ones to fit the extra space. The voltage depends on the number of turns per phase, flux per pole and frequency. Increasing the number of poles raises the frequency.( for a given rotational speed)

If you use bigger magnets then frequency doesn't change but flux per pole increases.

If you use more small magnets you gain from the increased frequency.

Using a speed increase makes up for the lower flux of ferrite without increasing the physical size, but without speed increase the size must go up considerably for the same output and efficiency.

I can't follow the torque argument, you can use a low torque high speed alternator and increase torque to the blades with a transmission or you can get the high torque at prop speed by a bigger and slower alternator.

Blade matching depends on cut in speed and electrical efficiency. This applies as long as both alternators have the same shape load curve, which for air gap axials is basically the case.

Single phase will load differently from 3 phase so there is an issue there. Two phase has more ripple than single phase and it will be a bit between the two, but much nearer 3 phase than single phase. I wouldn't expect the rectified 2 phase ripple to be a big issue compared with what happens single phase.

Flux

[defed]

I have to figure out why I'm getting that much voltage drop with that big wire.

what is the amperage at that voltage output?

at 150', 2AWG, 24v, 30amps, voltage drop is 7%.

if i use 36v to get closer to what you are actually making, as opposed to the system voltage, and go to 50 amps, the voltage drop is 8%.

this doesn't even take into account the losses in the 8/4.

[ChrisOlson]

I wouldn't expect the rectified 2 phase ripple to be a big issue compared with what happens single phase.

You're right there, Flux.  When I tested the single phase on the turbine it basically had very little braking torque.  But it still had surprisingly good output because the turbine ran very fast with it.  I ran it at much higher frequency (about 70 Hz) with a higher gear ratio.  When I stacked two of them for two-phase, the loading was considerably higher and I needed to go back to the lower gear ratio.  But I learned from it that a single phase generator can work just as good on a wind turbine as a three phase if you run it at high enough frequency, as far as matching the generator's output to the power curve of the rotor.

It's my opinion that the two-phase setup is better than three phase for getting a good "match".  Three phase generators of reasonable efficiency are too "stiff" with fixed air gap, permanent magnet design.  The two-phase is less "stiff" and it lets the turbine run faster.  And it's smoother running than the single phase, although the single phase wasn't all that bad at higher frequencies.

So again, I hope folks don't get too hung up on this efficiency thing and think three-phase is the only way to go because single phase is "bad", two-phase is "better" and three-phase is "best".  I'm looking at ways, that are not in the book, on how to let the turbine speed up more according to increases in wind speed.  Using this odd configuration that delivers power in "pulses" instead of the "superior" smooth output of three-phase does that.

I realize this will be never accepted as "good" because it's not three phase and all conventional wisdom says it just has to be three phase to be any "good".  But I have learned that getting the rip speed ratio of the blades from 5 up to 7 at 10 m/s yields more power even in the lighter wind speeds and the generator does not get hot.  I'm hoping to get my test setup working right so I can folks how it does it, even though it's supposedly "inferior" to three-phase

This is my hydraulic power unit with the bigger 9.0 Honda on it:



This is the Ross TF-series motor I've always used to drive my generators, but it's max rpm is 400:



I was looking for a way to drive the generator direct-drive to eliminate the geartrain in the transmission for testing.  But that Ross motor is the only one I have that has enough power.  So that complicates the testing because the actual torque input to the generator has to be calculated instead of measured.  With the big Ross motor I can measure torque input to the trans input shaft but it can't run fast enough to do it direct to the generator.

Hopefully I can come up with something where all the parameters can be measured with no theory or calculations involved.
--
Chris

[ChrisOlson]

I can't follow the torque argument

Flux, this is hard to explain, but I'll try.

Let's say we have a generator that operates at 5 rpm/volt.  So this would be like a 24 volt unit that cuts in around 120 rpm.  So at 200 rpm and 6 TSR the wind should be blowing 7 m/s.  The generator has increased its open volts since cut-in by 80/5 = 16 more open volts.  That 16 open volts translates to an increase of 768 watts @ 24 volt nominal, or with the generator at 75% efficiency an increase of 1.024 kW and 90 lb-ft of torque required to drive it.  This is more power than the rotor can make at 30% efficiency so the TSR is low because the blades are running up against their torque curve and can't reach the required 200 rpm.

Let's look at a similar geared unit that operates at 12 rpm/volt.  It still cuts in at about the same speed.  At the same efficiency the open volts still has to increase by 16.  It takes 192 more rpm from the generator to get that, and the same 1.024 kW.  This requires an increase in the input torque of 37 lb-ft, thru the gearing at a .4375 ratio, plus the loss in the gearing, which is an increase of 86 lb-ft at the input shaft to drive the generator.  The lower input torque requirement ( by 4 lb-ft)  requires more speed to make the same power, which the blades are fully capable of.  So as the wind speed continues to increase the blades stay closer to their ideal TSR and takes more speed to make each additional open volt required to push more current.

In mechanical engineering, high speed, low torque shafts are always the most efficient way to transfer power from one place to another, just like high voltage, low amps is the most efficient way to transfer power in electrical engineering.  In mechanical power transfer the high speed low torque situation requires lighter, smaller components that don't have as much rotating mass.  In electrical transfer the high voltage, low current requires smaller conductors and less copper mass.  The principal is somewhat the same.
--
Chris

[scoraigwind]

Catching up here

If we are talking about a given battery load with given cables etc and everything constant, then you are not going to be able to increase the speed without increasing the torque.

Hugh - no, that is wrong.

At the same load, comparing a direct drive to a geared drive, for instance; the geared drive drastically reduces input torque and increases speed for the same output voltage and current.  It takes two different design stators (a much more efficient one for the geared drive) for the different approach to driving the generator.

I was just pointing out that you can build a very high speed generator that has a low input torque requirement that is much more efficient than its direct drive counterpart.

With a given battery, resistance, generator combination where everything remains constant except for speed, then the torque requirement will increase with speed.  And I intend to show folks that relationship when I get this test setup running and tuned so I can test it with known accurate meters.
--
Chris

There is a lot of scope for misunderstanding here.  sorry I thought we were talking about a given generator and drive.  We all know that gearing increases the efficiency dramatically (as do more, larger, stronger magnets and larger coils...) .  However if you want to work over a wide range of speeds (to improve blade matching) then you need more resistance and you therefore have to avoid being too efficient.

On the more recent question about using a larger diameter rotor, Hirtvane, you might be interested in the analysis I did of what it would take to beat the dual stator design with one large stator and I found that it was possible with an increase of only 24% in the rotor diameter and a single stator.  The details are here http://scoraigwind.co.uk/?page_id=953  It's based on my own standard recipe but the lesson applies to any type of generator in principle.

If you were to put all the magnets on 2 rotors and all the coils in one stator it's equivalent to doubling the rpm but with no gearing, so yes I prefer this solution but Chris does not.  We have been around this loop a few times already.

[ChrisOlson]

There is a lot of scope for misunderstanding here.  sorry I thought we were talking about a given generator and drive.  We all know that gearing increases the efficiency dramatically (as do more, larger, stronger magnets and larger coils...) .  However if you want to work over a wide range of speeds (to improve blade matching) then you need more resistance and you therefore have to avoid being too efficient.

You're right - there is much scope for misunderstanding as I think most folks think the more efficient the better.  And this can get confusing going from battery charging to a grid-tie application too.  So yes, in a way that is what I'm trying to say - you need to be efficient enough at the full output of your prime mover to prevent overheating, basically.  Designing for that in a wind turbine can be tricky if you don't know how much power your blades actually make.

@ defed - I don't know how many amps the generator was putting out when I had the volt meter on it.  I suspect somewhere from 80-100.  I just ran the Honda up to full throttle to see if it would work.  Assuming it was 80 amps I found a couple nice online voltage drop calculators and it says I should get about 4 volts drop on the 2 AWG and 6 volts drop on each of the 40 amp single phase 8 AWG circuits to the rectifier.  That's a total of about 10 volts plus about a volt and a half on the rectifier, which is about what I measured.  So the wire run is fine - just seemed high when I threw the volt meter on it.
--
Chris