Windyboy and growl

[tieole]

Hello, I realise most of ya'll here are doing off grid machines with fat copper wire and not to many turns.  For the last couple of years we've been building grid tied machines here in France.  The first one we build was with Hugh for a trainning course in August 2007.  These are 3,6m-4m (13 foot) diameter machines with 300 turns of 0,90mm (19 gauge) wire.  32 magnets, 12 coils in three phase star.

We use a SMA Windyboy to let the voltage run up, pulling more current as the speed of the machine (and the voltage) increases.  

Well here is the question, the machine produces a transformer type howl when the Windy Boy is pulling energy.  Any ideas why and how perhaps to reduce this howl?  Seems on some machines is less, but not sure why.  One thought is the wires are vibrating, seems there is less noise when the coils are tightly wound.

Ideas?

Some of our machines can be seen at tripalium.org

Cheers,

Jay

[electronbaby]

I would say its because the windings are vibrating inside your casting.

I notice the same thing on machines that are not wound tightly, or potted thoroughly when cast.

I would make sure everything is nice and tight while winding. I happen to use a good amount of super glue while winding, and have found that the coils remain much more rigid, and I have not had a problem.

Not sure if this is necessary, but maybe wind tighter coils??

Sometimes the vibration from the stator can be amplified by the tower structure also. Maybe the tower is making it worse?...hard to say.

I have checked out your site before. Very nice. Keep up the good work :-)

[Flux]

All alternators growl or whine to some extent. Ideally the turns of the coil should be completely impregnated. If you pot a dry coil in polyester resin it will not completely impregnate the winding. I have little time for superglue but it is probably adequate. Impregnating with the same resin as you use for the potting as you wind the turns may be a better option if you can devise ways of dealing with wet coils or removing cured coils from the winding jig. I use release film on the jig and let the resin cure before removing the coil. ( slows the winding process somewhat).

Any lack of rigidity of the stator will lead to growling, the stator is a fairly effective loudspeaker diaphragm. The coils themselves if completely impregnated increase the stator rigidity very considerably. Adding marble flour or similar into the potting mix and perhaps a bit of glass strand may make the stator more rigid.

How you attach the stator may have a significant effect on the modes of vibration and the resonances.Very rigid stator fixing brackets may help, the common radial arms are a bit of a tuning fork to vibration. It may help to stagger the resonances by having different types of stator mounting rather than having the same and resonant at the same frequency.

If the resonances in the alternator excite mechanical resonance in the rest of the structure then things get amplified.

If you can bench test an alternator you can probably fairly easily find where the main noise is coming from. If it is quiet enough on bench test then suspect structural resonances excited by the coil frequency or harmonics.

None of my machines have produced enough noise to present a problem but they all do make noise at some points depending on the load. I don't see any fundamental reason why a high voltage winding should be worse and I have found that my last mppt machine is actually less noisy than those directly connected to batteries.

One factor may be that the line resistance on a high voltage set up is relatively negligible and the instantaneous peak charging current into the capacitors of the inverter may be higher than the peak into a battery at the end of a long resistive cable. Try a choke in the dc lead after the rectifier it may work wonders, I suspect something like 100mH will be enough to average the peaks.

Flux

[TheCasualTraveler]

     My first machine had a growling noise, I did a post on it here,

http://www.fieldlines.com/story/2008/4/22/231227/860

That stator was thin, had glass cloth and little talc.

     On my new machine I poured my stator with fiberglass mat instead of cloth. Also it is thicker and had lots of talc to the point it came out looking like gray marble instead of clear amber. Also it has 4 mounts instead of 3. Plus, I eliminated the ears where it attached to the mounts and made the whole thing larger. It is solid as a rock. I have no noise on this machine so far except for when I shorted the phases on a bench test and that made the whole thing vibrate.

[Airstream]

Your WindyBoy specifications allow DC input with up to 10% ripple voltage, have you checked the wind turbine output waveform for the peak-to-peak  voltages?  I know the inverter has a large smoothing capacitor but some of the growl may be from your funky coil choice (not-bad funky in this case?) and mag layout output waveform just overpowering the inverter input filtering. Neo magnets are brute force devices and your 300-winding high-voltage coils might be better modified for use in a 5-phase array to reduce ripple.

Also - Dual rotor axial air core alternators and conversions can get noisy when light duty diodes are used, well that is better stated as non-industrial "process designed" diodes are wired in, or, when the cheapest hobbyist or common consumer diodes get used...

To see what I mean about Diode choice look at a light duty rectifier diode's depletion zone as a 'perfect' on-off switch, when the it saturates and reaches "cut-in voltage" or "on-voltage""  a mathematically perfect  square wave strobe-pulses the circuit into activity.

The instant pulse cut-in and drop-out sets any moving electro-mechanical parts that were open-circuit prior to diode conduction into a torque loaded flux/coil work state until their induced voltage drops below the diodes on-voltage when torque is released.  

It's the cut-in and drop-out that amplifies into a resonance that we hear as a growl, pinging, ringing, squeal. Wild AC into a resistor bank will still growl or ring but its then mostly heard coming out of the resistors as buzzing if they have open coils wound around a ceramic frame or loose mountings.

So how do some diodes differ? The silicon junction can be manipulated by geometric scaling, doping level, electrode type to alter the progression into on-state.  When industrial high power diodes are designed most times they try to remove the `instant-on' behavior by forcing the solid-state junction conduction into a soft-on process, guiding the junction state change in a wave across the junction region instead of the entire region simultaneously like a camera flash strobe.

. ...Hmnn - 300+ words used to say "get a truly gnarly used High-Power diode off eBay and try it." Its worked before, one installation in California sounded close to a piezo buzzer alarm before they beefed up the diode bridge!

[boB]

When the turbine growls, turn off your 50/60 Hz inverter for a moment and see if the growl goes away.

I was just dealing with this yesterday.  I think you will find that the 100Hz or 120Hz ripple "current" is what makes it growl.   With or without diodes, that current is reflected back into the turbine where either the windings will vibrate like a Woofer-Hound (nice pun, eh?), or due to transformer-action, the ferrous materials will just use Faraday's law for an excuse to make noise.

Also, maybe some rubber-like vibration insulation between the turbine assembly and the tower will also help a lot I would think.

That's my take on this weirdness.

Then again, just eat Frosted Flakes and pretend you're a lion.

boB

[ghurd]

"What if"?  a large cap was installed near the rectifiers on the DC side?

Idea being the DC wire resistance would be the R in RC.

I have had some pretty strange things happen by adding a cap where it shouldn't make any obvious difference.

Just thinking out loud.

G-

[electronbaby]

You also could add inductance on the AC side of the rectifiers. Some manufacturers have done this to quiet out tower hum from badly matched alternator / feed line and highly capacitive inverter equipment.

The right value can be calculated, but it might take some experimentation. I have played with this a little.

[ghurd]

Speaking for those of us too old to recall RF theory... Uh, What?

Seems like the inductance on the AC side of a factory-produced (AKA iron-core?) machine would be pretty high already.

My thought is to catch the current in the cap.  And all the ripples.

DC amp feed from the tower cap to the inverter cap would be smoothed out.

Maybe a cap and 20mH?

I am not saying any of that is correct.

The Hz looks kind of slow and Jay's machine sounds like an air core.

Interesting stuff.

G-

[scoraigwind]

Grid tied machines do seem more prone to tonal noise, and I have heard it said that the rectifier is the culprit.  I am told that Eoltec are working on an active bridge and maybe I heard that the Jimp needed such treatment too while I was in France with you.  Downside is the high price.

It would be nice to think there is a lower tech way to damp the sound and so why not try out Flux's recipe with the inductance between diodes and capacitors.  You just have to ask him what a henry actually looks like (or several henries).

[tieole]

Thanks for the great comments.  I forgot to mention these are airgap machines, so not the lamination that's vibrating.  Also am using the Windy Boy Protection Box, and yet the howl seems to be noisier with this system than when I was using an oversize bridge diode.

Just a moment to gripe about the SMA wind protection box.  They could have built in a stop switch and lightning protection.  On the good side, I haven't killed a Windy Boy since.

The choke sounds good, reducing the AC ripple on the DC side going into the windings.

I've started my research and will let ya'll know.  In the meantime the newest tail vane is in the form of a wolf's head, that can say that wolf really howls.

Jay

[boB]

Just stuff a bunch of old towels inside the tower.

boB

[boB]

Oh, and/or old socks and underwear and anything you can find that will dampen the

tower from being a baffle and resonator.

boB

[Ungrounded Lightning Rod]

NASA used an energy-absorbing rubber with a lead strip backing it as an anti-vibration sticky tape.  Stick it to the middle of the vibrating area and it ate the vibration.

You could do the same sort of thing:  Find a spot on the tower where putting your hand on it significantly quiets the vibration.  Mount a vibration-insulating "foot" from a machine mount (i.e. typewriter) with a weight on the other side of the foot from the tower attachment.  Repeat, adding dampers to additional "loop"s of the standing waves until it's quiet enough for you.

[Flux]

Traditionally alternators haven't been used to drive rectifier loads into capacitors. The traditional rectifier load has been resistive or inductive ( machine fields).

Once you go to capacitive loads the machine is expected to deliver its power in short rectangular pulses during the time when the relevant phase voltage is above the mean dc. The peak phase currents are high for a given rms value. This impacts badly on efficiency and stator stress ( including noise).

I can't see any difference between battery charging and grid tie except for the grid tie being much higher voltage and hence the line resistance will have negligible effect.

By using some form of power factor rectifier you can maintain a reasonably sinusoidal ac current into the rectifier. I really can't believe that changing one conventional rectifier for a bigger one will make any difference but if the commercial rectifiers are something like the Vienna rectifier that incorporates power factor correction then of course it will have a very marked effect.

Another factor may be the way the  protection box works. If it limits volts by clipping on the ac side of the rectifier then it may be partially clipping even in normal operation. Short of looking with a scope at the machine 3 phase input I don't know how you can prove this but if it is clipping peaks then noise will be very likely.

Adding some inductance between the rectifier output and the inverter input capacitors will increase the conduction angle of the rectifier and improve the mean to rms current ratio and may dramatically reduce the noise but this may not be effective if the protection box is clipping peaks in normal operation. You may have to prove this on a day when you could safely run without the protection box if the choke gives no improvement.

The case is worse for an air gap alternator, all iron cored machines have significant leakage reactance and this causes overlap on the rectifier with a lengthening of the conduction period.

you could possibly try adding small ac reactors in the ac lines to simulate leakage reactance and force the rectifier to overlap but the inductance values will have to be small or you will get the undesirable reactance limiting characteristic of the iron cored machines.

If you go to a power factor rectifier such as the vienna or other forced commutated versions you can overcome some of this reactance limiting effect but the rectifier should then cure the noise anyway so ac reactors would be pointless.

These elaborate rectifiers are very lossy with too many diode drops at low voltage but should be excellent for the higher voltage grid tie, but they have fairly elaborate control circuits and you may need to involve SMA or someone who really understands these things unless you are a bit of an electronic wizzard ( it's basically inverter technology and ought to be part of the inverter but is probably too costly for small set ups).

Flux