Home
JAQForum Ver 20.06
Log In or Join  
Active Topics
Local Time 14:56 29 Mar 2024 Privacy Policy
Jump to

Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me.

Forum Index : Windmills : visual effect of capacitors

     Page 3 of 25    
Author Message
SparWeb

Senior Member

Joined: 17/04/2008
Location: Canada
Posts: 196
Posted: 07:43pm 09 Sep 2008
Copy link to clipboard 
Print this post

Hi again Gordon,
Nothing's stopping a user of WinSpice from putting an inductor element in series with the EMF of the generator. It's just a model. I don't have an oscilloscope, so this is the only means I currently have to "see inside" the generator.

You've got Dennis writing in all-caps in your thread about "caps". Can you two agree to disagree?

I'm going to look up "RFsim99", now.

Steven T. Fahey
 
GWatPE

Senior Member

Joined: 01/09/2006
Location: Australia
Posts: 2127
Posted: 09:26am 22 Sep 2008
Copy link to clipboard 
Print this post

Hi All,

PS edit: 3/12/08


An important consideration that should be noted, is that the capacitor coupling arrangements are charge pumps, and as such, these should not be left unloaded. An unloaded system, even if the mill is only producing a few ACvolts, has the potential to still produce hundreds of DC volts on the output over time.

Due care should still be taken and treat all connections as lethal.


this will have to be the start of a series of posts.

I have replicated Bryans test results. The capacitors are in series with the output, and not across the outputs. The mill is capacitor coupled to the load. I have recorded 19.5A @ 28.4V today, on many occasions. 550W, not bad for a standard 100S twisted poles in delta. I had measured a 5 degree temperature after an hour of operation, in the capacitors.

I will present some digital CRO outputs.

The first is the waveform between 2 phases on the mill side of the 230uF capacitors.



The mill did not seem to be spinning any faster and was noticably quieter.


This was the output waveform after caps prior to rectifier, just at cutin speed.



The waveform is just starting to be held by the battery load.



This is the output after the caps, prior to rectifier, at about 15A output current to the battery.

The series configuration seems to perform a VI conversion. I will not pretend to know exactly what is going on yet.

I now have a mill that easily starts in light winds, twisted poles, produces very useful power from a single stator and ferrite magnets.

I welcome any comments. The main criteria I would suggest is that a windmill should be set up with furling to be able to withstand all wind conditions with no load connected. Today the wind conditions were 20-25kph only during the testing. The mill is 2.4m rotor and is approx 50% furled at this windpeed.

Gordon.

Edited by GWatPE 2008-12-04
become more energy aware
 
oztules

Guru

Joined: 26/07/2007
Location: Australia
Posts: 1686
Posted: 11:47am 22 Sep 2008
Copy link to clipboard 
Print this post

Well done Gordon. It looks like Bryan hit on the right combination of caps and config for his mill (standard 100).

From the previous page:
"If we were to then drive the same alternator into a leading power factor arrangement (caps), we should see a rising rather than a lower to zero rate of climb, as the power factor straightened up."
..... this appears to be what is happening to a large extent. I think the sine wave gives the diodes a better waveform to work with as well which improves their efficiency. (I can't recall where I get this inkling from, but dimly in the past it was a truism to me for some reason) Maybe you can check the drops over the diode to check this in VA terms.

I should think Dennis may be well pleased that standard stators can be made to benefit from his idea as well, not just modified ones.

Further research could be done to determine the best values for each stator in various modes of modification I suspect, and it could then be put somewhere as a reference for future builders.... ie is that capacitance combo the best, or is there better to come.

I don't run an F&P, but if I did I would try and help answer these questions.

I am wondering if there is much more wattless current running around in the windings that could be given a voltage component in phase, and so become useful output, or have you harvested it all, and sorted out the vectors as good as it gets.

You may yet find that some smaller values in parallel may help still.

Well done Gordon, Bryan and Dennis



..........oztules

Village idiot...or... just another hack out of his depth
 
GWatPE

Senior Member

Joined: 01/09/2006
Location: Australia
Posts: 2127
Posted: 12:09pm 22 Sep 2008
Copy link to clipboard 
Print this post

The cutin speed with a series capacitor arrangement is similar to a standard 100S in delta, so this arrangement may benefit from a boost cct. This will not be a simple task, as the normal parallel arrangement used for a boost cct will not work. The parallel capacitor arrangement does reduce the cutin speed, but parallel resonance could lead to output reduction at higher rpm.

I am in the process of improving the wiring to cope with the new higher power levels.

A bonus of capacitor coupling of the mill is that a mill cannot in theory burn out through diode failure, as there is no DC current path to the windings.

I see this technique suitable for any 3phase setup.

I intend to trial on an 80 series stator as well.

Gordon.

PS

I have some data collected during wind gusts for a standard 100S in delta. The current seemed to plateau at about 10A. I will setup the same way with the caps and rerun. This will be a useful comparison. I expect the plateau will disappear and the current will better follow the wind energy.
Edited by GWatPE 2008-09-23
become more energy aware
 
GWatPE

Senior Member

Joined: 01/09/2006
Location: Australia
Posts: 2127
Posted: 01:37pm 22 Sep 2008
Copy link to clipboard 
Print this post

This is a comparison of before and after series caps on a 100S F&P in delta.

The comparison should be interpreted as the change in the ratio of the average of the F&P in yellow to the axial flux in purple. The F&P has 50% larger swept area.

Thge first graph is before caps with maximiser and the second shows the change in the yellow graph[F&P with caps], without maximiser. The ratio of the average of the data is the important detail.

GRAPH 1


GRAPH 2



The ratio in the first graph is 1.33
The ratio in the second graph is 1.46

The caps provide a better increase in average power than a boost maximiser.

There is a significant increase in the maximum power output, by almost a factor of 2. This would be of more use in stronger wind conditions.

A combination of capacitors and boost maximiser will be looked at next.

This should give benefit at the top and bottom of the windspeed range.

More to come.

Gordon.

PS

The graphs are battery current vs time, for the same timing period. The wind conditions had changed, but the 2 windmills were operation in the same winds, with no wind shadowing and are only 6m apart.Edited by GWatPE 2008-09-23
become more energy aware
 
GWatPE

Senior Member

Joined: 01/09/2006
Location: Australia
Posts: 2127
Posted: 11:51am 23 Sep 2008
Copy link to clipboard 
Print this post

From the marked lack of interest with this capacitor topic, I can only conclude that mill owners have perfectly operating mills and there is no desire to increase the power outout of their machines. Bryan and I may be an elite club with 500W output from a F&P single stator windmill.

Anyone who has a windmill with a normal 3phase AC output and has seen a plateau in the maximum current output, and is wanting a higher current is a candidate for adding capacitors to boost the top end power levels.

Capacitors do not make energy. The capacitors allow the windmill to produce unrestricted AC at a voltage and current that maximises the efficiency of the windings. Even though I have a 24V nom battery, my F&P mill was producing 170Vpk to pk or 60VRMS. The windings were producing a higher voltage. The capacitors seem to be working in conjunction with the rectifiers and are able to convert the voltage to additional current. The voltage across the capacitor does change and this is presumably where the additional current comes from. The system works by making surplus energy available to the load.

This is a picture of the revised windmill/battery interfaces.



At the top left is the chinese mill controller, used as an output regulator and main F&P rectifier.

The top right is the custom analogue MPPT and output regulator on my axial flux mill.

In the cabinet.

The mill output lead enters on the right. There is the mill shorting switch and above is the capacitor block and then the capacitor shorting switch. In the centre is the secondary rectifier for the boost maximiser, at the top left. Below is the 4channel data logger. The wiring allows the caps to be switched in or out and the maximiser can be switch disabled. The DC output exits on the left. The cabinet is placed to allow the lid to be used as a bench for a laptop etc.

there was a lot of chat earlier about capacitors somehow producing current and there was conflict about efficiency etc, etc.

My theory is: the capacitors are not really about power factor correction. There is probable blade, wiring, iron eddy current and diode efficiency improvements. Current that would not normally be able to travel to the battery is stored in the capacitors, like a battery. This current is able to be transferred as additional DC through the rectifier action. The emf is able to push against the capacitors in a way that allows the rotor to change speed with the windspeed. The power from the increase in windspeed and then rotor speed, produces a higher voltage. Normally the current increases and there is additional voltage loss in the stator. With capacitors in the wiring, the voltage is not held down by the battery. This allows a lower current to produce the same power, and allows the rotor to accelerate and better follow the wind energy.

In theory above a certain frequency, the capacitor becomes a very low impedance. This will allow the load to then heavily load the mill. The extra current is probably related to this relationship. I will be performing some more tests with additional capacitors across the windings in addition to capacitors across the bridge rectifiers. I am unsure if there will be much benefit. Maybe only a few less switching spikes.

Gordon.




become more energy aware
 
oztules

Guru

Joined: 26/07/2007
Location: Australia
Posts: 1686
Posted: 02:00pm 23 Sep 2008
Copy link to clipboard 
Print this post

Gordon,
I have never considered the humble capacitor as a current/voltage conversion device before, but on reflection, it happens all the time ... I just haven't looked at it that way before....... A simple real life example we all know is:

If I trickle charge a capacitor with 100ma of current for a minute or so through a resistor, (eg timing circuit) it may eventually reach full charge. If I discharge it quickly, I may see many amps at a lower voltage (splat the terminals so to speak)......Dinges made a spot welder using just this technique.

The F&P is a good candidate for this as there is plenty of volts and frequency and odd looking waveforms to play with... the caps appear to be acting as a "torque converter" in a mechanical system.

In short, your hunch that caps can convert Va to vA Is I think a valid argument, and one that I may ponder on some more.......It will also explain the spike in the leading edge of the square waveform you displayed.

If you look at your square wave, you have a period of time where charge is accumulating on the plates before giving a massive injection of voltage (and by necessity current) before it settles down to driving at the battery voltage.

This time I suspect is where the time constant comes in. It accumulates the coulombs, then discharges them in the burst we see on the waveform.

It is possible that the window has a lot to do with the other two out of phase legs conducting at the same time, if we were to do the vector addition and the phase inversion from the caps, we may find an answer in there somewhere. A three trace scope would be handy.

You must be disappointed that you finally have a working MPPT without a microprocessor in sight..

It's late, and I may not be thinking straight, will revisit this in the morning and re-evaluate my ramblings then, but at this stage I would consider larger caps to extend that burst from currently about 12% of the wave to something a bit wider???


.........oztules Edited by oztules 2008-09-25
Village idiot...or... just another hack out of his depth
 
SparWeb

Senior Member

Joined: 17/04/2008
Location: Canada
Posts: 196
Posted: 05:32pm 23 Sep 2008
Copy link to clipboard 
Print this post

Hi again Gordon,
I for one am still interested - but it's hard to keep up with a thread that has 4 pages. I'd personally appreciate whatever analytical or mathematical weight you care to apply to the experiments.

I'm tempted to play around with this stuff, too, but I'm a bit daunted by the 4-pole motor conversions that I have to work with. Compared to the FP's you have, I need caps of an entirely different colour!

Help me read those scope plots, please: The flats in the second and third plots are at about 28-29V, correct? And the scale of the first plot puts the first peak at about 22V? The time scale gives me about 180 Hz (or is it 90?), so with your 230uF caps, that's a reactance of about 3.8 ohms (voltage lagging current, ignoring the rest of the system). What's the phase resistance per leg of a FP stator anyway?

If there are spikes of voltage as shown on plot 3, does that imply a large inrush current at that moment?

BTW, I looked for RFSim99, and the whole thing was way above my head. Probably explains why I'm asking so many stupid questions, here, too.

Steven T. Fahey
 
Dinges
Senior Member

Joined: 04/01/2008
Location: Albania
Posts: 510
Posted: 05:49pm 23 Sep 2008
Copy link to clipboard 
Print this post

  GWatPE said  From the marked lack of interest with this capacitor topic

Don't equate lack of comments with lack of interest. Personally I don't feel I can make much of a contribution to this thread, so, no comments from me. That doesn't stop me from reading though.

Also notice that this thread has had 2571 views at the moment. Hardly a sign of lack of interest...

Edited by Dinges 2008-09-25
 
brucedownunder2
Guru

Joined: 14/09/2005
Location: Australia
Posts: 1548
Posted: 09:21pm 23 Sep 2008
Copy link to clipboard 
Print this post

Hi Gordon,, thanks for the post.. I'm very interested, have taken a back seat lately as I've many "irons-in-the-fire" so to speak . Nearly finished the varnishing the house ,,so next will be trying ,with lots of help , to get this capacitor set-up going. Have all the gear ,,just a matter of time..

Thanks to all of you guys for your experiments.

Bruce
Bushboy
 
oztules

Guru

Joined: 26/07/2007
Location: Australia
Posts: 1686
Posted: 10:52pm 23 Sep 2008
Copy link to clipboard 
Print this post

Gordon, an interesting observation from Tecker on fieldlines (where Byran has generated some interest over there) regarding those confounded spikes.......

Tecker:"The series Caps work while running caps around the same rating would not . You can see from the wave form what is happening the voltage rises to a spike before leakage to the other cap in the series string starts and the combined charge starts to accumulate just before the current flow . If you put running caps in there they would discharge slowly and screw with the current flow lagging the current by 30 to 40 degrees . This spike and light discharge puts you up on voltage and current you can see the current as it flattens the sine on the backside ."

Interesting viewpoint. It is a shame we dont have access to big run caps to test this out... he may have something here.

Dinges, Underground Lightning Rod mentions that he may give caps a go on an induction motor conversion. I hope he does, as he has an incredible grasp on most things electrical.... particularly on inductive devices and capacitors.......

I happen to know a certain person with lots of BIG caps and a blue motor conversion lurking around in the northern hemisphere................ He also seems to know a bit of maths to get the cap values close for a first test...(lots of series parallel configs for you)

Open discussion usually ends up with the correct answers, but also a lot of incorrect musings along the way, but if we keep talking about it, all viewpoints and schemes are worthy. The wrong ones will be teased out and it will all become clear in the end..... only been about 30 rethinks on my part so far.... and counting.

I have a lurking suspicion, that when the dust settles, a simple chart in the projects pages will become F&P folklore. There won't be an F&P mill out there without caps to match it properly... No point making a mill and not getting the most out of it.

But it still needs to be fully understood.


.........oztules Edited by oztules 2008-09-25
Village idiot...or... just another hack out of his depth
 
GWatPE

Senior Member

Joined: 01/09/2006
Location: Australia
Posts: 2127
Posted: 11:25pm 23 Sep 2008
Copy link to clipboard 
Print this post

Hi oztules,

PS edit: 3/12/08


An important consideration that should be noted, is that the capacitor coupling arrangements are charge pumps, and as such, these should not be left unloaded. An unloaded system, even if the mill is only producing a few ACvolts, has the potential to still produce hundreds of DC volts on the output over time.

Due care should still be taken and treat all connections as lethal.



The cap cct acts as a voltage reducing charge pump. The impedance of the cap to DC is the key. The windmill is not held back by the low impedance of the battery above cutin. This is the key to the wind energy tracking function.

I would not suggest MPPT. Balanced loading may be more appropriate.

My windmill seems to perform exactly the same, at the lower wind power levels. The benefits arise once the battery starts to overload the stator. The AC still gets through, where the DC would cause heating in the stator.

This is similar to what happens with a cheap audio amplifier and a speaker cone. ie clipping. This causes large DC currents to flow in the voice coil, with usual burnout.

If I knew the inductance of the stator coil, then a precise capacitor value could be determined. From the data so far, the inductance has to be around 8mH.

Hi sparweb,

20mS is 50Hz
these graphs were without a boost maximiser.
the 3rd graph is 5mS, or 200Hz, approx 18A 430rpm
the 2nd graph is 11mS or 90Hz, cutin 190rpm
the 1st graph is 5.5mS or 180Hz, approx 15A 380rpm

the blades appear to be more responsive to the wind energy since the pole fingers were twisted to reduce the cogging. The low rpm cutin does not seem to be as important now. The mill keeps spinning, as the wind lulls and accellerates quickly with the gusts.

Hi Bruce,

you still have the caps from Dennis. The caps are first separated from the end to end ring arrangement. A cap is placed as a series component in the wire between each mill output winding and the wire to the bridge rectifier.

Run the bench tests again. Benefits will be seen at the higher rpm, where normally power output would plateau.

Hi Dinges,

I get the impression that the results presented do not spark any feedback responses. I hope readers are not daunted by the findings. The only issue I have with increasing the power output, is that the diversion and all other loading has to be sufficient to accomodate the additional power available.

Gordon. Edited by GWatPE 2008-12-04
become more energy aware
 
GWatPE

Senior Member

Joined: 01/09/2006
Location: Australia
Posts: 2127
Posted: 11:33pm 23 Sep 2008
Copy link to clipboard 
Print this post

Hi oztules again,

I suspect that all AC windmills supplying a battery load will benefit from this cap arrangement. I would not restrict to F&P mills. The high pole count does help reduce the cap size though as the frequency v rpm is high.

Gordon.

become more energy aware
 
GWatPE

Senior Member

Joined: 01/09/2006
Location: Australia
Posts: 2127
Posted: 12:35am 24 Sep 2008
Copy link to clipboard 
Print this post

I have just simulated the cap arrangement on RFSim99 to estimate the inductance of the windings given the performance and the cap size used. I estimate the inductance of the 100S delta configuration to only be 2mH, not 8mH.

If I increase the inductance, then the frequency peak drops proportionally.

The inductance is the key element. Useful rpm is 100-400 for a F&P unit. This gives 50-200Hz AC output. I suspect that the 100S in delta is the only unmodified F&P stator that has a low enough inductance for caps to be of benefit. Rewires of other series stators is required. There is benefit of caps to lower voltage systems[12-24V]. Higher voltage systems ie, 48V, may not benefit due to the higher inductance of the stator. A rewired 80 series to 7S2P will have an inductance of 1/4 of the original stator.

A low pole machine with higher inductance can still benefit with caps, but these will need to be much larger to compensate for the lower frequency.

This is still not a quick fix for all machines. The inductance/voltage/pole count determine suitability and size of caps.

Time to go fishing or shopping or something;

Gordon.

become more energy aware
 
Jarbar
Senior Member

Joined: 03/02/2008
Location: Australia
Posts: 224
Posted: 12:51am 24 Sep 2008
Copy link to clipboard 
Print this post

Gordon,as I read your discription of capacitors effects I also visualized as a tourque converter as suggested by Oztules.As for totally grasping what you have done,I like Dinges am only able to say all capacitance power to you.As you are the capacitor to my understanding.Keep up the great work,the planet needs it.What would happen to a large Megawatt mill with this technology.And would large enough capacitors exist?.
"Creativity is detirmined by the way you hold your tounge".My Father
"Your generation will have to correct the problems made by mine".My Grandfather.
 
GWatPE

Senior Member

Joined: 01/09/2006
Location: Australia
Posts: 2127
Posted: 02:25am 24 Sep 2008
Copy link to clipboard 
Print this post

Hi jarbar,

this is a complicated process and a lot hinges on the inductance and the pole count and cycle frequency operation range.

I suspect the 50Hz limitations of MW machines will limit usefulness there. A lot of attention to matching source and load would already be done.

The problems with RE and battery systems is the low impedance of the battery. The series capacitors sort of makes the table level again, by isolating the direct effect of one on the other.

I hope this helps.

Gordon.
become more energy aware
 
Tinker

Guru

Joined: 07/11/2007
Location: Australia
Posts: 1904
Posted: 10:31am 24 Sep 2008
Copy link to clipboard 
Print this post

Gordon, a lot of times 'capacitors' was mentioned in this thread but I have not yet spotted what uF/voltage rating capacitors you are using in your experiment.

Thanks for enlightening me - interesting topic - I have been on holiday for the last 3 months and not read all the messages since June.
Klaus
Klaus
 
GWatPE

Senior Member

Joined: 01/09/2006
Location: Australia
Posts: 2127
Posted: 11:28am 24 Sep 2008
Copy link to clipboard 
Print this post

Hi Klaus,

The capacitor values were given by Dennis on the original capacitor thread. 220uF or about at 200-400VAC rating. I suspect caps will only work on a 24V system, as at higher voltages, inductance is too high for caps to work effectively.

Gordon.
become more energy aware
 
Phil581
Newbie

Joined: 25/09/2008
Location:
Posts: 1
Posted: 03:16pm 24 Sep 2008
Copy link to clipboard 
Print this post

Ya know after thinking about the series capacitor a while ,it sounded a bit familiar so I went to the Amateur Radio Handbook and looked up power supply circuits.
The closest single phase circuit that looks like what your talking about (only in your case there are 3 of these, one for each phase), it sure sounds like a common "voltage doubler" circuit with the series pump capacitor charging to 1.414 of the peak available voltage and battery acting as the output capacitor.

Just my 2 cents.
 
Gill

Senior Member

Joined: 11/11/2006
Location: Australia
Posts: 669
Posted: 03:24pm 24 Sep 2008
Copy link to clipboard 
Print this post

I suppose seeing how everyone is putting 2 bobs worth in there's no harm adding another amateurs view.

From my understanding the key is resonance. I see Flux beat me to promoting this though I see it in simpler terms and disagree on a couple of points with his views.

For those less familiar with this state I'll refer to this pic.



I see two major influences in this arrangement.
The first is a better matching of wind power with gen load. This is stage 1 of the idea I promoted on MPPT.
The second is increased power through lower resistance.

The first influence applies through a reduced generator loading at a non linear rate. this can be seen in the diagram as the impedance load reduces as frequency increases. In my MPPT proposal this was done by switching off the battery at a duty cycle that reduced as rpm increased. Here the generators high impedance reduces output at low rpm and increases it as rpm rises. this is equal to having a variable resistor in series and reducing it as rpm increases. We note the impedance drop off is not linear and whilst I have no idea if this is a perfect mathematical match of wind power it goes close if not. On this first factor I would not expect to see much change at low rpm perhaps a very slight rise in cut-in. I disagree with flux as caps being suitable for fixed rpm generation only as it's changing impedance is exactly in the manner required by wind power generation.

For the second influence I see a reduced impedance up to resonance. As has been stated time and again by someone oztules knows well, reduce coil resistance, get it as low as possible. Well resistance is measured in ohms, the same as impedance and this is not for some whimsical reason. Reducing impedance by adding caps to achieve a resonance at peak rpm is the same as reducing coil resistance as rpm picks up. I understand the high voltage peaks Flux warns of at resonance though disagree with runaway after that, at least no more over-speeding than would occur with an uncapped mill and that's why they all need furling, no more , no less. From the formula for resonance F = 1 over 2 Pi sqrt LC or it's derivative LC = 1 over 4 times Pi sq times F sq we can theoretically calculate caps for all series F&P.

One Thing I would do is not to back-to-back caps but rather connect and match to the rectifier diode's bias.
This may change the trace but more importantly not stress the caps with reverse potential and so increase working life.

Any raise on 2 bob's worth?
Edited by Gill 2008-09-26
was working fine... til the smoke got out.
Cheers Gill _Cairns, FNQ
 
     Page 3 of 25    
Print this page
© JAQ Software 2024