Optimizing Coil Design for DIY Generator Efficiency

[sankafola]

Hello All,
am trying to make my own Generator 5kw , am trying to find out how many coils i need , i have a drum which is full of magnet 48 NSNSNS...

for the coil i have 8 cores U shape U86A size of section is 15 cm ( 56x28mm )
how to find out what gauge for the magnet wire and how many turns
and how much they will give max as voltage and amps ?

for example if i put AWG 18 wire ( 1 mm ) what would be the amps and voltage ? my hand drawing showing the drum where all the magent will be turning and the coil inside fixed

 

[essenmein]

There is no where near enough information in the post to even begin to answer that question.

If DIYing something, the approach I'd take is put a known number of turns, eg 1turn per coil. Rotate the system at the desired speed then measure the resulting voltage. From there you can calculate the number of turns needed to reach the target voltage, once you know the turns count needed, you can work out your maximum wire size to get to a decent fill factor in the coils.

 

[sankafola]

what wire gauge ? let say its gauge 18 how many turns to do 50 volts? if the section is 15 square cm

 

[sankafola]

is there is a formula for coil ? so i can put the right wire gauge and the right turns ?

 

[essenmein]

V=N*(dΦ/dt), N = turns count, dΦ/dt is rate of change of flux the coils are exposed to.

Maximum current is determined by the heat dissipation capacity of the machine.

 

[sankafola]

a lot of unknown this way , if i assume the Φ ( thickness of the wire ) =0.72 x Squar of the amps so if i want 2 amps then the thickness of the magnet wire need to be 1 mm is that right ?

and if i assume the section to be 15 cm square then n=k/s so k=45-50 so number of turns is = 45/15 = 2.8 turns per volts

so for 50 volt i would need 18 turns of wire AWG 17 , how far am i ? is that calculation is right >?

 

[enorbet]

It is interesting that you seek to achieve 5KW since that brings back memories of a kit i bought about 40 years ago from a company listed in The Whole Earth Catalog. The kit was for a set of instructions for adapting a Dodge Dart w/ Air Conditioning alternator to produce 5KW-6KW by crafting a wooden block as a coil form and increasing the gauge of the existing wire. Since this was for a Wind driven electrical generation system and those have enjoyed increased popularity, I imagine an internet search would easily result on some hits for a more modern source. The advantage as I see it is all the practical, concrete hints at how to turn the mathematics into an actual physical project. Maybe you would find that helpful as well as encouraging.

FWIW, mine worked quite well but I needed more than 6KW and I was lucky enough to find an old alternator owned by the operator of a lumber mill who used a water wheel that put out over 20KW at low rpm that I happily bought for $30 USD and powered with an old VW motor at fast idle of roughly 1200 rpm.

 

[Baluncore]

so for 50 volt i would need 18 turns of wire AWG 17 , how far am i ? is that calculation is right >?

Find the numbers by winding a 10 turn coil from thin wire. When you spin the alternator at its expected lowest operating RPM, measure the open circuit voltage produced.

You can then work out the number of turns needed. Work out the current from amps = watts / volts. Select from wire tables a gauge that will carry that current.
https://en.wikipedia.org/wiki/American_wire_gauge#Tables_of_AWG_wire_sizes
Work out the length of wire needed to put that many turns on your coil former. From that work out the resistance. Some voltage will be lost in that resistance; Vloss = ohms * amps. Add another turn or two to make up that lost voltage. Power loss will be; watts = amps * ohms²Work out the cross section of the winding and check that it will fit on the core. Select a bigger core, or recompute with thicker wire until you have an optimum solution with minimum power loss.

 

[Baluncore]

The LG motor has the number of poles on the three phase stator closely related to the number of PM poles on the external rotor. The sketch you provided shows few (7) magnetically closed coils being passed externally by many (48) closer pitch magnetic poles. That will be very inefficient. You will need to modify your proposed magnetic circuit. Start by tracing the magnetic paths in the LG motor. Mark the alternating N S poles on the rotor with different colours, then notice how they pass the stator poles and the way the poles are wound.

The large diameter of the LG motor means it is only a low RPM motor. It will not survive high RPM. You might consider rewinding the original LG stator to get the voltage you require at the RPM you will drive it. The stator section available to enclose the windings limits the power that can be generated efficiently. I would be surprised if the LG stator could generate more than the 750 watt it was engineered to convert. To increase power you would need to proportionally lengthen the motor, (wound stator and PM rotor), to increase the area available for the windings.

 

[sankafola]

@Baluncore hi, so what is the rule to match the coil with the poles? i have 48 poles NSNSNSNSN... do i need 48 coil ? the LG motor have 36 coil and 48 poles

the coil in the LG motor is not good to do a generator , max is 1200 Watt , so what i want to do is i do an ANT LEnz coils a new design ,

can you tell me what is the rule of the number of coil versus the number of the poles ,

 

[marcusl]

Your cores look like ferrite, which would be essentially worthless for your generator project.

 

[sankafola]

can you explain why ? what if its nanocrystal ?

 

[Baluncore]

The LG direct drive motor looks like the F&P smartdrive motor developed in NZ.

the LG motor have 36 coil and 48 poles

Check the counts. 36 coils will probably make three phases, each of 12 coils in series. The magnetic rotor will need to have 12 N-S magnets that repeat every three coils. That makes 24 magnetic poles, not 48. The 24 pole rotor could be a ring of 12 flat magnets stuck on a magnetic backing plate. Hold a magnetic wire near the rotor to feel and identify the magnetic poles.

Each coil is wound on a stator pole. Follow and visualise the magnetic field lines from a N pole on the PM rotor, across the gap into the stator core, then through the laminated core surrounded by a winding, probably splitting in two before returning back across the gap into the adjacent S poles on the PM rotor.

Each time the rotor moves by one magnetic pole, the field reverses through each coil.
Each time the rotor moves by one N and then one S pole, it produces one cycle of AC voltage.

I expect every third coil on the LG motor is wound in series to produce one phase. All those poles in series produce the output voltage in phase. Three phase requires the number of coils be a multiple of three. Every third coil should be opposite a N pole at the same time.

7 coils will not work efficiently with 24, 36 or 48 poles because 7 is not a factor of 12 and the poles need to line up. Notice that as the PM rotor moves through each magnetic cycle, it sequentially passes three evenly spaced windings, to generate the three phase voltages separated by 120°.

 

[sophiecentaur]

can you explain why ? what if its nanocrystal ?

You seem to be very (even over) optimistic about this. You need to buy a book on alternator design and become very well informed about the whole business. (Becoming an EXPERT even) You can't hope to get enough from Q and A on PF - despite the wide knowledge and experience base that PF has. You need to know what questions to ask and not to miss out vital ones. Anything you come up with can have one Achilles' Heel and that could make it fail to meet requirements and that would be upsetting. The upside is that you will surely get some power out of the thing and, if you make the stator connections re-configurable, you will be able to optimise the output.
DO you have the facility to machine things to make the gap as narrow as it needs to be?
PS Copper is quite expensive and it is usually very hard to re-use any wire from one version in the next one and to get the best packing density of your windings. So it could cost a lot.

 

[sankafola]

@sophiecentaur i have that Motor LG washer machine and as a prototype i put on the outter surface NYD 52 NSNSNS... and i made a small coil " Anti lenz " ONE coil and it was making 24 volts 1 amp and raising the input of the motor 3 watts only so open circuit 24 watt for 3 watt and my core was just alan key , so i think the nanaocyrstal will be better and i will put the coils inside instad of outside , i work in electromechanics but am not into motors and generators

so now i need to do factor of 12 for my coils , my question is how to determine the number of turns i decide to go with AWG 18 , or maybe i can go with AWG 24 + AWG 18 together one for volts and one for current but how many turns ?

 

[sophiecentaur]

i work in electromechanics

In which case you will know about Magnetic Circuits. The fields have to follow a loop from the magnet, through the moving coils and back again. Any gap is an additional series Reluctance and will reduce the effective field through the coils. The cores would ideally have a radius on the faces with the magnetic rotor to get a 'better magnetic contact' across the gap.
There ia a lot of stuff n alternators but, as you'd expect, the direct answer to your question isn't there. I did find this link - any use to you?

 

[sankafola]

i have a generator here 30kw what is the proper way to test if its really 30kw can you tell me , without a load , its suppose to output 600 v DC 50 Amps what is the best way to test if its really 30kW ?

 

[sophiecentaur]

I would say NO. You can’t just look at emf and short circuit current. You could always use an array of water or space heaters (borrowed or from a hire shop.) I could do that easily in UK.

 

[sankafola]

I would say NO. You can’t just look at emf and short circuit current. You could always use an array of water or space heaters (borrowed or from a hire shop.) I could do that easily in UK.

But if am testing without load I should be able to get the 600 v dc why am only getting 400 v dc at 2000 rpm

 

[sophiecentaur]

When you say “should”, what is your basis for the statement?

 

[sankafola]

When you say “should”, what is your basis for the statement?

It’s because I bought it rated 30kw 600volts dc 50 amps

 

[sophiecentaur]

That’s a pretty good reason! Did you have the right revs? (That’s the only excuse I can suggest. )

 

[sankafola]

well they wouldn't say at what RPM i made a youtube video on it :) so you tell me what is the best way to find out i would say the voltage at the end of this video i tested at 2000 RPM i got 376 v DC

 

[sankafola]

so that is why i decided to make my own Generator :)

 

[sophiecentaur]

So you reckon you can do better than a proper one? If the one you bought is faulty or sub- spec then take it back.

 

[sophiecentaur]

Why try 2000rpm? 3600 is common iirc. A whole range is needed.

 

[sankafola]

ok will try that

 

[Baluncore]

But if am testing without load I should be able to get the 600 v dc why am only getting 400 v dc at 2000 rpm

To get 1.5 times the voltage you need to cut the field at 1.5 times faster, so maybe you should really be spinning it at 3000 RPM.

2 pole at 50 Hz, synchronous = 3000 RPM, induction with 4% slip = 2880 RPM.
2 pole at 60 Hz, synchronous = 3600 RPM, induction with 4% slip = 3450 RPM.

 

[sophiecentaur]

You need a revs / volts curve with no load and with whatever load you can get hold of.
You surely can’t believe that you can do better with a home made machine at any particular speed.