When winds are low

[willib]

When winds are low ,M1 is off , M2 is is boosting into the right 12V battery

when winds are med , M1 is on , M2 is off , and the left battery is charging





When winds are high M1 is on , M2 is off and any current that is excess will flow into the right battery also , minus a diode drop .

huh huh

what do ya think

 should work

M1 is also be pwm controlled .

do ya think i need a third Fet under the righmost batt

Cap is optional. But stores power in low winds , for boosting.

[Flux]

Should work. I can't see why you want 2 batteries but fine if you do.

I suspect a reasonable sized low esr electroytic will work better than that 2 Fard thing which will not be a true capacitor.

Flux

[willib]

thanks Flux for the reply

hmmm

A single battery version.

M1 is optional , but i would use it to let the blades reach their optimum tsr.

this is assuming the alt is more powerful than the blades, ie the alt could stall the blades in all but a very strong wind.





by varying the pulse width into M1 the current into the battery is adjustable .

so in low winds M1 would be full on ,M2 would be pulsing current into the batt

in med and high winds ,the pulse width into  M1 could be varied to let the blades reach their tsr,and M2 would be off .

[Flux]

One or two points I don't like. The basic boost converter is fine, you can control its pwm to track the alternator load to the prop during boost.

If you make absolutely certain M1 can't operate while the boost is in operation it may be ok, otherwise the boost converter will produce monster voltage spikes when M1 is open.

I suspect it will be ok on a small scale but I wouldn't want to do it with a large machine. Either use boost in low winds and accept the match in high winds without M1 or convert M1 to a true buck converter for control of load matching and wind the alternator not to need the boost.

In theory you could use a buck/boost converter but that is not without its problems with polarity inversion.

You are on the right lines, just prove one stage at a time.

Flux

[s4w2099]

Yeah, and not to mention that those high voltage spikes could fry M2. Watchout

[commanda]

Sounds like you're about to discover the SEPIC converter.

Quote from wikipedia;

(http://en.wikipedia.org/wiki/SEPIC_converter)

A SEPIC (single ended primary inductor converter) is a DC-DC converter which allows the output voltage to be greater than, less than, or equal to the input voltage.

Amanda

[willib]

you give me way too much credit

i found this when following your link

http://www.national.com/an/AN/AN-1484.pdf

its how to design sepic converter

thanks

[willib]

i thought i would add a dump load , to round out the process.





with Fet M1 off there is no drain on the battery during dump.

[RP]

I wonder if these spikes could be a problem for the supercap or the rectifiers also.  It might be a good idea to parallel the supercap (or electrolytic if you substitute one) with a polyester or tantalum to grab the high frequency stuff so it doesn't punch holes in the dielectric of the big cap.

[s4w2099]

I dont know much about the cap but the rectifier are not as sensitive as the MOSFETS. You can do the following:

The diode across the coil is to short any reverse voltages caused by sudden current changes. That should take care of that. The Transient Voltage suppressor will leak current into the gate of the MOSFET whenever the voltage exceeds the specified voltage of the device and with this current and voltage in the gate the MOSFET will conduct shorting the output thus lowering the voltage and killing the spike.

The value for the suppressor should be chosen depending on the max voltage charasteristic of your MOSFET. If your MOSFET will only take 100v to fail choose a value for the transient voltage suppressor a bit lower than that, usually 30% lower will be safe. Now everything beyond that MOSFET will be safe from high voltages.

Hope it helps

[windstuffnow]

  Willib, I'd like to try my luck at building a simple version but have a couple of questions... Is L4 a torrid?  how many turns of wire should it have?  I don't know anything about the PIC your using, could the mosfet be driven by a 555 and if so at what frequency would be best?

  Keep in mind I don't know alot about electronics, I'm just an old fart that likes to tinker with things and learn...

.

[willib]

just sent ya an email

[Stynus]

That diode over the coil is destroying the working of the circuit.

The reverse voltages are used to boost the voltage.

And the other 2 diodes is your drawing are useless to because there is never going to flow current trough them like that.

[willib]

Ed i havnt worked with a 555 timer in a long time.

but i googled ' 555 pwm ' and one link that poped up said something that i remember

the 555 will have a hard time turming fully on, so if you use a 555 for pwm you will have a hard time turning a Fet completely 'on ' and 'off' , ie , without it pulsing a little bit ,

example : when turned all the way up the 555 may be pulsing 95 percent on, and 5 percent off at the frequency you choose , which may cause problems ( which Flux was talking about) if you are pulsing M1 above and boosting at the same time.

on the other end , when the 555 pwm should be all the way off , ie, no boost , the fet should be all the way off , and i'm not sure if the 555 can be all the way off , i just dont remember.

another option is i could program a pic for you , all i do is hook up a pot ( varible resistor) to the pic and a oscillator (clock) to a programmed pic , turn the pot all the way down , Fet is guarenteed to be off , turn it up a little and the Fet starts to pulse  , if i turn it all the way up the Fet will be fully on , no pulsing but a very low resistance RDS(on).

another thing ,there are Fets that can be turned on with a low voltage ,called logic level fets like the IRF3703 and similar ones , i use the IRF3703 with a direct connection to the pwm output of the pic. i got my order from microchip the other day and one thing i ordered was some Fet drivers , i cant wait to try them on the boost inductor , just to see if the output improves..

in the meantime look around for some old pc monitors , they have a lot of usefull parts like transformers (ferrite ones) , i just cut it off the circuit board ,throw the whole tranny in a jar with acetone in the bottom , close the lid , and shake occationally , after a little while the tranny comes apart easily.

but dont force it they are easy to break .

if you want to go the 555 route

http://cache.national.com/ds/LM/LM555.pdf

page 8 & 9 are a good start

a 1K resistor and a 1micro farad cap should get you in the range from 100Khz

down to 100 hz.

the range you are looking for is 20 Khz to 100 khz.

i think Flux recomended 30 Khz to 50 Khz?

the higher the pwm frequency the smaller the inductor needed.

if you have a 'scope' about , it will help alot.

[s4w2099]

Ok, you are wrong.

First we do NOT need the negative side of the spikes in that part of the circuit. They will cause damage to your MOSFTET. In this case since its 3 phase it should not be very necessary.

As for the other two diodes. Transiento voltage supressors are designed to work backwards. They will not conduct electricity untill you reach the reverse voltage rating that I was speaking about. This diode will conduct normally in forward as any other diode. Knowing that I added the other diode after the transient voltage suppressor connected to the gate just in case the first diode connected across the coil fails to kill the negative voltage, it wont go through the second NTE125.

This is a proved design that I use for the same purpose in power inverters. The same concept applies. I copied this design from Trace power inverter plans. Its bets to study electronics from what other people havo done. With a scope you can see how it works.

[s4w2099]

I forgot to mention, if you need the negative side of the spikes for the circuit to work for any odd reason unknown to me you can just remove the diode across the coil. the rest of the circuit will work the same. I just placed that diode there for extra security against reverse voltages across the FETS. That technique is often used with relays.

[willib]

I've moved the two 1 farad caps to the boost side , because, during bench testing , if i let the boost continue till tha caps were spent , the blades would have to start up in a stall condition.

 till tha caps charged up.

the caps seem happier where they are now

they can take any extra boost and feed the battery during a lul.

still waiting for my new programmer kit to arrive , hopefully tomorrow.

tomorrow or this weekend look good for testing , we're expecting very windy conditions.