Zero Volt Diode

[South Easter]

I would like to be able to measure my generator output voltage, even when it is below the battery voltage (at the moment, the only way to do this is to disconnect the battery).  I could add a diode, but that means a voltage drop and power loss.

I found the following circuit - can anyone comment on whether its will work, is worth building, or if better components could be used?

[Flux]

Not sure what you mean.

 If it is an alternator you can measure the ac voltage on the input to the rectifier, or add 3 auxiliary diodes to the bridge to measure dc volts below cut in.

If it is a dc commutator generator you need a series diode to prevent it motoring below cut in so why add another one.

It is possible to use mosfets as low drop diodes but I think you are making a big issue about nothing and making lots of work and potential trouble.

Flux

[South Easter]

Thanks for comments.  Its a PM 3 phase generator.  The rectifiers are on top of the pole, and getting extra wires down would be a hassle.

I covered the option of extra diodes in my post - this loses some voltage and power.

Such a circuit would allow data logging of generator (rectified) voltage and battery voltage separately.  Together with wind speed this would be useful.

Comments on the circuit itself would be much appreciated!

[Nando]

Since you have not given enough information, the only thing to do is to wild-guess what you have.

You have a generator up the tower with rectifiers and just the DC power is brought down and connected directly to the battery bank.

Since you do not WANT to disconnect the battery, then you have a problem when wanting to read the generated voltage.

A solution is to have a relay between the DC power coming down and the battery, read the voltage and when the voltage is just above battery voltage, close the relay contacts to charge the battery bank.

The logical way is to measure the voltage before the diodes, to have the battery isolation needed to determine the generator voltage when below the battery voltage level.

The alternate way is to have the generator in a test bench with capabilities to drive the generator at known RPM to get the voltage generated or the Volt/RPM factor and from there, the only thing needed is to have a small circuit to measure the RPM which give the Voltage output at any RPM level.

The circuit you have is a Flip-flop,and does not do what you want and since I see that is not practical for your needs I will not spend the time to explain how it works.

Nando

[Flux]

Yes you do have a problem if the rectifier is up the tower.

A series schottky diode would not cost you anything in real terms. There will be a power loss in it but you will not notice it as the slight increase in prop speed will probably bring you away from stall and there will be no detectable loss by including it.

I am not sure about your circuit, I can't read it without saving the drawing in  something to enlarge it. I don't think it will be completely satisfactory. I think you need an op amp or comparator to be precise enough to detect the turn on point.

For mosfets to have an on resistance much lower than a schottky you need lots of them in parallel. Ok for a tiny machine but if you are looking at lots of amps then you need several mosfets even if you choose those with very low rds on values.

Unless you are using some form of mppt to keep the prop on the peak of the curve, line loss and a bit of diode loss is something you have to accept. You can remove it at great cost but you can't extract the power that would otherwise be lost as you stall the prop and drop its aerodynamic efficiency more than you can gain on the electrical side.

Flux

[South Easter]

Nando, all your guesses are spot on.  I ran a simulation of the circuit using LTSpice/Switcher CAD III (the diagram is a screen shot of that), and it seemed to work as intended.  Not sure how to make diagram clearer - its all very clear when I view this post.  Any suggestions welcome.

But it sounds like a schottky diode is the way to go?  Or maybe germanium?  Flux, can you recommend a specific part number?

Is there any reason not to rectify (at the top of the pole) using schottky diodes?

I did bench testing, and found a KBPC3510 rectifier gave DC at 2.4V below the AC voltage.  A SB356 gave DC at 1.3V below. I was surprised at that difference!

I work at 24V, so that kind of drop means there is a lot (ok, some) of the time when the generator could be delivering current but isn't due the voltage drop across the diodes.

(for those who want all the details, see http://windpower.org.za/benchtests.html

and for the coils layout etc see http://windpower.org.za/drawings/drawings.html

or for Everything you Possibly Want to Know and More see http://windpower.org.za )

--

[Flux]

I would forget germanium diodes, they are no longer readily available and to be honest they are not as good as schottkys.

If you can put up with the dreadful TO220 package you have a lot of choice, if you want a decent package the choice is very limited and the cost is high. You don't say anything about your current so I really cant give much more help. For the series diode as you are needing it a 40v device will do.

Yes you can also replace the main bridge with schottky but I don't recommend it. They are rather close on voltage for a 24v system and if for any reason you loose the load you will blow them. If you can guarantee the output will never go open circuit then they will be ok but you will need 100v devices and their volt drop is not so low.

There are significant differences in ordinary diodes but it will make precious little difference to your output ( it will make a big difference to the size of heat sink needed). The potted bridge thingies are convenient but they are not the best choice, they have higher drop and need more heat sinking. Decent individual diodes of the stud mount or pressfit type will give your more current carrying with less heatsink.

In terms of actual output you will see no difference on a 24v system, it may seem rather serious on a test bench with constant speed drive but under wind conditions it will make no detectable difference, at cut in the power from the blades just overcomes the system losses and as soon as the diodes conduct you will just drag the speed down, it will take considerably more wind to give you enough to show a useful charge. ( if you are worried about a few tens of mA then ok, but real charge it makes little difference). Even in high winds the main gain from more efficient diodes comes down to less heat sink, it will not significantly affect the output.

At 12v things are starting to have real effect and schottkys may pay off, but at 24v  I think that in the end conventional diodes will prove just as good and more reliable.

For the same reason I think your interest in volts below cut in is a red herring unless there is something seriously wrong with your cut in speed, there just is no useful power in the lower winds. Use a schottky in series for your tests to check your cut in speed is near ideal, then short the thing out and forget volts below cut in, bearing friction will have more effect than that diode drop.

Flux

[Nando]

Samoa:

Your voltages indicated are undefined and the circuit itself is not well determined.

for one include the Fet intrinsic parasitic diode and you will see the differences and limitations.

The V1 voltage will be inserted into the circuit when its voltage is within V2 less 0.6 volts, also the 2N3906 Base Emitter junction is in reverse mode allowing the breakdown ( Zener voltage of it ) flow current causing transistors behavior to be affected.

Then if the V2 voltage is greater by 0.8 volts than the V1 voltage, the Fet intrinsic diode to be turned ON ( reverse mode operation)- which may give you problems if V2 does increase radically.

If you want and really need to measure the voltage at any time, then another circuit is needed, like to be installed up the tower and brought down either with wires or wireless device to send the data.

Proper data taken is BEST to define the generator Volt/RPM and RPM measurements will give you the generated voltage even under any load and if you have the generator internal resistance, then you may be able to calculate the stator watts dissipation at any instant.

One needs to think in practical terms the parameters and the design of a circuit to see if it is applicable for one needs.

Nando

[scottsAI]

South Easter,

Looking over your web, looks like this is a 50watt PMA?

The cost of doing something must be weighted against the benefits.

If the gen spends lots & lots of time just below cut in then doing something about it may be worth it.

Integrate the power and time, I do not think the cost of the schottky diodes will be worth it.

Remember the diode rating must be above the max voltage seen. At furling the voltage will be several times higher than cut in. 100v to 150 volt schottky are not cheap. I design using the max expected voltage to be no more than 1/2 to 2/3 of rated device voltage.

You should consider a voltage booster before I would worry about schottky diodes.

Have fun,

Scott.

[South Easter]

Thanks for advice Scott!  I have seen a max of 10A coming down the pole so far (240W) and hope to tweak that up a but, so I would like to allow for 20A, and so use components maybe rated for 30/35A.

From what I understand then, I could use 40V schottkys.  Trawling through a local components store's website, I found:

Model    :    STPS1545CT

Device    :    Schottky

Type    :    Power

Forward Voltage @ Forward Current    :    0.57V

Device Package    :    TO-220AB

Termination    :    Plastic

Repetive Peak Reverse Voltage    :    45V

Max Forward Rectified Current    :    7.5A x 2

But that seems to have a 0.57V drop!?  So not much better than normal diode anyway...unless maybe drop is much less at lower amps.  Any comments appreciated.

Is a 'voltage booster' the same thing as a DC-DC converter?

[Flux]

I think a better device would be 40CPQ040 which is TO247.

With both halves in parallel it would drop just under .5v at 25deg C with 20A current. The volt drop falls quite a lot with temperature and will be about half this at a junction temperature of 100C.

Try one, then short it out and let me know how much your charging current increases under real wind power ( not on a rig). If you parallel 2 devices the volt drop will come down a little bit. Typical conventional diodes will drop about twice that if lightly loaded or about 3 times if pushed.

Flux

[Opera House]

No one likes to talk about the reverse leakage of Schottky diodes that can make a reading incorrect unless a load is placed on it.

[Flux]

You are absolutely right.

 I never trust the reading of a bridge rectifier either into a digital meter without a nominal load of a few mA. some of the waveforms into a 10M ohm meter are very strange indeed.

To measure the dv volts below cut in with a schottky in series I would load the meter with about 4.7k or use a good old analogue meter.

Flux

[scottsAI]

South Easter,

40. v may sound right but it will not last, I'm quite surprised nobody busted your chops for saying it:-)
    
24. v PMA will generate much higher voltages.
    
    

The diodes must be rated for the worst peak voltage.

PMA rated for 10amps what is the generators output resistance?

With the wiring lets say it's 1 ohm. To inject 10 amps into the battery you will need to generate the 28v battery and 10 volts = 38v, two volts shy of the diode's limit.

Your hopes of 20 amps will put you well over the voltage limit.

Further more, a PMA running at higher winds than cut in will be operating in stall, or RPM limited.

If for any reason the battery is disconnected the output voltage will scream up, 4 times normal operating voltage are not uncommon.

Several have shown placing resisters in series with PMA in higher winds did not reduce the current into the battery. Their blades turn faster at a more efficient speed able to supply more power into the battery and heat up some water. I do not expect this to work for everybody. Just a point to understand what is going on! Most PMA systems have a rather linear output vs wind speed, when it should be a cubic function.

Reverse voltage rating has nothing to do with current rating.

When a diode's reverse voltage rating is exceeded it acts like a zener diode!

Which then causes it to go bang.

Read around, many dead diodes reported here. Three ways to kill a diode.

1. over current
   
2. over reverse voltage
   
3. over temperature.
   
   

Good luck and hope lightning does not do them in.

For the cost of a Schottky, you can buy a low resistance MosFET and get the drop out even lower, does require some additional circuitry as you have shown.

Is a 'voltage booster' the same thing as a DC-DC converter?

Yes it can be or it can be a transformer or star / delta switch. Most do not like star / delta.

Actually most do not like voltage boosters of any kind. Electronics required inside!

Have fun,

Scott.