Transformer isolation for measuring turbine rpm

[zap1]

Hi,

I would like to monitor my turbine's rpm using Hugh's circuit "http://www.scoraigwind.com/circuits/index.htm". What kind of transformer should I use to isolate the generator ac wires from the circuit? I tried without the transformer but the turbine would not spin freely. I think the problem is that one of the AC wires is grounded by my rpm circuit. The circuit is powered by a 12V battery.

My setup:

Wind Turbine -> 3 phase AC -> rectifiers -> 12V Battery

2 AC wires from generator -> LM2917 circuit Fin and Gnd.

Thx,

Zap1

[willib]

the problem as i see it is that your transformer is loading down your alt.

we would need a sketch of what you have done to provide a better answer

[Flux]

The 2917 needs about 200mv and will tolerate 28v so you need a transformer to give something in that range.

You didn't mention your system volts. Remember the transformer may have to handle quite low frequencies so something like 220v to 6 v  or 12v should be ok. If it is a 12v system then 120 to 6v should do fine. It only needs to be a few VA.

Flux

[zap1]

It's a 12V system. The 2917 has a zener in it to limit the input voltage from the battery.

Does the transfromer resistance matter? If the resistance is too low, won't it load down my generator? Besides, the voltage step down spec, what else should I consider when picking a transformer?

How about using optoisolator like 4N28? The problem with opto is that it won't turn on the led if the rpm is too low.

[Flux]

The zener in the 2917 is for supply voltage regulation, it has nothing to do with the maximum input voltage.

Transformer resistance is mostly irrelevent, it is the winding inductance that determines the load it presents. That is why you need a reasonably high voltage transformer, so that it still lakes a reasonable magnetising current at low frequencies.

There is nothing very demanding about the transformer at all, anything in the region of 120v to 3v up to 20v should do. Even an ac output wall wart should be adequate.

It is possible to capacitively couple the 2917 to the diode bridge and it is also possible to use an opto isolator but in either case you really need to know what you are doing. The 2917 is a really nice device but its input circuit is not the easiest thing in the world to drive from messy waveforms associated with rectifiers, it will easily respond to harmonics in the waveform and give wrong answers. The transformer is far easier and even then you may need a bit of filter under certain conditions.

The input of the 2917 must cross zero, if you use an opto isolator you will need a capacitive coupling to the input.

Flux

[Ding123]

This is a good post.Thanks.  

[zap1]

Thanks a lot, Flux.

Can I use something like this from radioshack? http://www.radioshack.com/sm-isolation-transformer--pi-2103994.html

I'll post my circuit drawing later.

[Flux]

No not that thing, that is designed to work at audio frequencies and at millivolt level.

You need a mains transformer intended for a small power supply, anything from 3 to 100VA will do . you will most likely find something suitable in a dead radio or cd player or a wall wort. Failing that I believe you have Mouser electronics or Digikey in N America. I am sure some local folks can point you in the right direction. Even a small battery charger should give you a suitable transformer.

Flux

[Ungrounded Lightning Rod]

Another thing you could do is use a resistor from one of the AC output lines to a  couple of diodes - paralleled facing opposite directions - to ground.  Hang a cap across the zeners to suppress harmonics.

(I presume your negative rectified output is tied to ground somewhere.  Asuming that, where I say "ground" I mean the negative side of the bridge rectifier.)

Assuming your voltage will never go over 200V RMS even if the batteries come unhooked, a 47k resistor will not dissipate more than a half-watt.  (So use a 2 watt carbon composition resistor.)  That will give you about a quarter miliamp at 12v and about 4.25 ma at 200V.  You can use 1n914 switching diodes.  (They'll survive 5 ma continuous at up to 165C and 45 at 100C.)

Run another resistor (same value) from each of the other two AC lines from the genny to ground if you want stable operation at frequencies below cutin (though of course your meter will stop reading if the mill gets too slow.)

If you have trouble with harmonics you can hang a capacitor across the diodes.  (Picking the capacitor might be tricky, since you don't want it to wipe off the signal of interest when the mill is running at high speed.)

[Ungrounded Lightning Rod]

Hang a cap across the zeners to suppress harmonics.

I'd started with zeners but realized that switching diodes would provide adequate voltage, leave most of the dissipation in the resistor, and would be easier to find.

If you can't find 1.914s virtually any switching diodes will do.  Or you could use rectifier diodes like 1N4001.

[commanda]

My circuit is here. R1 to ac side of bridge rectifiers. Component values for 12 volt.

Amanda

[wooferhound]

A transformer is a Dead Short to DC voltages. If you have really low frequencies this would be a problem. It would really make a difference at startup . . .

[Flux]

Yes large transformers used for power matching can delay start up and it is one thing to watch when using high voltages and transformer matching. If the flux density is reasonable they will start without removing the transformer.

In this case, using a tiny transformer to drive a frequency meter there will be absolutely no problem. Any suitable transformer chosen for this would have such a high primary resistance that it would never be noticed during start up.

Flux

[BigBreaker]

Why not put a small neo on the back rotor plate and use a coil to pickup the signal directly?

That also avoids using a possibly dead coil to sense an RPM that may be at freewheeling levels.

[zap1]

My first RPM circuit has a current limiting resistor (100K) connected to one of the AC output lines, 2 1N4004 diodes paralleled facing opposite directions to clip the voltage seen by the LM2917. It works great (even at low rpm) with just a turbine and rectifiers to test no load rpm.

But once I've connected a battery as a load to the turbine, the turbine will spin very slowly. It feels like 2 of the ac lines are shorted. The battery negative, rectifier negative and one of the ac lines to LM2917 are all grounded together. Could this be the problem? That's why I need a transformer to isolate the alternator ac line from the circuit ground. But with the transformer, it won't detect low rpm precisely.

[willib]

have you thought about feeding the tacho chip with the output from a muffin fan circuit board ,

i'm not sure if the tacho chip needs to cross zero or not , but an opamp can cure that.

the muffin fan has a magnetic (hall effect switch ) built in , it also has two transistors to drive the fans coils .

the hall switch detects the position of the rotor and tells the fan when to energize its coils.

this way you would not need a transformer at all.

only put the hall switch near your  alt. magnets , i've even seen it switch from the residual magnetism comming out the back of the rotors , so it wouldnt need to be placed on or near the stator.

the muffin fans can be found in any pc power supply , and run happily on 5V as well as the 12V they are rated for.