Matching to the impedance of an alternator.

^^^^^^ This should have been "first", the way I posted the links.

The web is full of statements that AC synchronous motors are not self-starting, and many explain why. What's to not understand?

I explained why they usually won't self-start, before I looked for confirmation on the web. Believe what you like, but don't expect the world, or the motors, to agree.

How is this wrong?

Starting methods

Synchronous motors are not self-starting motors. This property is due to the inertia of the rotor. When the power supply is switched on, the armature winding and field windings are excited. Instantaneously, the armature winding creates a rotating magnetic field, which revolves at the designated motor speed. The rotor, due to inertia, will not follow the revolving magnetic field. In practice, the rotor should be rotated by some other means near to the motor's synchronous speed to overcome the inertia. Once the rotor nears the synchronous speed, the field winding is excited, and the motor pulls into synchronization.

John

Lunch yesterday, at Ra, relatively modern multi-speed bike chained to the wrought iron fencing around Ra's outdoor eating area.

NO dynamo, hub or edge, but it had lights, head and tail... obviously battery-powered... perhaps with switchers for efficiency.

Any cyclist here to comment? ...Jim Thompson

Great! ...Jim Thompson

Ian implied that he wanted to derive more charge current. Is that correct, Ian?

I still think PowerMOS as rectifiers would be the best solution. ...Jim Thompson

Improving the load matching (as opposed to shunting with bulbs that overwhelm the alternator - or a SLA battery which more or less shunts the O/P voltage) by using a buck converter to draw less current and allow the more lightly loaded alternator to attain a higher O/P voltage - and hopefully obtain some improvement on the rated 1.8W output.

It is intended that the buck regulator output will be set around 7.2V to charge the SLA which is currently direct from the bridge rectifier.

The digital temperature outside probe is telling me 10.5 DegC, but there's 3 of them (when they all work) and have a spread of about 3-4 DegC between them.

The source is operating almost constant-current into a low-voltage load like the batteries. Improving the rectifiers won't help much.

Is that the thing you are intending to post here today?

John

Consumer-grade temperature measurement stuff is all over the place. Just look at a bunch of thermometers on a store shelf.

I used 1K platinum thinfilm RTDs for my remote cabin automation system, measured ratiometrically against 1K 0.1% resistors. That's accurate to about 0.2 C.

John

It'd buy some improvement at low speed which I'd opine is where he spends a lot of his time. (His route can't be downhill both ways :-)

Some real curves from a Sturmey Archer hub would be interesting.

No. ...Jim Thompson

The 43F reading is from a large dial probe type thermometer rescued from a scrap commercial fridge out of a food/catering shop.

My K-thermocouple DMM is in a tea chest somewhere at the back of the garage.

Without knowing the length, dia & tempco of the winding wire, I doubt absolute accuracy of temperature is all that important.

intending

The increased current that you would get from ideal rectifiers is minimal. The source impedance is too high. As Ian says, the issue is to better match the alternator impedance to extract more *power*. That involves picking a higher point on the power transfer vs Rload curve. I chose to stay low on the curve and use a simple current multiplier, a transformer or a fixed-duty-cycle buck switcher. You might prefer to move to the high end of the curve, rectifying to a high DC voltage and switching down. That has complications, and doesn't take advantage of the design of the alternator, but is do-able.

The stuff Marcel posted, in German, is excellent.

OK, show us. But don't neglect your honey-doo assignments.

John

Copper is only 0.4% per degree C.

John

You could build the simple fixed-duty-cycle buck. A trimpot could set the current multiplier ratio. Or find a transformer with some windings that give a few nice ratios.

John

Same current for longer time makes for more _charge_. It's not clear from Ian's post how far he is from maintaining a good battery charge.

Combining my past post with some other data...

The current input requirement to a "normal" buck switcher (with a feedback loop setting output voltage) is the _full_load_current_ until Vin > Vout, thus cannot start up and reach "current multiplication" if fed from a current source than has a value less than the load current demand.

I screwed around with thinking in terms of not connecting switcher until the rectified voltage on a large storage cap exceeded the required output voltage, tested by forcing a .IC in the attached simulations. (I also tested a decreasing resistive load and observing drop-out.)

What I've concluded is the best way to do this is to rig (*) the switcher so it can't exceed 50% duty-cycle, then it can start up, but yet be able to go into regulation when the output voltage exceeds the requested voltage.

(*) I'll have to study some LTC data sheets and see how one might do that... in my custom world it'd be a piece of cake.

Of course, as Fred Abse opined, and Larkin reiterated, an autotransformer would deliver the current. But you still need to add some means of regulating to avoid over-charging the battery or zorching the bulbs.

(My behavioral model isn't a "cartoon", it's working just exactly like a real switcher would... except it's perfectly efficient... but still useful as a simulation tool. I slammed together a GVALUE source that I need to clean up, then I'll post the model. Maybe I can figure out how to add max duty cycle as a parameter :-) ...Jim Thompson

The peak to peak and RMS voltages don't correspond.

Similarly for the others. I think something's awry, there.

You appear to have an 8-pole machine. The device under discussion has 20 poles (Sturmey Archer Dynohub), but that is designed to run at wheel speed. Tire contact types run faster.

You're seeing the effect of the magnets on the B-H curve of the iron.What would be of real use would be a series of measurements of voltage and current into a resistive load (not a lamp), over a range of speeds, as well as into open circuit.

"Jim Thompson" wrote in message news: snipped-for-privacy@4ax.com...

My LED lighting set barely scratches the surface, but I'd like to run a 3W star LED "main beam" for short periods while navigating unlit cycleways.

Here a short summary of "Modellierung_eines_Nabendynamos__Mathias_Magdowski_.pdf" (without guarantee for correctness of either the information nor translation :)):

The author found the following model parameters (see page 4, for the model used) for the hub dynamo he measured

k_U = 0.0861 Vs Li = 0.1076 H R_Fe = 76.581 ohms Ri = 7.3633 ohms (corresponding equations on page 7)

by fitting the model parameters to the measured curves of open circuit voltage and short circuit current over velocity.

There are also some graphs where he validated his results on page 10 (open circuit voltage top, short circuit current bottom).

The table at the top of page 11 contains measurements under load (the load used seems to be a 6V bulb with 2.4W). The bottom table contains data calculated with the model for this load. The graph on the following page compares the measured voltages at the load with the calculated ones.

Page 13: Checks if the dynamo complies to German laws.

Page 14: Analyzes failure of head or tail lamp (based on model apparently).

Page 15: Analyzes how to obtain higher voltage from the dynamo with an external series cap. (For his simulation the author used 330uF, no real measurements.)

Page 16: Analyzes how to lower the voltage at high speeds with an external series inductor (model only).

Page 18: Analyzes an LED as load (model only).

Page 19: Summary and a scope measurement of open circuit voltage.

Hint for conversion between rpm and velocity: he used a wheel with a 2 meter perimeter (see page 4).

If someones needs assistance with the other article, I'll try to help if requested.

Adding a MaxDuty parameter does, in fact, allow it to start up from zero...

Now just jazz it up with temperature tracking the SLA charge curve and you've got all you need. And with the added benefit of longer bulb life :-) ...Jim Thompson