I built this motorcycle voltage regulator

http://jhau.maliwi.de/mot/voltreg.htm

quoting from the designer:

"Capacitor C1 stabilises the circuit at higher load. I have observed that the switching behaviour of the circuit without this capacitor is pretty "sharp" and stable if the load draws less than ca. 0.5 A (typically); however at higher current I observed a tendency to oscillate. This can be observed best with an oscilloscope, or a 12 V lamp as load: it will start to glow, instead of being switched completely on or completely off. C1 eliminates this instability; a value of 100 nF to 220 nF should do the job."

It does indeed oscillate, as I found, even using a mosfet instead of the darlington.
The designer's solution of putting a cap on T1 slows down the switching, so I'd like to find a better way. I guess the field switch doesn't cause the problem, because it happens whether you use a darlington or a mosfet. That points to T1. Why it would not oscillate with a small field current and then start oscillating at higher field current I'd like to understand.
Perhaps a base resistor or a tiny bit of inductance would help?

 

Is there a reason you want it to switch faster?

 

The cap slows down the circuit in more than one way. It not only puts a damper on the switching frequency; it also slows down each switching transition, which causes resistive losses in the switch.
I'm looking for some way to reduce switching frequency -- in other words damp the oscillation -- that doesn't have the side effect of substantially slowing switching speed.

 

Add a capacitors across the power leads at the circuit 10-20uF and 100nF, try a small cap across the load, at the board end, as well. What frequency are the oscillations?

 

I had to put my oscilloscope and the rest of my equipment in storage recently, so I can't do any bench work.

 

It looks like it's supposed to oscillate to me. It's a switching regulator and there's something wrong if it doesn't oscillate.

__________________
I also post at the following sites:
http://www.stop-microsoft.org http://www.heated-debates.com
Screen name: Aloone_Jonez
And http://www.silicontronics.com, same screen name as here.

 

It has hysteresis.
If you read through Joerg's description of testing the circuit, you may see.
When adjusting the voltage it should have two different voltages where it switches. At the lower of the two voltages, the field turns on. The positive feedback through the 120k resistor causes the transistor base to see a different voltage depending on whether the field switch is open or closed. At the higher of the two voltages, the field turns off. The positive feedback makes the circuit lock into that state, until the battery voltage changes enough to overcome the hysteresis. The whole point of the design is to prevent oscillation. The field can only switch as fast as the battery voltage drifts up and down through the hysteresis window. The size of the hysteresis window and the rate of change of battery voltage in response to the switching of the field determine switching frequency.
But in testing, you can move the pot as slow as you want. If you leave it alone, the circuit shouldn't switch at all. It should only switch when the voltage passes one of the hysteresis set points.
And it works for low field current, but at higher currents something causes the small transistor to oscillate.

 

Again, try adding capacitors as the higher current problem indicates a voltage regulation problem. Bikes and cars have noisy electrical circuits and once the amps increase there will be an increased voltage drop on the incoming power leads. As the voltage drops the noise will become more significant. Have you tried a crystal earpiece? this may make the oscillations audible, if it is "motorboating" then it will most likely be a power problem.