Classic Wein Bridge Oscillator Bulb Selection

[Norm Dresner]

Is there any method(ology) for selecting an incandescent bulb that will work
in this circuit or is it just a matter of using what someone else has found
works?

TIA
Norm

 

[John Popelish]

Is there any method(ology) for selecting an incandescent bulb that will work
in this circuit or is it just a matter of using what someone else has found
works?

TIA
Norm

Since the bulb has to have a significant temperature rise driven only
by the output of the oscillator, a very low power bulb is very
helpful. If you want ot generate low audio frequencies or below,, the
thermal time constant of the filament becomes important. Low voltage,
high current filaments have longer time constants that low current,
higher voltage bulbs, but this conflicts with the first requirement.
So some compromises are better than others. But many oscillator
circuits can be made to work fairly well with many different bulbs,
with a bit of tweaking.

 

[Roger Lascelles]

Lamps rated 50mA at 3V worked here OK driven from an output stage, 56R to
the feedback point and the lamp to ground. Two lamps in series was useful,
too.

I have played with 6V at 50mA and 12V at 50mA lamps too, with success. The
secret is to adjust the feedback so that the lamp is glowing very diimly; I
believe this gives a longer time constant.

Lamps are pretty crude and add lots of distortion at low frequencies. Diode
detection with FET or Light Dependant Resistor as the controlled resistance
is so much easier to control.

regards
Roger

 

[Winfield Hill]

Roger Lascelles wrote...

Norm Dresner wrote ...

Lamps rated 50mA at 3V worked here OK driven from an output stage,
56R to the feedback point and the lamp to ground. Two lamps in
series was useful, too.

I have played with 6V at 50mA and 12V at 50mA lamps too, with
success. The secret is to adjust the feedback so that the lamp
is glowing very dimly; I believe this gives a longer time constant.

Lamps are pretty crude and add lots of distortion at low frequencies.

Norm asks if there's a way to independently evaluate lamps, and yes,
of course there is. One can easily create a little test setup to
determine the lamp's small-signal resistance as a function of signal
current and frequency, measure any distortion of this resistance, and
the time-constant for each condition. Armed with this information,
one could calculate the detailed performance of different oscillator
configurations, and the effect of changing component values. If one
was very bold, one could even attempt to create a spice model. :>)

Diode detection with FET or Light Dependant Resistor as the controlled
resistance is so much easier to control.

Amen, brother Roger.

Thanks,
- Win

(email: use hill_at_rowland-dot-org for now)

 

[Jim Weir]

Take your lead from Hewlett and Packard and use a Christmas tree bulb?

Jim



"Norm Dresner" <[email protected]>
shared these priceless pearls of wisdom:

->Is there any method(ology) for selecting an incandescent bulb that will work
->in this circuit or is it just a matter of using what someone else has found
->works?
->
->TIA
-> Norm

 

[John Woodgate]

Take your lead from Hewlett and Packard and use a Christmas tree bulb?

Yes, if you use their toobs or similar. (;-)

Too much power required for small-signal transistors or op-amps.

 

[John Jardine]

Is there any method(ology) for selecting an incandescent bulb that will work
in this circuit or is it just a matter of using what someone else has found
works?

TIA
Norm

Pretty much any bulb, can with a lot of buggering about be squeezed into
doing this. Main problem using opamps etc, is that the lowish, cold/hot
filament resistance change has to padded out to with extra resistance while
still allowing the feedback resistors to just straddle the <3.000X thru

3.000X gain threshold.

If there's not much signal power available, then feedback resistors of say
2k and 1k would give a ideal gain of 3.000X. Having measured the bulb it may
be be 20ohm cold to 50ohm hot. That 1k resistor now needs reducing to (say)
970ohms and the bulb plonking in series. This gives a circuit hot/cold gain
change from 2.960X to 3.02X. Oscillation (3X) is when the bulb is 'warm' at
(critically) 30ohms. The bulb is a small difference playing amongst big
numbers and the feedback resistors need to be precise, or more usually the
oscillation point is found by tweeking a preset pot.
My GenRad bridge osc' thermistor failed. Eventually got the damned thing
working using 3X 12V 50ma bulbs in series. 5 bulbs=barely any thermal time
constant, so the osc' squegged. Two or less and not enough power and too
just too small an R to usefully use.
(Nowadays I'll use an LDR/LED combo. just uW needed to control kohms of
linear resistance change.)
regards
john

 

[Stephen Powell]

Is there any method(ology) for selecting an incandescent bulb that will work
in this circuit or is it just a matter of using what someone else has found
works?

I've written up my experience at:

http://octopus.freeyellow.com/sine.html