开云体育

The difficult bit is deriving the temperature from the applied voltage, or? the filament current. However, I think you can use a 'backwards'? trick, by defining the temperature as the initial condition. It should be well above room temperature, but not so high that the filament glows. Try 300 C. I posted some data about the tempco of tungsten a few days ago, which gives you alpha.? Using the equation R=Rcold?(1+alpha?(T?Tcold)), with R0 = room temperature resistance and To is room temperature you can get R at T = 300 C.

I put this together. Hope you need it.

?

R0 = lamp voltage rating^2 / power rating
T0 = start temp
T = temp
alpha = temp coefficient of tungsten/filament material

?

R=R0?(1+alpha?(T?T0))

开云体育

Yes, the low Q of the Wien bridge is not good for low distortion. I intend to try out that MFB circuit when I get a chance.

Perhaps you'd entertain a non-Wien Bridge oscillator: MFB_Oscillator. A working example sine wave oscillator schematic using an MFB (multiple feedback) topology. It uses no more parts than the usual Wien bridge and offers significantly better (distortion) performance.
?

--
Regards,
Tony

On 17/03/2025 20:33, Bell, Dave via groups.io wrote:

I was trying for the “classical” hardware implementation (cf. HP201C, e.g.)

?

I got a LTspice model working well, using a diode and filter feedback to adjust a behavioral resistor, but that’s cheating.

开云体育

Perhaps you'd entertain a non-Wien Bridge oscillator: MFB_Oscillator. A working example sine wave oscillator schematic using an MFB (multiple feedback) topology. It uses no more parts than the usual Wien bridge and offers significantly better (distortion) performance.
?

开云体育

That sort of error is likely to be flagged by all versions of LTspice. I suggest you examine your .ASC carefully to see if you? can see an error. I suppose you have not added a model as text in the .ASC and also called it with an ',NC or .LIB directive?

Was this from the Lamp Test.asc in the zip file?

I didn’t have any errors from that, with XVII

?

From: [email protected] <[email protected]> On Behalf Of Richard Andrews via groups.io
Sent: Monday, March 17, 2025 1:06 PM
To: [email protected]
Subject: EXTERNAL: Re: [LTspice] Incandescent lamp model

?

DS1 is the name of the lamp you are testing as per the schematic.?

?

When I try to get an expanded listing I get "Fatal Error: Multiple instances of "B:ds1:cf"".

?

Mind you, I use LTspice 4.

开云体育

开云体育

开云体育

开云体育

开云体育

开云体育

That model is probably for a gas-filled lamp, but a pilot light is probably a vacuum lamp. That could make quite a difference. I'm surprised that a 6 V 10mA lamp is still made. Do you really need low distortion? Diode clipping can get down to about 1% with careful design. A JFET as a voltage-controlled resistor (VCR) can do better, but only for a one-off, because the FETs vary too much unless you use a very costly part designed and selected as a VCR.

I’m trying to adapt and use the model in LampModel.zip, uploaded by analogspiceman last September.

?

The test circuit supplied works fine for a 120VAC lamp.

I want to adapt it to a small pilot lamp, as a stabilizer for a Wein network oscillator.

?

There are a number of parameters defined in the model, but except foe ambient temperature, not well explained.

I commented the .param block and passed them externally, and that works exactly the same, as expected.

?

Now, can anyone suggest what needs “tuning”, to get? about 600 ohms at 10mA, after warm-up?

Ka, ambient temp (Kelvin) is easy.

Kc looks like design voltage (120RMS), but doesn’t really act like it; left alone.

Kf seems to be a heat transfer coefficient, maybe good for warmup timing, but slightly changes stable resistance.

CTf is a coupling factor from Tambient to Tfilament, sets warmup timing; left alone.

RTf is filament resistance, but changing that parameter makes only a very small difference to the preset, ~150 ohms.

?

Thanks,

Dave