电灯泡稳幅的 1KHZ 文氏振荡电路

millwood

I liked the 2-opamp solution: using a separate opamp to provide the gain simplifies the design a little bit.

Here is my implementation:



U1 is the oscillator; U2 provides the gain (gain = R3 / (R4 + U3)). The AGC circuit consists of U4, Q1 and U3. The output signal is rectified by D1 and smoothed by R6/C5 (large rc constant). U4 amplifies it to drive Q1 which in turn controls the current going through the LED (of U4), which changes the equivalent resistance of the phototransistor, which in turn controls the gain of U2, thus stablizing the oscillator.

Because of the opamp / negative feedback, the circuit is quite robust vs. component values selected. For simplicity, you can replace the AGC with a (logic level) mosfet as well).

The upper pane has the output waveform (green trace). The blue trace is the output of U4 - you can see that it goes more negative, dimming the led thus lowering the gain of the oscillator.

The lower pane shows the voltage across C5 (=U4's input) - it goes up gradually.

The dual opamp set-up makes it easier to wire in a cheap opamp like ne5532, or TL0xx.

and you don't have to deal with light bulbs.

高数59

昨天朋友给我一个玻封的热敏电阻灯珠那样，我想试下。

e3po

millwood 发表于 2013-6-24 06:53
I liked the 2-opamp solution: using a separate opamp to provide the gain simplifies the design a lit ...

     I   ran the above in LTspice.

     THD is above -40dB.

     Not Good.

millwood

many factors could have driven the thd figures. one of them is the asymmetry of the photo-transistor. I think it would perform better with a photo-resistor.

it is to show how you could implement the agc. the oscillator side should be the same.

e3po

millwood 发表于 2013-6-24 13:14
many factors could have driven the thd figures. one of them is the asymmetry of the photo-transistor ...

      btw,  how would you simulate thermal copulation in LTspice ?

补充内容 (2013-6-24 17:18):
Sorry, "thermal  coupling".

millwood

"thermal copulation"? what's that?

e3po

millwood 发表于 2013-6-24 17:05
"thermal copulation"? what's that?

            well,  the heat.  " thermal  coupling",  I think I used the wrong term.

            The heat from one component will affect another component if they are very close, or stick together.


            Some literature said  thermal  coupling  can go up to KHz.


           I could not find reference on how to simulate this effect in the LTspice.


           You might know the trick, I think!

millwood

LTSpice allows you to sweep the TEMP variable. But that's global (ie. impacting every part).

To my knowledge, there is no easy way to change TEMP of a part or a select group of parts; and there is no way to simulate thermal diffusion from one part to another.

Some literature said  thermal  coupling  can go up to KHz.

I don't know about "kHz" but it is entirely possible and probably likely that high current going through a part, particularly a part with low thermal mass (like a SOT23 or TO92 part) can cause (localized) heating that in turn changes the part's parameters.

Quite similar to the microphonic in tubes or mica / ceramic capacitors.

e3po

millwood 发表于 2013-6-25 11:42
LTSpice allows you to sweep the TEMP variable. But that's global (ie. impacting every part).

To m ...

        Thank you for your input!

      
        I found something online, but not necessarily relevant :


http://www.vrg.utoronto.ca/~ngwt ... SPSD/2003/PS5P3.pdf

http://www.vishay.com/docs/73554/73554.pdf

http://www.beigebag.com/Thermal_Analysis.pdf

http://www.vishay.com.hk/docs/72794/72794.pdf

http://link.springer.com/chapter ... 2-11802-9_11#page-1

http://ieeexplore.ieee.org/xpl/l ... Farnumber%3D1283700


http://ieeexplore.ieee.org/xpl/l ... 3Farnumber%3D293302



        

e3po

tlxli 发表于 2013-5-31 16:05
我做的版本,使用小電流的彩燈．
熱時間常數大的燈能得到更低失真．

      


      您的音频界面很厉害啊。


     俺手头的音频界面到了 -100dB 附近都给底噪淹没了。

e3po

       换了个音频界面， 用了光敏电阻做简单的 AGC (跟随器带两个 白光LED 照射 光敏电阻 ).

      还是没能突破 -100dB 大关。

      也许没有用稳压电源给振荡器供电是问题根源。

      另外也许犯了个大忌， 用了陶瓷电容做退耦。


     


     

tlxli

e3po 发表于 2013-7-7 15:40
您的音频界面很厉害啊。

AW2 ,本來拿來聽的,感覺自己耳朵沒那么講究．倒是發現拿來采集信號不錯．計劃趕不上變化．

一直沒有閑工夫進行嘗試．．．我也想過用光敏電阻和燈進行增益控制,看看能不能達到更低．

e3po

tlxli 发表于 2013-7-8 06:21
AW2 ,本來拿來聽的,感覺自己耳朵沒那么講究．倒是發現拿來采集信號不錯．計劃趕不上變化．

一直沒有閑 ...

       等您的 K HZ 版本的测试结果。

epawp

早期的音讯1-甲和后来的国产XFD-6电子管低频信号发生器就是用钨丝灯泡作为稳幅元件的；
正温度系数或者负温度系数的元件都可以起到稳幅作用，就看你把它放在什么位置上。

e3po

tlxli 发表于 2013-5-31 16:05
我做的版本,使用小電流的彩燈．
熱時間常數大的燈能得到更低失真．

       实验了 OP amp +  2sc1815/2sa1015 钻石缓冲 + 灯泡,

       如果频率在300赫芝, 单个 OPAMP 就能接近 -99dB.

      把手头的运放大致试了一遍：

     OPA1611,  OPA134,  AD797, LT1115, LT1468, OPA627,  ADA4627-1, OPA277,  OP07, TL081 ...


           也不需要高级的 OPamp,  TL081 就可以了。


        如果也是 150Hz,   估计突破 -100dB 用 TL081  也可以。