From: Rick Hellicar
Subject: Re: Transistor amplifiers v. frequency in SPICE
Date: 13 Jan 2003 11:44:57 GMT
User-Agent: tin/1.4.6-20020816 ("Aerials") (UNIX) (CYGWIN_NT-5.0/1.3.18(0.69/3/2) (i686))
In sci.electronics.design Winfield Hill wrote:
> Rick wrote...
>> Winfield Hill wrote:
>>> John wrote...
>>>> I couldn't make anything else out of it. But the OP claimed 1 V
>>>> out for 2 mV in at low frequencies. At 80 uA, re = 490 ohms, so
>>>> he's trying to get a gain of 300 out of one stage, assuming that
>>>> that the load impedance is high compared with 150 kohms, which
>>>> it won't be, of course.
>>> It'll be infinite in spice. :>) At low frequencies the expected
>>> gain could be 150k/re = 150k/312 = 480 (note, 25/0.08 = 312), so
>>> yes, 300 is reasonable, assuming an infinite transistor collector
>>> resistance (a very high Early voltage), which isn't correct, IIRC.
>> At the low frequency mentioned by the OP (100KHz) the emitter bypass
>> is much higher impedance than re and will therefore reduce the gain
>> to considerably less than 300, nearer 80 or so.
> Good point Rick. The design has so many problems, it's a
> possible candidate for our "Bad Circuits" section. :>)
It's more like a good circuit gone awry:
1. The input resistance is NOT well-defined, because beta/gm is
comparable, or even less than, than 150k | 150k; this makes the
input high-pass -3dB frequency sensitive to beta variations.
2. Re is very high, making the emitter voltage very high, thus
reducing the voltage-swing available at the collector - I'd
personally reduce the base voltage and reduce the emitter resistor.
3. The emitter sees |Xc|, which is much larger than 1/gm (re) at lowish
frequencies, so the gain is lower than the expected value of gm*Rc.