Russell McMahon wrote:

Thanks Russell for the wealth of informaton.  Allot of it is over my
head, but every little bit helps.  I'm currently using a bypass cap and
that does help.  Grounding the emitter has much the same effect, but
with a slightly lower collector voltage when idling.  Your values would
probably work well for me, but I don't have that much power available in
the "budget".

Upping the base pullup to 6M7 and lowering the collector pullup to 33K
is giving me quite a bit of voltage gain now.  Right now a 100mV p-p
signal will drive the output into clipping.  I would guess that my
voltage gain now is something close to 100.  This is probably the best
that I can get without extra components, but the rise time is still
kinda slow (>150uS to clipping), but the fall time is rather quick on
the output.  Ideally, I would like the circuit to act as a comparator or
switch with a hair trigger.

michael brown

>> I have a transistor (2n3904) that is used as a
>> buffer/amp/conditioner in my software based 1200 baud modem
>> receiver.  AFAICT, the input is in the 200mV range and I need a 5V
>> swing on the output.
>
> 5.0/0.2 = Gain of 25.
>
> Max gain possible is about 38.4 x Vcc ~= 200 so it can be done.
> (People will query that gain calculation ! :-) )
>
> The **** FIRST **** thing to try is to place an electrolytic
> capacitor from emitter to ground - say 10 uF should do. More wont
> hurt.
>
>> Right now, the
>> transistor is connected as follows.  The base is fed the DC blocked
>> audio signal.  The base is pulled up to Vcc with a 3M3 resistor
>
> This is not a legitimate way to bias silicon transistors.
> It will work after a fashion sometimes maybe.
> It's legit for Germanium.
> (People will query ... :-) ).
>
> A silicon transistor should have the bias point set explicitly. In
> this case with two R's - one each to supply and ground from base.
>
>> and has
>> a 5V zener to ground to protect it from the ringing signal voltage.
>> The
>
> probably not needed but OK
>
>> collector is tied to Vcc with a 1M resistor.  The emitter is routed
>> to ground thru a 2K resistor.
>
> Gain with emitter bypassed with cap as above would be about
>     Rc/26 * IcmA
>     Here Icmax = 5/1m = 5  uA = 0/005 mA
>     So gain = 1m/26* 0.005 = 200 = OK
>
> Try the emitter bypass capacitor !!
>
>
>> As I understand this, the transistor is held
>> in saturation by the self biasing action.
>
> if it germanium :-)
>
>> The collector "idles" at
>> about 200mV and the base at about .5 - .6V The problem is that this
>> only gives me a 2V swing at the collector with a 200mV drive applied
>> to the base.
>>
>> After some tinkering, I changed the collector resistor to 330K and
>> the emitter resistor to 1K and this gives me 4V peak at the
>> collector, but I don't really know why.  Ideally, I want to have as
>> much gain as possible to the point of massively clipping the signal
>> in order to shape the sine wave into a pseudo square wave for
>> driving the PIC pin digitally.  I don't need any real current from
>> the transistor, just some more voltage gain and faster switching.
>>
>> Can someone recommend something using the same 2n3904 that will give
>> me considerably more gain and consequently a steeper rising edge to
>> the PIC.  I'm using a 16F628, but don't have enough pins available
>> to use the internal comparator to do the job.  I'd also like to
>> accomplish this with just discrete components and no extra ICs.
>
> You can get a truly lovely result with 2 transistors in a long tailed
> pair (Google ...) but letssee if we can make the transistor worlk OK
> (although it may be already with that cap.)
>
> Set Vc at 1/2 supply quiescent.
> Aim at say 1 mA Ic to start
> Rc = (Vcc-Vcquiesc)/Ic = (5-2.5)/.001 = 2500 ohms.
>
> We want a smll emitter resistor to allow base to float away from
> ground clamped position.
> Say Re = 100r gives 100/2500 * 2.5 = 0.1 volt dc at 1mA so OK.
> Place 10 uF+ across Re
>
> Ve = 0.1v so Vb = 0.6 + 0.1 = 0.7v or so.
> V Rb-+ = 4.3v
> V Rb-gnd = 0.7v
> ratio of Rs about 6:1
> 10K:1.5K
> 12K:2k2 etc using R12s
>
> 10:1.5K
> Vb = 1.5/11.5*5 = 0.65 too low
>
> 12k:2k2
> Vb = 2.2/14.2 x 5 = 0.767v = OK
>
> Try
> R C-V+ = 2500r
> R B-V+ = 12K
> R B-gnd = 2k2
> R E-gnd = 100r
> C E-gnd = 10 uF
> C coupling in = whatever (10 or 100 uF?)
>
> Required Ib = Ic/beta = 1 mA/100 say = 10 uA
> I in 12K Rb-V+ is about 4.2/12k = 300 uA so plenty for base.
>
> Howzat ?
>
>
>         RM
>
>
>
>
>
>
> Try Rb-Vcc = 10k
> Rb-gnd = 2k7
> Vb = 2.7/12.7 * 5 =
>
>
>
>
>
>
>
>
>>
>> michael brown
>>
>> --
>> http://www.piclist.com#nomail Going offline? Don't AutoReply us!
>> email listserv@mitvma.mit.edu with SET PICList DIGEST in the body

--
http://www.piclist.com#nomail Going offline? Don't AutoReply us!
email listserv@mitvma.mit.edu with SET PICList DIGEST in the body