Hi Vasile! Hi Jinx, I saw your 100mA constant
current laser modulator circuit. I think you CAN get
good high freq modulation from a cheap laser but it
might mean a slight re-design in a more "agricultural"
fashion. ;o)

Suggestions:
* power driver/cc MUST be very low capacitance
* power driver must be simple
* decoupling and layout like RF power stage;
 * short thick tracks
 * attention to ground planes
 * parts close together (and away from others)
 * power decoupling with RC
 * NON inductive resistors

To try and get these objectives i'd try a totally
different approach:

+24v --------R--------------------------------------
(reg?)            |        |           |
                  |        R           |
   (RC decouple!) |        |           E
                  C        |---------B    fast PNP
                  |        |           C
                  |        R           |
                  |        |           |
                  |        C           R  constant current
                  |  in- B    NPN      |   resistor
                  |        E           |
                  |        |           | +
                  |        |          LED  (laser)(C?)
                  |        |           | -
gnd  -------------------------------------------------

The small amount of current regulation you lose
by changing to just a 24v source and resistor is
minimal, and you can expect brightness/freq
illinearities in transmitter, receiver, atmostphere
etc, so a couple of percent really won't matter.
Depending on your needs the 24v could be reduced
as needed, reductions will reduce the R C(laser)
issue, even though a good laser should have quite
a low capacitance.

BUT the big gains in simplicity and very low circuit
capacitance will give good HF performance. Hopefully.
The power stage needs to be driven hard to give the
speed, so expect high base currents into both the
transistors.

Also I think real attention needs to be paid to the
power decoupling, both to provide fast current pulses
to the load and to stop it chewing up your PSU
rails, obviously using a 24v rail and high freq
decoupling will help heaps. The entire power circuit
as shown should be almost circular with the decouple
cap and PNP,R,LED in a close ring. I'm sure the RF
guys will add some gems of wisdom to my crude
ramblings. :o)
-Roman


Jinx wrote:
>
> > I've guess the schematic you've point could go at max. 50KHz.
> > Can you measure the output modulation with a PIN photodiode
> > ( or a standard photodiode if you haven't a PIN one ) and a scope?
>
> Suppose I could when I have a little time
>
> As you say, the 4093 will limit the frequency, although a 74F132
> would be OK to 50MHz, if you really need to use a gate at all on
> the bypass' base
>
> Is the BD139 too slow ? > 200MHz ? Anyway, there are plenty
> of HF power transistors around if you don't like the BD139

--
http://www.piclist.com hint: The PICList is archived three different
ways.  See http://www.piclist.com/#archives for details.