Re: OPA657 50Mhz photodiode preamplifier problems

From: Yannick (yannick_de_wit_at_pandora.be)
Date: 09/16/04 

Date: 16 Sep 2004 14:58:20 -0700

> I'm using AGC - reducing the bias voltage for the APD and also the local
> oscilator drive voltage seems to have the least impact on phase, as long as
> i dont drop below 145 V at wich point it is no longer fully depleted, wich i
> asume maens the capacitance changes dramaticaly.
>

Hmmm i ordered also an avalanche photodiode because i am now still
working with the BPX65 wich is fine for a PIN photodiode with 5pf at
20v reverse bias but has only 0.5A/W, and with the avalanche
photodiode i ordered you get about 100A/W at 150V reverse bias if i
remember correctly, i was thinking about using an oscillator (sinus)
and then just a transformator to have a voltage gain from whatever
voltage to 150V, the current will then ofcourse be reduced with the
same factor but that wont give much problems because you only have
nanoamperes to measure... is this the same way you do it to reverse
bias at these high voltages ??

>
> With path lengths of a few inches you can usualy get away with little regard
> to termination, at 20mhz anyway.

yes i know...but i want it to be perfect:) (as you see my analog
design techniques are poor therefor i also choose this project and i
try to do many thinks to up my skills but indeed they arent always
nessecairy like the impedance matching...

> If the receiving end is matched then you dont realy need to match the
> driving end anyway, as there is no signal coming toward the driver to get
> reflected.

hmmm, but if you drive with 25 Ohms, and the receiver end is 50 Ohm,
half of the signal will be reflected and this is coming back to the
receiver but then will be fully dissipated into the 25 Ohms? or does i
see it totally wrong (never had the theory of this, suppose it is in
my last year)

but anyway if you are going to drive it with 25 ohms that would
> be fine IMO, otherwise you just lose half you signal, otherwise you can put
> an extra 25 ohm resistor in series if u realy want it matched to 50 ohm at
> both ends.
>
yes ofcourse.

> Im not clear wich u are using as buffer but an emiter folower is the same as
> comon colector, a comon emiter is not realy a folower.

No ofcourse, i use common collector as buffer because of the high
input impedance and low output impedance (1/gm). But i used a common
emitter after the common collector but i now see that was stupid, i
now did it only with a common collector were i splitted the emitter
resistance in 2 resistors (560 ohm and 56 ohm) soo i have a signal
reduced effect of factor 10 wich is needed because phase detector has
-60dbm to 0dbm at 50ohm and my signal input (at the buffer) is about
1v pk-pk.

> Problem with most available jfets/mosfets is the lower gm/capacitance means
> they dont perfom so wel at high frequencies, unless you realy need the low
> dc input curent, an emiter folower (comon colector) is ideal, obviously you
> have to make sure the emiter resitor is low enough to provide enough curent
> to pull the capacitance on the output down, otherwise it may be distorted.
> if your signal is too large you can take a tap from the emiter resistor, the
> output resistance is generaly equal to the Re of the transistor wich depends
> on Ic so isnt always wise to rely on this if you need an acuratly defined
> output resistance.
>
Yes i know, Rout = 1/gm were gm = Ic/vt

> You could always clean it up with a tuned LC circuit, or indeed with another
> crystal. if your using a comon emiter then you probably got distortion
> becuase the gain is to high, or especialy if the transistor is saturating,
> or your are using a tranny with to low a Ft.
>
I see

> incidently rather than use a DDS have you considered using a few crystal
> oscilators instead if you only need 3 diferent frequenices?
>

yes i considered it but i want to do more then 3 meausurements because
then i phase errors will have less influence , the more samples i get
the better.

But for now i just stick with 1 frequency just to test...

Thanks for your time, hope i will have a simple distance measurment in
the coming month wereafter i can go up in frequency etc....

Yannick

Relevant Pages

  • Re: LED impedance
    ... It won't bother the voice stuff, but it might affect the ring voltage, ... Phone DC resistance: 33 ohms ... Some red and/or reddish orange LEDs ... A resistance of mostly about 10-20 ohms ...
    (sci.electronics.design)
  • Hysteresis needed in 741 Op-Amp circuit
    ... (determining how the change in resistance ... What that means is that if the thermistor has a resistance of 10000 ... ohms at 25°C, then at 15°C it'll have a resistance of 14,883 ohms, and ... THE VOLTAGE DIVIDER ...
    (sci.electronics.design)
  • Hysteresis needed in 741 Op-Amp circuit
    ... (determining how the change in resistance ... What that means is that if the thermistor has a resistance of 10000 ... ohms at 25°C, then at 15°C it'll have a resistance of 14,883 ohms, and ... THE VOLTAGE DIVIDER ...
    (sci.electronics.basics)
  • Re: Does heater voltage affect grid current?
    ... "With cathode bias, the grid circuit may have a resistance ... provided the heater voltage is not ... The parameter (a resistance in ohms), will be lower for fixed bias-- ...
    (alt.guitar.amps)
  • Re: Amp output Z
    ... >>>In the real amp it must be something other a pure resistance of 10 ohms ... > voltage because the amp is unstable with no load. ... He believes that the output Z is about 10 ohms ... I think the 10 ohms output resistance is a guesstimate too. ...
    (sci.electronics.design)
Quantcast