OPA695 current feedback amplifier

[Phil Hobbs]

That'd be 25mV into 50 Ohms. We're getting far less,
or at least that I measure.
When I'm lucky, I get 3mVpk from 1mW average. Well,
the diode is held into a reflex from a crystal or into the
leakage from a mirror. While the leak of 1mW may be 4mm in
diameter, the diode is not that big. The diode doesn't have
the bandwidth of the pulse and acts as integrator.
Then the dutycycle of the pulses is in the order of 10^-4.
And there are 2m of cable between the sensor and the
amplifier electronics.

Rene,

The responsivity of a good photodiode is on the order of 1A/W, depending on
wavelength--at 1.24 um, the energy of 1 photon is 1 eV, so 1W equals 1A of
photocurrent if the quantum efficiency is 1.0. My 1 mW pulses at 1.55 um
produce about 800 uA of peak photocurrent, because there's no antireflection
coating on the photodiode I'm using. There's definitely some buried treasure
in your setup.

Using a photodiode as an integrator for the pulse is usually a bad idea,
unless the pulse energy is small--the local carrier concentration goes so
high that all the carriers recombine before having a chance to separate, and
the response becomes very nonlinear.

A small integrating sphere is a much better method for stretching the pulse
out--you can get on the order of 50% photon efficiency if you're careful, and
the pulse will be stretched out to about 30*(sphere diameter)/c, which can
easily be several nanoseconds even with a very small sphere.

I've seen people use silicon PDs with long-wavelength lasers e.g.
CO2--running them in forward bias, and detecting the change in forward
voltage with laser power, but that's the only way I know to get sensitivities
as low as you're reporting.

Cheers,

Phil Hobbs

 

[Rene Tschaggelar]

The responsivity of a good photodiode is on the order of 1A/W, depending
on wavelength--at 1.24 um, the energy of 1 photon is 1 eV, so 1W equals
1A of photocurrent if the quantum efficiency is 1.0. My 1 mW pulses at
1.55 um produce about 800 uA of peak photocurrent, because there's no
antireflection coating on the photodiode I'm using. There's definitely
some buried treasure in your setup.

Using a photodiode as an integrator for the pulse is usually a bad idea,
unless the pulse energy is small--the local carrier concentration goes
so high that all the carriers recombine before having a chance to
separate, and the response becomes very nonlinear.

A small integrating sphere is a much better method for stretching the
pulse out--you can get on the order of 50% photon efficiency if you're
careful, and the pulse will be stretched out to about 30*(sphere
diameter)/c, which can easily be several nanoseconds even with a very
small sphere.

I've seen people use silicon PDs with long-wavelength lasers e.g.
CO2--running them in forward bias, and detecting the change in forward
voltage with laser power, but that's the only way I know to get
sensitivities as low as you're reporting.

Thanks Phil,
I'll have a look into that. It is actually my customer that mounts the
diode and supplies the pulse. Till now I kept telling them that -40dBm
was a bit too low. As said, the laser pulses come with 100MHz and 3ps or
so pulse length. That gives 10W or so per pulse from 1mW average.
The diode is always much slower than the pulse with 3ps length.
The diode is specified to have a risetime of 1ns and however fast it is,
the 2mm coax is having some losses at these frequencies too.

Yes, there are faster diodes with 50GHz bandwidth. Beside that they are
much more sensitive to higher power levels, they are a hundred times
more expensive. So cost is a factor.

Rene

 

[Yannick]

yes i agree, sory i misunderstood slightly, the resistor inevitably adds
noise so leaving it out means you have less noise,
wich is what i think i said as well anyway, but i thought you were still
meaning the higher frequency wich you mentioned in the previous paragraph. i
wasnt fully awake when i read it lol.

hehe no problem

If i were you i would measure what signal you get when u reflect it off a
target, my estimate of 10% reflected back towards the lense was pure
gueswork. the real chalenge comes when u have a smooth dark surface that is
angled away from the detector.

indeed , the good news is that my phase meausurement works, i now can
do distance measurement with direct light path on the photodiode. I am
now gonne make a microcontroller interface wich does the ADC,
calculations(phase-distance algorithm) and drives a DDS frequency
synthesiser. After this i am going to replace my PIN photodiode with
an avalanche one and the difficult part will began : measuring the
reflection on a diffuse surface... i almost cant wait

177 na at 20 mhz acros 10pf = 140uv wich sounds ok compared to the 18uv
noise for a 20mhz bandwidth from the amplifier, but this wil stil give u
quite a bit of jitter, of course this can be averaged out over many
milliseconds, but ive found it quite dificult to get as good as results as i
wld expect from this simple calculation. i have quite narow bandwidth too,
but i find most of the problem lies in noise picked up. in particular my
high voltage bias generator frequency seems to apear a lot on the signal,
despite a sheild over the rf input section and a sheild over the hv
generator and using a 7 stage multiplier so i need a lower voltage of the
200khz squarewave to the step up transformer, however you may have les
problem here as my input is tuned so is very much higher impedance.

I see, i will have to be carefull with that, are you sure this
crosstalk is from HF electromagnetic waves and not via the same power
supply?

Also i mentioned in another post one day i had it resting on my keyboard
wich is wirless and so was transmiting constant keypresses at 27mhz wich
swamped the output and had me looking for the cuase for ages. i normaly
switch off my computer/monitor flourescent lamp etc. when i try to measure
low noise performance, but hadnt considered the keyboard.

Hehe, that's frustrating, another question: how do you measure noise,
because i have a digital 60Mhz scope with high impedance probes,these
high impedance probes give me high voltage noise (4ktR) soo i cant see
my circuit noise.I suppose i need another probe (low impedance) ???

you sugested using a sinewave wich i gues would make more sense as would not
puting the thing so dam close to the detector like i did lol, i just thought
best to have a short a track as posible at 250v.

Of course the rf signal that drives the laser also apears on the signal
despite this being further away and shelded also. Although I havnt however
fuly soldered the sheilds in place yet as then it would be hard to make
changes. it is however only noticable when the gain is turned up max and the
detector totaly blanked out, also as i think i said before stray reflections
albeit invisible were also an issue at one point.

these thinks are the hard part i suppose, do you use a common mass
plane or did you split it up ?

Yannick

 

[Rene Tschaggelar]

hehe no problem




indeed , the good news is that my phase meausurement works, i now can
do distance measurement with direct light path on the photodiode. I am
now gonne make a microcontroller interface wich does the ADC,
calculations(phase-distance algorithm) and drives a DDS frequency
synthesiser. After this i am going to replace my PIN photodiode with
an avalanche one and the difficult part will began : measuring the
reflection on a diffuse surface... i almost cant wait

Better don't wait at all.
I tried this almost 10 years ago and failed due to the little light
coming back. It was rather frustrating to see how little came back
from a white paper in 20cm distance. The ultimate surface was not
meant to be white and perhaps a bit farther away than 20cm.

Rene

 

[colin]

indeed , the good news is that my phase meausurement works, i now can
do distance measurement with direct light path on the photodiode. I am
now gonne make a microcontroller interface wich does the ADC,
calculations(phase-distance algorithm) and drives a DDS frequency
synthesiser. After this i am going to replace my PIN photodiode with
an avalanche one and the difficult part will began : measuring the
reflection on a diffuse surface... i almost cant wait
good

I see, i will have to be carefull with that, are you sure this
crosstalk is from HF electromagnetic waves and not via the same power
supply?

i have gone overboard on the supply decoupling, upto 3 stages of rc/lc
filtering to each section of the circuit, and further split with seperate
regulators, i doubt its geting through those, although each suply line might
pick it up again, but its filtered again inside the sheild also.

Hehe, that's frustrating, another question: how do you measure noise,
because i have a digital 60Mhz scope with high impedance probes,these
high impedance probes give me high voltage noise (4ktR) soo i cant see
my circuit noise.I suppose i need another probe (low impedance) ???

I wouldnt worry about noise from the 10M resistor in the scope, the scope
gain wldnt nortmaly be high enough to see it anyway, and if u have a open
probe it picks up far more hum and other emi, as soon as you conect the
probe to a low impedance then its no longer 10M ohms.

these thinks are the hard part i suppose, do you use a common mass
plane or did you split it up ?

I have a simple 2 sided pcb, the backside is a gnd/shld plane and is
undisturbed.

The top side is aranged so that as much coper is left for a top gnd plane
wich encircles each of the sections. top and botom planes are joined at
strategic points, on a suck it and see basis. careful atention paid so that
any identifiabnle curent loops that include gnd are as direct as posible so
as to enclose as little area so least inductance.

Also the top gnd plane is split so the laser oscilator/driver has its own
gnd plane on the top wich is conected to the other planes by a single point
and smal ferite bead inductor. the receiver also has its own gnd plane wich
is split from the rest but also is conected to the back plane in many places
near the center inside the sheild. the sheild is conected to its own
seperate gnd too, this last change surprisingly made most improvement.

I find now i working on reducing the noise from the HV generator, there is
about 50mv of noise wich looks too jittery to be ripple, more like noise in
the control loop, its too low freq to be easily filtered out by r/c network,
might also be a noisey capacitor, ive had a few of those. also the 100k APD
series resistor means that stray light cuases a 100hz ripple on the HV
supply, im reluctant to lower this as its there to limit the curent to a
safe value, maybe i need a transistor curent limit.

Colin =^.^=

 

[Phil Hobbs]

Better don't wait at all.
I tried this almost 10 years ago and failed due to the little light
coming back. It was rather frustrating to see how little came back
from a white paper in 20cm distance. The ultimate surface was not
meant to be white and perhaps a bit farther away than 20cm.

Rene

A bit of retroreflective tape on the target is good for a 50-70 dB
(electrical) signal increase.

Cheers,

Phil Hobbs

 

[colin]

A bit of retroreflective tape on the target is good for a 50-70 dB
(electrical) signal increase.

nice, i think i might rush out and get hold of some, isnt that the stuff
that cyclists and pedestrians use to be seen more easily in the dark by car
drivers?

however my initial thought of an aplication for this was to replace a
mechanical dial gauge to measure run out on (slowly) revolving shafts or
surveying or robotic positioning feedback, i gues only on the later it might
be suitable.

Anyone got any other sugestsions of aplications?

Colin =^.^=

 

[colin]

Well, I was pleased. Most MMICS seem to run low, mid-high 30's often.
I wonder why... maybe that optimizes gain or noise figure? The 3586
can be tuned to exactly 50 ohms by fiddling the device current; it's
the only MMIC I've found that can. Add a small series RC from input to
ground, and it becomes a very good bounceless wideband match.

John

Some of the MMICs are no more than 2 transistors with a few bias resistors.
obviously at microwave frequencies the monolithic construction gets rid of
problems with lead inductance such as instability but some discrete
microwave transistors have even better noise performance, wich was the
aproach i went with to start with.

The only significant problem i have had with instability with microwave
transistors is with emiter folowers, or comon colector (such as one half of
a long tailed pair), they seem to be unstable at any reasonable colector
current, im not sure exactly why this particular configuration should be so,
i looked at a simulation with lead inductances and capacitances in circuit
and it seemed to be awefuly unstable, however i found by a bit of trial and
error that a modest resistor in series with the colector helps a lot, i gues
its like a colpits oscilator or something, with ft> 5ghz u dont need much
inductance or capacitance.

You cant see the oscilation on the scope as its far to high, it manifests
itself by strange and eratic dc offsets on parts of the circuit or even
cuases a dc ofset in the scope when its on ac coupling, obviously the input
stage is just acting as a detector at this frequency.

Colin =^.^=

 

[John Woodgate]

I read in sci.electronics.design that colin

Some of the MMICs are no more than 2 transistors with a few bias
resistors.

Quite right. The difference is that they work, unlike other similar
combinations of two hot transistors and some bias resistors.

 

[John Larkin]

I read in sci.electronics.design that colin


Quite right. The difference is that they work, unlike other similar
combinations of two hot transistors and some bias resistors.

The idea of an 8 GHz Darlington amp has always seemed improbable to
me, but there they are.

John

 

[Yannick]

i have gone overboard on the supply decoupling, upto 3 stages of rc/lc

filtering to each section of the circuit, and further split with seperate
regulators, i doubt its geting through those, although each suply line might
pick it up again, but its filtered again inside the sheild also.

ok, and do you connect mass to the shield ?

I wouldnt worry about noise from the 10M resistor in the scope, the scope
gain wldnt nortmaly be high enough to see it anyway, and if u have a open
probe it picks up far more hum and other emi, as soon as you conect the
probe to a low impedance then its no longer 10M ohms.

ofcourse, stupid from me

I have a simple 2 sided pcb, the backside is a gnd/shld plane and is
undisturbed.

The top side is aranged so that as much coper is left for a top gnd plane
wich encircles each of the sections. top and botom planes are joined at
strategic points, on a suck it and see basis.

strategic points?? is this important , i lately heard that for lower
noise they sometimes use seperate gnd planes wich they connect with
only one point with the other gnd plane, but i don't understand why
this gives lower noise, you are doing the same , can you explain this?

Also the top gnd plane is split so the laser oscilator/driver has its own
gnd plane on the top wich is conected to the other planes by a single point
and smal ferite bead inductor. the receiver also has its own gnd plane wich
is split from the rest but also is conected to the back plane in many places
near the center inside the sheild. the sheild is conected to its own
seperate gnd too, this last change surprisingly made most improvement.

interesting, even though i don't understand the reason...

Yannick

 

[Yannick]

Better don't wait at all.
I tried this almost 10 years ago and failed due to the little light
coming back. It was rather frustrating to see how little came back
from a white paper in 20cm distance. The ultimate surface was not
meant to be white and perhaps a bit farther away than 20cm.

Rene

but it has to be possible... i found range meters wich could detect
distances to 5km with 1cm accuracy , i am wondering how they do
that...

Yannick

 

[John Larkin]

but it has to be possible... i found range meters wich could detect
distances to 5km with 1cm accuracy , i am wondering how they do
that...

Yannick

Time-domain, maybe. Watts of pulsed laser, fast pd, picosecond timing,
lots of signal averaging, good optics. One advantage of pulsing is
that your transmitter (or close-in backscatter) doesn't blind your
receiver.

John

 

[colin]

but it has to be possible... i found range meters wich could detect
distances to 5km with 1cm accuracy , i am wondering how they do
that...

Yannick

i think tank laser range finders use single pulse TDR with about 20 watts of
pulsed laser power, judging from some surplus stuff i seen for sale at
display electronics, were using about 5mw, thats a big diference.

Colin =^.^=

 

[colin]

ok, and do you connect mass to the shield ?

what do you mean by mass? earth or gnd? i dont actualy have a gnd apart from
that provided by the scope earth clip, as the power suply is double
isolated, but the sheild is at what i consider 0v, i might however have gnd
at 3.5 v so the laser diode body is grounded, or have a seperate -3.5v suply
for the laser and a level shifter.

strategic points?? is this important , i lately heard that for lower
noise they sometimes use seperate gnd planes wich they connect with
only one point with the other gnd plane, but i don't understand why
this gives lower noise, you are doing the same , can you explain this?

I sometimes put the whole pcb etc on a aluminium sheet which i conect to
0v/gnd and this gives slight improvment, tbh only realy noticable when a
scope probe is left floating it picks up less 50hz humm.

power planes arnt perfect by any means, if u look at a digital gnd plane by
using a 2 contact probe 0.1" pitch and conect this to a scope you can see a
smal voltages over just 0.1" - any noisey curent you 'dump' into the power
plane makes it noisey. this voltage spreads out from the curent path acros
the entire gnd plane. if you have a seperate gnd plane conected at one point
this stops it from sprreading into the seperate plane, otherwise a sensitive
amplifier wld easily amplify this smal diference wich would inevitably apear
as a diference between diferent parts on the amplifers gnd wich would get
superimposed on the inputs unless you were very carefull.

It can need a bit of trial and error as its not always easy to spot wich
curents are going to be the most troublesome, wich is what i was jokingly
refering to as the strategic points lol.

Colin =^.^=

 

[Yannick]

what do you mean by mass? earth or gnd? i dont actualy have a gnd apart from

that provided by the scope earth clip, as the power suply is double
isolated, but the sheild is at what i consider 0v, i might however have gnd
at 3.5 v so the laser diode body is grounded, or have a seperate -3.5v suply
for the laser and a level shifter.

Yes 0V, why should u use the voltage of the laser body for shielding?
i was thinking this voltage of the shield has to be as low as possible
(relative to the used voltages) soo most EMI will not get past.
are doing the same , can you explain this?

power planes arnt perfect by any means, if u look at a digital gnd plane by
using a 2 contact probe 0.1" pitch and conect this to a scope you can see a
smal voltages over just 0.1" - any noisey curent you 'dump' into the power
plane makes it noisey. this voltage spreads out from the curent path acros
the entire gnd plane. if you have a seperate gnd plane conected at one point
this stops it from sprreading into the seperate plane, otherwise a sensitive
amplifier wld easily amplify this smal diference wich would inevitably apear
as a diference between diferent parts on the amplifers gnd wich would get
superimposed on the inputs unless you were very carefull.

Soo if i understand correctly by using only one point most of the
noise current in the first gnd plane is mostly reflected and only a
little can pass due only using one single connection, right?

Yannick

 

[Rene Tschaggelar]

but it has to be possible... i found range meters wich could detect
distances to 5km with 1cm accuracy , i am wondering how they do
that...

Yes, I've seen one of these too. It was a Q switched YAG that
was able to emit single pulses with a peak power in the order
of 20kW and a pluse length in the 200ps region, perhaps below.
I was told it was not yet covered by the law, as the law doesn't
(didn't) define single pulse laser intensities.

Rene

 

[colin]

Yes 0V, why should u use the voltage of the laser body for shielding?

The laser diode has the anode conected to its case and realy needs a
heatsink, its just a little inconvenient to have the heatskink siting at a
noisey 3.5v isntead of 0v and it would be awkward to isolate the heatsink.

i was thinking this voltage of the shield has to be as low as possible
(relative to the used voltages) soo most EMI will not get past.
are doing the same, can you explain this?

Whatever voltages you are using it doesnt matter if gnd is the lowest or
not, it doesnt efect EMI. ECL for instance uses negative voltages with
respect to gnd, but usualy everything is all referenced to the same voltage,
whatever voltage that is, and that is then usualy called gnd or 0v.

Soo if i understand correctly by using only one point most of the
noise current in the first gnd plane is mostly reflected and only a
little can pass due only using one single connection, right?

I dont think the current is reflected as such, just simply stops becuse the
conduction path ends, its curent in the gnd plane, not like EM waves in a
non conductor, but yes I gues u could think of it a bit like a harbour at
sea, a small opening alows the water inside the harbour to be much calmer.

Colin =^.^=

 

[Rene Tschaggelar]

Yes 0V, why should u use the voltage of the laser body for shielding?
i was thinking this voltage of the shield has to be as low as possible
(relative to the used voltages) soo most EMI will not get past.
are doing the same , can you explain this?

It makes sense to use the potential of the laser body as shielding
potential. Otherwise an electric conntection between the two, eg
they being mounted onto the same sheet of Alu or such, would let a
current flow.
The potential of a surface is relative to others and has no influence
on how the EMI is spread.
Isolating the laser body creates some optics precision problems plus
perhaps heat problems.
If the laser body is the positive laser terminal, the current source
is negative. Is there a problem ?

Rene

 

[Yannick]

Also the top gnd plane is split so the laser oscilator/driver has its own

gnd plane on the top wich is conected to the other planes by a single point
and smal ferite bead inductor.

this inductor is to filter out the higher frequenties soo these doesnt
come into the other gnd plane?

the receiver also has its own gnd plane wich

is split from the rest but also is conected to the back plane in many places
near the center inside the sheild. the sheild is conected to its own
seperate gnd too, this last change surprisingly made most improvement.

soo u use a 'guard ring' around your PCB wich you connect with one
single point to the gnd plane , and this guard ring you connect to the
shield, is that correct?


Also i am wondering about daylight rejection, i just bought the book
photodiode amplifiers from gerald greame, he never indicates how to do
this (or i have missed, didnt read it all in detail (only have the
book since yesterday)) but i am now doing it with a simple coupling
capactior between the diode anode and the inverting input of the
opa657 , and for keeping the reverse bias voltage on the photodiode i
first used a coil (in serie with a resistor) but it seems due
capacitive effect(and EM pickup) of this coil it is better to just use
a large resistor(200K) in serie with the photodiode, as the input
impedance of the transimpedance is much lower then this 200K almost
all current will flow into the amplifier and not into the powersupply.
This works great but i never heard anyone using this technique and no
book is advising it soo i am thinking there has to be some negative
part about it wich i dont see because otherwise everyone would use
this to reject daylight.

Yannick


Yannick