DDS Selection

[spflanze]

I am designing designing a system that uses changes in the
transmissivity light in a detection chemistry to detect gas. The light
source is a sine wave modulated LED source. After the light passes
through the detection chemistry it will be measured by a photodiode
and a transimpedance amplifier circuit. The output of the
transimpedance amplifier is measured by a 16 bit AD7606 ADC which does
a DMA data transfer into a Blackfin DSP where a Fourier Transform is
done to extract an amplitude at the LED's modulation frequency.

There will be six detection channels. Each channel is to be modulated
at a different frequency to reduce crosstalk. The length of time the
Fourier Transform is done over will be in multiples of 100ms for
maximum rejection of both 50Hz and 60Hz. The modulation frequencies
are chosen in multiples of the inverse of this length of time for
maximum crosstalk rejection.

Between the output of the transimpedance amplifier and the ADC is an
anti-aliasing filter. Its cutoff frequency and number of polls are
chosen so the attenuation at the ADC's sampling frequency is equal or
less than the ADC's LSB divided by the ADC's total number of states
(2^16).

The system needs an option for battery power so current draw is an
issue.

I need to choose a DDS chip to generate the sine wave reference for
the LED's modulation. I have been looking at 14 Bit DDS chips at:
http://www.analog.com/ps/psthandler.aspx?pstid=10068&la=en
I notice there is a dramatic increase in power requirements the more
bits the DAC part of the DDS has. I would like to choose a 10 bit one
such as AD9838 that uses little power but I am concerned about the
quantization noise that appears as a spurs in the DDS output beginning
at twice the output frequency. That spur would be within the anti-
alias filter's passband. Theoretically this spur would appear right in
one of the notches of the sync function that is the frequency response
of a Fourier Transform and so would be taken out by it. In practice
can I actually count on it doing that? The DDS has the same frequency
reference as the ADC's sample rate so the spur would be precisely
located there.

Or do I need a reconstruction filter for the DDS that would have a cut
off between the modulation frequency and the first spur at twice this
frequency? I am sure it would need many poles. If so could the anti-
alias filter, although it is at the output of the transimpedance
amplifier instead of the DDS, double as a reconstruction filter if it
would have the same cutoff frequency and poles?

[John Larkin]

Why not use a synchronous detector (lockin) architecture?

John

[miso]

>
> Or do I need a reconstruction filter for the DDS that would have a cut
> off between the modulation frequency and the first spur at twice this
> frequency? I am sure it would need many poles. If so could the anti-
> alias filter, although it is at the output of the transimpedance
> amplifier instead of the DDS, double as a reconstruction filter if it
> would have the same cutoff frequency and poles?

If we did a mental exercise and reduced the bits in the DDS, eventually
it would be a square wave. So wouldn't the first noise appear at 3x the
modulation frequency if the modulation frequency and clock rate have an
integer relationship. If so, then your post filter is substantially
easier to build.

[Jamie]

All that just to modulate a LED with sine wave output? Sounds a bit
over kill to me.

One item on the list you need to consider, the LED is not linear in
its transition to current verses emissions. The first item on the list
should be a driving circuit using a photo feed back to ensure you have a
linear representation of photons emitting through your chemistry.

Generating the SINE wave in short, can be done with out the use of DDS
technology, which just adds to the pile of over kill.

Using an analog method of generating that sine wave with a
synchronous signal for the receiver would be more to what I would expect
to use.

Jamie

[Vladimir Vassilevsky]

spflanze wrote:
> I am designing designing a system that uses changes in the
> transmissivity light in a detection chemistry to detect gas.

You designing a homework.

[Bill Sloman]

Probably not. A student following a sensibly designed course wouldn't
be thinking of using DDS to get a sine wave, nor of applying a fourier
transform to the data collected to find out the amplitude of the sine
wave getting through.

This is almost certainly a graduate student who has done some reading,
but of the wrong kinds of books. They need to read Horowitz and Hill's
"The art of electronics" which was designed as an electronics text for
bright physics students (at Harvard) and covers useful subjects - like
lock-in and phase sensitive detection - in some detail.

God only knows why spflanze thinks you need to use sine-wave
modulation on a signal being used to measure an extinction coefficient
relatively slowly.

We could usefully know more about what is actually being measured -
what are the six separate detection channels actually doing?

--
Bill Sloman, Nijmegen

[John Larkin]

Right. Then instead of doing an FFT, just synchronously demodulate
(digitally, if you want) and lowpass filter or boxcar average. You
don't need a DSP source or a Blackfin for that.

John

[Phil Hobbs]

Nah, LEDs are pretty linear above about 10 uA. I second the suggestion
to use square wave drive and a lock-in per channel, i.e. an AM radio
with synchronous detection. (AM radios are pretty good at adjacent
channel rejection, for such simple devices.) For instance, the OP can
pick drive frequencies of, say, 12, 13, 14, 15, 16, and 17 kHz, with 100
Hz-ish lowpass filters after the demodulator. One wants the low order
harmonics of each modulation frequency to miss each other, because that
helps a lot with the spurious signal reduction. If he doesn't want to
lose all that nice software work, he can write a program to find the
optimum channel assignments that minimize crosstalk.

Square wave lock-ins are sensitive to odd harmonics, but a reasonably
good lowpass filter after the TIA will get rid of those, as well as
noise contributions from their neighbourhoods.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510
845-480-2058

hobbs at electrooptical dot net
http://electrooptical.net

[miso]

Why do a simple analog solution when you can throw a computer at the
problem. ;-)

I'm probably not following the need for a lock-in since he has all the
oscillator sources at hand. You don't have to lock, just demodulate the
photo signal using the modulation source. Quite possible I'm missing
something here.

If the modulation circuit also produced a quadrature clock source, then
the photo signal could be IQ demodulated. One path would be the signal
level, and I suppose the other path would be a figure of merit of the
noise. [All assuming no significant phase shift in the photo signal.]

[John Larkin]

"Lock-in amplifier" refers to syncronous demodulation followed by
low-pass filtering.

http://en.wikipedia.org/wiki/Lock-in_amplifier

John

[Bill Sloman]

On Jan 23, 2:30 am, John Larkin
<jjlar...@highNOTlandTHIStechnologyPART.com> wrote:

> "Lock-in amplifier" refers to synchronous demodulation followed by
> low-pass filtering.
>
> http://en.wikipedia.org/wiki/Lock-in_amplifier

Strictly speaking, you've got to have a modulation on the signal being
detected before you can lock-in onto that modulation and synchronously
detect it.

Normally you excite the system under test with some sort of
alternating signal - sine wave are often handy, square wave are easier
to make, but repeated narrow pulses at regular intervals can be more
appropriate for some problems.

--
Bill Sloman, Nijmegen

[miso]

>>
>> "Lock-in amplifier" refers to synchronous demodulation followed by
>> low-pass filtering.
>>
>> http://en.wikipedia.org/wiki/Lock-in_amplifier
>
> Strictly speaking, you've got to have a modulation on the signal being
> detected before you can lock-in onto that modulation and synchronously
> detect it.
>
> Normally you excite the system under test with some sort of
> alternating signal - sine wave are often handy, square wave are easier
> to make, but repeated narrow pulses at regular intervals can be more
> appropriate for some problems.
>
> --
> Bill Sloman, Nijmegen

Reading the description, lock-in sounds like more than just sync demod
plus LPF. But we are in agreement that is what would be done to recover
the signal. Generating I and Q with the modulation source would make
this task relatively simple, plus yield a figure of merit signal.

[miso]

> http://www.boselec.com/products/siglimwhat.html

This is actually a better description. There is really nothing to lock.
You have the source already, so just feed it to the demod.

[Phil Hobbs]

Lock-ins don't phase-lock to the input--otherwise you couldn't measure
the static phase. You supply them a reference, which in this simple
situation is the same as the drive signal, phase shifted to match the
pre-detection filter (e.g. the RF filter in an AM radio).
Alternatively, since it's only the amplitude he cares about, he could
phase-lock to the detected signal. That can get exciting in a
multichannel setup, though.

[miso]

> Lock-ins don't phase-lock to the input--otherwise you couldn't measure
> the static phase. You supply them a reference, which in this simple
> situation is the same as the drive signal, phase shifted to match the
> pre-detection filter (e.g. the RF filter in an AM radio).
> Alternatively, since it's only the amplitude he cares about, he could
> phase-lock to the detected signal. That can get exciting in a
> multichannel setup, though.
>
> Cheers
>
> Phil Hobbs
>
>

But you don't really care about the static phase shift, other than if it
is significant, then there is an error in the amplitude if the demod is
not quadrature.

I think we're in agreement. You have the reference and demod the photo
signal, beat down to DC. With quadrature demod, you could be phase
independent presuming the I and Q are demodulated.

[Phil Hobbs]

The point about the static phase is to distinguish between a lock-in and
a PLL, which you seem to have been conflating.

And in the present case, you don't want to measure the phase, but you do
care about it--you want it to be zero, ideally in an architecture where
the temperature drift cancels to leading order.

You don't want to use I and Q, because if the phase is adjusted right, Q
contains precisely zero information, and if you let the phase be just
anything, you have to reconstruct the amplitude from I and Q, which
costs hardware and extra engineering time, and gets you nothing
whatever. It also puts a serious premium on the linearity of the phase
detector, and on the accuracy of the waveform. Switching at the
zero-crossings is about the most accurate way to use CMOS transmission
gates for this job.

On a custom chip, I'd want to use a PLL, because the amplitude channel
is insensitive to jitter, and making performance independent of normal
process variations in RC products and GBW would improve yield. You'd
need a Q channel for that, of course.

In a built-up circuit, though, complexity costs money, and accurate RCs
are cheap, so I'd make the TIA wideband, and use a filter architecture
that has zero phase shift at the desired operating frequency, such as
cascaded 2-pole HP and LP filters.

[George Herold]

Fun discussion,
Sorry, but I can't help you with the DDS selection.
I'd fourth the lockin/ demod approach.
Are your six LED's all different wavlenghts?
(and if not why six?)

To the group... is there some point where it is advantageous to use
the FFT technique.
(what if he had more sources)
(I'm thinking of FIR spectroscopy and the Fellgett advantage.)
At least with one detection channel you don't have to worry about
channel to channel variation.

George H.

[Phil Hobbs]

At some point making all those separate filter channels and PSDs becomes
more of a pain than doing it all digitally. A windowed FFT gives you a
whole bunch of identically-shaped filter channels centred on each
frequency sample, which is a very nice feature.

[spflanze]

On Jan 21, 11:09 pm, miso <m...@sushi.com> wrote:
> If we did a mental exercise and reduced the bits in the DDS, eventually
> it would be a square wave. So wouldn't the first noise appear at 3x the
> modulation frequency if the modulation frequency and clock rate have an
> integer relationship. If so, then your post filter is substantially
> easier to build.

That would be true if it were a perfect DAC with infinite bandwidth.
But there is a finite bandwidth and finite slew rate so its output is
other than a succession of perfect squares. Triangular elements are
involved and those have even harmonics.

Also there are other spurs that can be present the time length of each
square will vary slightly depending on the lower order tuning bits.
The variations happen when the bits that are discarded between the
phase accumulator and the phase to amplitude converter vary. And that
happens when any of the corresponding bits in the tuning word are set.

http://www.edn.com/article/478438-DDS_design.php

[spflanze]

On Jan 22, 9:55 am, John Larkin
<jjlar...@highNOTlandTHIStechnologyPART.com> wrote:

An FFT is a digital demodulation with a box car average. An FFT
multiplies by a sine wave and extracts the 0 Hz difference frequency
by averaging.

[John Larkin]

Sure. But it's a lot of work, and relies on the source and DDS time
references to be equal. And there's the I/Q thing to deal with. And
the complexity.

John

[spflanze]

On Jan 22, 7:01 am, Jamie

An analog feedback control circuit for driving the LED's using
feedback from a monitoring photodiode is being considered. This is not
just for linearity, but to keep the light output constant as the LED
ages and also for noise reduction. The light output of an LED for the
same current declines with age. I suspect also there is noise other
than shot noise and noise in the driven current. I suspect something
similar to 1/F noise in opamps from the LED alone that such a current
control circuit would counter. I have never seen such a noise
specification from an LED manufacturer. Attempts to get this from one
went nowhere. The LED manufacturers make them for lighting and
indicators. They neglect those of us who use them for optrodes.

I left the current control out of my initial description because it
was not essential to the concern I started this thread about.

[spflanze]

I was a bit careless in that first sentence. But me design no
homework. Me design for company chemical warfare agent detector. :)

[George Herold]

> was not essential to the concern I started this thread about.- Hide quoted text -
>
> - Show quoted text -

Unless you really over drive them LEDs are pretty much just shot
noise.. at least the few I've looked at. You have to drive them
from a low noise source. (of course)

George H.

[spflanze]

We are looking for the smallest possible changes in transmissivity. So
I am looking to do state of the art in noise performance.

Modulation is necessary to take out offset error drift in the
electronics. Drift is really the lower end of the 1/F noise spectrum.
I will choose a modulation frequency right where the 1/F noise
spectrum gets buried in other noise. beyond this there is a point
where noise starts increasing due to the zero made with capacitance
around the TIA's virtual ground and the TIA's feedback resistor. I am
doing sinewave modulation to keep the carrier frequency components
beneath that zero. It is simpler if there is only one frequency
component.

If both the modulating and demodulating waveforms are square wave
there will be contributions to the output at odd harmonics where the
signal to noise ratios of the system is less. The square wave is the
equivalent of doing both a carrier signal and Fourier Transform at the
fundamental and again at all odd harmonics and summing the results.

The odd harmonics can be filtered. A filter with enough attenuation at
the first of the harmonics and very little attenuation at the carrier
frequency would require many poles, and would require many op amps.

A sine wave can be generated by op amp circuits. But because for noise
reduction I am looking to extract the carrier in as narrow band a
filter as possible it means extracting it with an FFT. So the analog
circuit generator has to be precise in frequency. That means phase
locking its output which requires many parts and board area. A DDS can
do it all in one chip.

The response time of this system is crucial. The box car averaging in
the FFT offers the best compromises between sensitivity (determined by
noise) and response time. The more time the FFT is done over the
narrower its bandwidth.

The decision I need to make is whether I need 10, 12, or 14 bit DDS.

[spflanze]

Not all the LEDs will be the same wavelength. Each will illuminate a
different chemistry. For those LEDs of the same wavelength it is too
difficult to split the beam to where the light needs to go.

Also it is known there will be cross talk. How much isn't known yet.

[Bill Sloman]

> An analog feedback control circuit for driving the LED's using
> feedback from a monitoring photodiode is being considered. This is not
> just for linearity, but to keep the light output constant as the LED
> ages and also for noise reduction. The light output of an LED for the
> same current declines with age. I suspect also there is noise other
> than shot noise and noise in the driven current.

LED's are photodiodes - if not very good one's - which may complicate
life a bit

> I suspect something
> similar to 1/F noise in opamps from the LED alone that such a current
> control circuit would counter.

If the LED output is decaying during it's operational life, then the
decay mechanism (whatever it is) is going to show up as a sort of 1/f
noise. If the light output that you monitor doesn't come from exactly
the saem part of the LED as the light output you are using to measure
transmission, the feedback control of brightness may not cancel out
this kind of 1/f noise. The decay mechanism could be localised in the
position or the direction of the element of the light source that's
changing, so you might not be controlling exactly what you want to
control.

> I have never seen such a noise
> specification from an LED manufacturer. Attempts to get this from one
> went nowhere. The LED manufacturers make them for lighting and
> indicators. They neglect those of us who use them for optrodes.

Because you don't form a significant portion of the market. You might
do better if you bought laser diodes, but they are a lot more
expensive.

> I left the current control out of my initial description because it
> was not essential to the concern I started this thread about.

--
Bill Sloman, Nijmegen

[Bitrex]

A careless designer building a chemical warfare agent detector. That's
comforting.

[Bill Sloman]

On Jan 23, 6:20 pm, spflanze <art...@wavenet.org> wrote:
> On Jan 21, 11:09 pm, miso <m...@sushi.com> wrote:
>
> > If we did a mental exercise and reduced the bits in the DDS, eventually
> > it would be a square wave. So wouldn't the first noise appear at 3x the
> > modulation frequency if the modulation frequency and clock rate have an
> > integer relationship. If so, then your post filter is substantially
> > easier to build.
>
> That would be true if it were a perfect DAC with infinite bandwidth.
> But there is a finite bandwidth and finite slew rate so its output is
> other than a succession of perfect squares. Triangular elements are
> involved and those have even harmonics.

A triangular wave is just the integral of a square wave, and has the
same - odd only - harmonics, but with the amplitude dropoing in
proportion to the square of the harmonic number.

But yes, the output from a DDS is potentially messier than the output
from simpler switching circuit.

Have you thought about a delay-line filter? You clock your square wave
through a shift register, hook up a suitable resistor to each stage,
and sum the outputs. IIRR a sinc function tapered with a Hamming
window to kills the Gibbs oscillation can give you a very clean sine
wave; you've got residual high-frequency harmonics, but the resistors
can be 0.1% parts (with a bit of padding to get the exact values) and
the lower harmonics can be held to better than 60dB down.

<snip>

--
Bill Sloman, Nijmegen

[Phil Hobbs]

Well, you have to live with the results, so do what you're comfortable
with. The lock-in will blow the doors off the digital approach for
noise.

[Mikko OH2HVJ]

spflanze <art...@wavenet.org> writes:

> We are looking for the smallest possible changes in transmissivity. So
> I am looking to do state of the art in noise performance.

What is your target on transmissivity change and available light power ?

Can you take part of the light before the chemistry and use that as
the reference for the intensity ? This with lock-in detection (done
in HW or SW) has given good results for us in similar measurements.

--
Mikko OH2HVJ

[tm]

"spflanze" <art...@wavenet.org> wrote in message
news:e2aab768-3c58-4381...@t13g2000yqg.googlegroups.com...

> If both the modulating and demodulating waveforms are square wave
> there will be contributions to the output at odd harmonics where the
> signal to noise ratios of the system is less. The square wave is the
> equivalent of doing both a carrier signal and Fourier Transform at the
> fundamental and again at all odd harmonics and summing the results.
>
> The odd harmonics can be filtered. A filter with enough attenuation at
> the first of the harmonics and very little attenuation at the carrier
> frequency would require many poles, and would require many op amps.

Use a notch (or several as needed) as well as the LPF to get rid of the
harmonics.

[George Herold]

> Also it is known there will be cross talk. How much isn't known yet.- Hide quoted text -
>
> - Show quoted text -

Cool, sounds like a fun project. One thing that stinks about LED's is
you don't get a lot of light out of them.

George H.

[lang...@fonz.dk]

guess most of devices like that are mostly for security theater
anyways,
so it only has to work enough to convince those who write the checks
in DHS/TSA

-Lasse

[Fred Bartoli]

George Herold a écrit :

Don't know for noise spec but I once encountered a low frequency issue
which was traced down to convection combining with the more often spec'd
LED output efficiency tempco. (the LEDs were driven withing their rated
specs)


--
Thanks,
Fred.

[George Herold]

So how 'good' can you do Mikko? I've looked at these really tiny
changes in our optical pumping apparatus,
(a straight DC light level measurement, no modulation or lockin) and
seen changes at smaller than a part in 10^5. (That was averaging
several sweeps though.. Signal to noise in one sweep is about 1 at
changes of a part in 10^5.)

George H.

[spflanze]

On Jan 23, 9:36 am, John Larkin

It is not a complex thing for the ADC and the DDS to have the same
external clock reference. If both I and Q are required because there
is enough phase shift in the electronics, then implementing it is just
a matter of software.

[Bill Sloman]

Notches aren't a good idea inside a lock-in amplifier set-up. Too much
phase shift.

You start needing to add all-pass networks to your filter networks to
keep the phase-shift low and stable.

--
Bill Sloman, Nijmegen

[John Larkin]

Why not use square wave modulation and a receive TIA with fairly wide
bandwidth, take lots of ADC samples, and do the lockin thing
digitally? That will dither the ADC codes into oblivion. TIA phase
shift will be small enough to ignore.

John

[spflanze]

The photodiode would look at the LED sideways, which would be outside
its lens's cone, but would still have plenty of light due to its
proximity. After Bill Sloman's response I am not sure about noise
suppression. I would expect a photodiode in such an arrangement to get
more light from some areas of the light emitting surface than others.
I have some lab experience where the feedback seemed to suppress noise
on the output of a Luxeon III. I do not know where the monitoring
photodiode was relative to this LED because both were in a sealed
unit. The two LEDs being considered now are:

http://partfinder.avagotech.com/Avago/Avago.jsp#!partSearch=&partno=asmt-fj10-adh00
http://www.cree.com/products/pdf/LEDlamps/C503B%281079%29.pdf

There is not yet a specific target.

[John Larkin]

LEDs have rotten optical-output tempcos.

John

[spflanze]

It is the sawtooth that has the even harmonics. You are right that the
triangular wave can be considered the result of a square wave with its
higher harmonics being attenuated by an integrating filter.

I can see how an array of resistor values on a shift register can be
chosen to replicate the impulse response of a filter. I do not follow
how a sync function can be implemented with this. A reference would be
appreciated.

What advantage would a delay line filter have over a Sallen-Key filter
with many poles?

[George Herold]

> Bill Sloman, Nijmegen- Hide quoted text -
>
> - Show quoted text -

Nah, they're fine. You don't want to make the Q too high, but a nice
Q of ten can really help with the dynamic range. You, of course,
have to adjust everything, center the filter and adjust the phase.
It's tough to get down to the noise level without a bandpass somewhere
in the system.

George H.

[George Herold]

On Jan 23, 6:12 pm, John Larkin

> John- Hide quoted text -
>
> - Show quoted text -

Yeah seems like the same equipment can do the lockin thing and the
FFT.. it'd be interesting to see the results. I did this lock-in vs
amplitude detection with similar time constants and all. (to OP, the
great thing about lockins is you can get a really long time constant...
(narrow linewidth)) The lockin was only about square root 2 better.
The phase part of lockin.

George H.

[George Herold]

> http://partfinder.avagotech.com/Avago/Avago.jsp#!partSearch=&partno=a...http://www.cree.com/products/pdf/LEDlamps/C503B%281079%29.pdf

>
> There is not yet a specific target.

Dang optical units. The Avago part has the following unit's

50mA currrent ~2V Vf
600nm wavlength, 8 degree view angle.
and the 'winners'
luminous efficacy 355 lm/W
luminous efficiency 27 lm/W.

So if I stuck a big photodiode in front of this at 50mA how many watts
or amps of current will I measure?

George H.

[George Herold]

On Jan 23, 6:13 pm, John Larkin

> John- Hide quoted text -
>

> - Show quoted text -

Yeah for a good source ya gotta control the temp too... but that's
true most optical sources to some degree* or other.

George H.
*NPI

[spflanze]

Is that a rotten tempco for wavelength or output power or both?

[John Larkin]

Or use a Hobsonian Laser Noise Canceler.

John

[spflanze]

I do not need any frequencies about the fundamental or carrier. Why
put a notch at a harmonic when I could use a low pass to take out all
the harmonics?

[spflanze]

On Jan 23, 5:05 pm, George Herold <gher...@teachspin.com> wrote:
> On Jan 23, 6:05 pm, spflanze <art...@wavenet.org> wrote:
>
>
>
>
>
>
>
>
>
> > On Jan 23, 11:59 am, Mikko OH2HVJ <oh2...@sral.fi> wrote:
>
> > > spflanze <art...@wavenet.org> writes:
> > > > We are looking for the smallest possible changes in transmissivity. So
> > > > I am looking to do state of the art in noise performance.
>
> > > What is your target on transmissivity change and available light power ?
>
> > > Can you take part of the light before the chemistry and use that as
> > > the reference for the intensity ? This with lock-in detection (done
> > > in HW or SW) has given good results for us in similar measurements.
>
> > > --
> > > Mikko OH2HVJ
>
> > The photodiode would look at the LED sideways, which would be outside
> > its lens's cone, but would still have plenty of light due to its
> > proximity. After Bill Sloman's response I am not sure about noise
> > suppression. I would expect a photodiode in such an arrangement to get
> > more light from some areas of the light emitting surface than others.
> > I have some lab experience where the feedback seemed to suppress noise
> > on the output of a Luxeon III. I do not know where the monitoring
> > photodiode was relative to this LED because both were in a sealed
> > unit. The two LEDs being considered now are:
>
> >http://partfinder.avagotech.com/Avago/Avago.jsp#!partSearch=&partno=a...
>

> > There is not yet a specific target.
>
> Dang optical units.  The Avago part has the following unit's
>
> 50mA currrent ~2V Vf
> 600nm wavlength, 8 degree view angle.
> and the 'winners'
> luminous efficacy 355 lm/W
> luminous efficiency 27 lm/W.
>
> So if I stuck a big photodiode in front of this at 50mA how many watts
> or amps of current will I measure?
>
> George H.

The monitoring photodiode is looking at the LED sideways, not in
front. The detection photodiode has the LED aimed at it but the
detection chemistry is between it and the LED.

[George Herold]

On Jan 23, 8:24 pm, John Larkin

<jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
> On Mon, 23 Jan 2012 17:08:14 -0800 (PST), George Herold
>
>
>
>
>

Oh I was thinking about the wavelength.

You can measure the output power lotsa ways.

George H.

[George Herold]

> the harmonics?- Hide quoted text -

>
> - Show quoted text -

Oh I'm sorry perhaps I misread the post. (Sorry Bill)
I was thinking of a band-pass filter not a notch.

George H.

[miso]

I only mention IQ is that the "null" signal is a figure of merit. If it
is quiet, you have a quality signal. Quadrature isn't required, but useful.

[George Herold]

On Jan 23, 8:29 pm, spflanze <art...@wavenet.org> wrote:

> the harmonics?- Hide quoted text -

>
> - Show quoted text -

Opps sorry Bill, I misread your post. I was refering to a band-pass
filter centered one the fundamental.

George H.

[John Fields]

On Mon, 23 Jan 2012 17:32:56 -0800 (PST), spflanze

---
With the exception of Dr. Hobbs, you can expect everyone else's reply
to be self-serving, so caveat emptor.

--
JF

[George Herold]

> Is that a rotten tempco for wavelength or output power or both?- Hide quoted text -

>
> - Show quoted text -

More wavelength than output power. (but they don't have similar
units)
I don't know about the leds you posted. You've got to measure them.

George H.

[George Herold]

> JF- Hide quoted text -

>
> - Show quoted text -

You're absolutely right John. I was hoping Dr. Hobbs was logged on
and would give me a 'real' number. Or at least there would be a reply
by the time I logged on in the morning. lumens and candela's are the
worst!

George H.

[lang...@fonz.dk]

On 24 Jan., 00:12, John Larkin

I guess you could say an FFT is similar to lockin, is just a mix with
sin/cos and an average, just done at "all" frequencies at the same
time

-Lasse

[John Larkin]

Both. You can compensate for the amplitude drift with a dual-path
system. What sort of wavelength stability matters to you?

John

[John Larkin]

Be wary of those intensity-angle curves. The data sheets always draw
them symmetric and perfectly centered on 0 degrees. In real life, they
tend to be all over the place.

John

[Phil Hobbs]

Fred Bartoli wrote:
>
> George Herold a écrit :
> > On Jan 23, 12:37 pm, spflanze <art...@wavenet.org> wrote:
> >> On Jan 22, 7:01 am, Jamie

> >> An analog feedback control circuit for driving the LED's using
> >> feedback from a monitoring photodiode is being considered. This is not
> >> just for linearity, but to keep the light output constant as the LED
> >> ages and also for noise reduction. The light output of an LED for the
> >> same current declines with age. I suspect also there is noise other
> >> than shot noise and noise in the driven current. I suspect something
> >> similar to 1/F noise in opamps from the LED alone that such a current
> >> control circuit would counter. I have never seen such a noise
> >> specification from an LED manufacturer. Attempts to get this from one
> >> went nowhere. The LED manufacturers make them for lighting and
> >> indicators. They neglect those of us who use them for optrodes.
> >>
> >> I left the current control out of my initial description because it
> >> was not essential to the concern I started this thread about.- Hide quoted text -

> >>
> >> - Show quoted text -
> >

> > Unless you really over drive them LEDs are pretty much just shot
> > noise.. at least the few I've looked at. You have to drive them
> > from a low noise source. (of course)
> >
>
> Don't know for noise spec but I once encountered a low frequency issue
> which was traced down to convection combining with the more often spec'd
> LED output efficiency tempco. (the LEDs were driven withing their rated
> specs)
>
> --
> Thanks,
> Fred.

Drift has a characteristic 1/f**2 dependence, so it usually dominates at
very low frequency.

Cheers

Phil Hobbs
--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510
845-480-2058

hobbs at electrooptical dot net
http://electrooptical.net

[Phil Hobbs]

Actually notches are great inside feedback loops, because their phase
shifts go away at frequencies more than a few notch widths away. Long
long ago, when I was designing atomic force microscopes, we got a factor
of about 4 increase in speed by notching out the mechanical resonance of
the silicon cantilever.

If you have a resonant actuator, a notch filter can be a _big_ win.

[Phil Hobbs]

You can do that sometimes, especially when the noise is nearly white.
Photodiode capacitance makes the noise of TIAs rise steeply with
frequency, though, so once you get past the point where the
differentiated TIA voltage noise e_N*(2*pi*f*C_diode) dominates the
shot noise, you start losing SNR by increasing BW.

Cheers

Phil "CW guy at heart" Hobbs

[spflanze]

On Jan 23, 5:24 pm, John Larkin

<jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
> On Mon, 23 Jan 2012 17:08:14 -0800 (PST), George Herold
>
>
>
>
>
>
>
>
>

What is a Hobsonian Laser Noise Canceler? Is it a Phil Hobbs invention?

[Phil Hobbs]

http://electrooptical.net/#cancellers

Cheers

Phil Hobbs

[spflanze]

This is not necessarily true. FFT can be done at a single frequency.
If the number of samples per cycle is a power of two there is a
symmetry in that there are repeated values of the sine function to
multiply the sample by. They repeat symmetrically about the 1/4 point
in the first 1/2 of the waveform, and again at the 3/4 point in the
second half of the waveform. Also the absolute values of the first 1/2
of the waveform are the same as in the second half of the waveform.
This enables the multiplications to be factored out of all the
summations, greatly reducing the number of multiply operations. This
is what I intend to do at the carrier frequency of each channel. I am
not interested in the other frequencies.

Depending on the sample rate the repeat values may be spread out over
two or more periods. The more periods the less efficient the FFT but
also the lower and more random the quantization noise.

[spflanze]

On Jan 23, 11:51 am, Phil Hobbs

> --
> Dr Philip C D Hobbs
> Principal Consultant
> ElectroOptical Innovations LLC
> Optics, Electro-optics, Photonics, Analog Electronics
>
> 160 North State Road #203
> Briarcliff Manor NY 10510
> 845-480-2058
>
> hobbs at electrooptical dot nethttp://electrooptical.net

What kind of lock-in are you referring to? Is it a Homodyne where the
input signal to be measured is essentially multiplied by a square wave
in analog circuits?

[spflanze]

I am now considering filtering higher harmonics and noise by using a
switched capacitor filter such as described here:
http://www.maxim-ic.com/app-notes/index.mvp/id/2081
Instead of inputting a square wave I could input the output of a DDS
and by so doing start with lower harmonics to filter out.

I am concerned about switching noise. In prior experience with
switched capacitor filters I have observed switching noise in the
output. I would like the switching frequency to be well above the
cutoff of the anti-alias filter. For this to be I figure I would need
at least 10,000:1 ratio of switching frequency to cutoff frequency. In
the filters I have seen 100:1 is typical and 1000:1 is rare. I have
not found one with 10,000:1. Does it exist?

Failing that I could make the switching frequency a multiple of 1/T
where T is the length of time the FFT is done over. That would put the
switching frequency in one of the notches in the sync function that is
the frequency response of a Fourier Transform.

I could create a switched capacitor filter out of several chips. But
it would make the board too large. I would need it in a single chip.

There would also be the 6 different switching frequencies to generate.
I know how to do this with several counter and logic chips. But any
lower chip count suggestions are welcome, especially a singe chip with
several channels of output.

[Bill Sloman]

On Jan 24, 12:31 am, spflanze <art...@wavenet.org> wrote:
> On Jan 23, 11:51 am,BillSloman<bill.slo...@ieee.org> wrote:
>
>
>
>
>
>
>
>
>
> > On Jan 23, 6:20 pm, spflanze <art...@wavenet.org> wrote:
>
> > > On Jan 21, 11:09 pm, miso <m...@sushi.com> wrote:
>
> > > > If we did a mental exercise and reduced the bits in the DDS, eventually
> > > > it would be a square wave. So wouldn't the first noise appear at 3x the
> > > > modulation frequency if the modulation frequency and clock rate have an
> > > > integer relationship. If so, then your post filter is substantially
> > > > easier to build.
>
> > > That would be true if it were a perfect DAC with infinite bandwidth.
> > > But there is a finite bandwidth and finite slew rate so its output is
> > > other than a succession of perfect squares. Triangular elements are
> > > involved and those have even harmonics.
>
> > A triangular wave is just the integral of a square wave, and has the
> > same - odd only - harmonics, but with the amplitude dropoing in
> > proportion to the square of the harmonic number.
>
> > But yes, the output from a DDS is potentially messier than the output
> > from simpler switching circuit.
>
> > Have you thought about a delay-line filter? You clock your square wave
> > through a shift register, hook up a suitable resistor to each stage,
> > and sum the outputs. IIRR a sinc function tapered with a Hamming
> > window to kills the Gibbs oscillation can give you a very clean sine
> > wave; you've got residual high-frequency harmonics, but the resistors
> > can be 0.1% parts (with a bit of padding to get the exact values) and
> > the lower harmonics can be held to better than 60dB down.
>
> > <snip>
>
> It is the sawtooth that has the even harmonics. You are right that the
> triangular wave can be considered the result of a square wave with its
> higher harmonics being attenuated by an integrating filter.
>
> I can see how an array of resistor values on a shift register can be
> chosen to replicate the impulse response of a filter. I do not follow
> how a sync function can be implemented with this. A reference would be
> appreciated.

It's not a synchronising function but sinc - sine(x)/x which gives the
resistor value at each tap.

Google wasn't too helpful, until I remembered that the structure is
also known as a transversal filter

http://www.rane.com/note122.html

shows one at Fig 25

http://www-sigproc.eng.cam.ac.uk/~op205/3F3_5_Design_of_FIR_Filters.pdf

goes into the theory.

> What advantage would a delay line filter have over a Sallen-Key filter
> with many poles?

It's more compact, and you can get 0.1% tolerance resistors, while 1%
tolerance capacitors are as good as you can buy. It's also a finite
impulse response filter, with a linear phase shift.

--
Bill Sloman, Nijmegen

[Bill Sloman]

On Jan 24, 6:29 am, Phil Hobbs
<pcdhSpamMeSensel...@electrooptical.net> wrote:
> BillSlomanwrote:

>
> > On Jan 23, 10:09 pm, "tm" <No_one_h...@white-house.gov> wrote:
> > > "spflanze" <art...@wavenet.org> wrote in message
>
> > >news:e2aab768-3c58-4381...@t13g2000yqg.googlegroups.com...
>
> > > > If both the modulating and demodulating waveforms are square wave
> > > > there will be contributions to the output at odd harmonics where the
> > > > signal to noise ratios of the system is less. The square wave is the
> > > > equivalent of doing both a carrier signal and Fourier Transform at the
> > > > fundamental and again at all odd harmonics and summing the results.
>
> > > > The odd harmonics can be filtered. A filter with enough attenuation at
> > > > the first of the harmonics and very little attenuation at the carrier
> > > > frequency would require many poles, and would require many op amps.
>
> > > Use a notch (or several as needed) as well as the LPF to get rid of the
> > > harmonics.
>
> > Notches aren't a good idea inside a lock-in amplifier set-up. Too much
> > phase shift.
>
> > You start needing to add all-pass networks to your filter networks to
> > keep the phase-shift low and stable.
>

> Actually notches are great inside feedback loops, because their phase
> shifts go away at frequencies more than a few notch widths away.  Long
> long ago, when I was designing atomic force microscopes, we got a factor
> of about 4 increase in speed by notching out the mechanical resonance of
> the silicon cantilever.
>
> If you have a resonant actuator, a notch filter can be a _big_ win.

Agreed. You just have to want to notch out a frequency that has
nothing to do with your modulation frequency, which doesn't happen all
that often.

--
Bill Sloman, Nijmegen

[Nico Coesel]

spflanze <art...@wavenet.org> wrote:

>I am now considering filtering higher harmonics and noise by using a
>switched capacitor filter such as described here:
>http://www.maxim-ic.com/app-notes/index.mvp/id/2081
>Instead of inputting a square wave I could input the output of a DDS
>and by so doing start with lower harmonics to filter out.

What frequencies are you talking about? I'd sample the signals using
an ADC and do the whole filtering and processing digitally. In the
analog domain you get extra noise and errors with every stage you add.

--
Failure does not prove something is impossible, failure simply
indicates you are not using the right tools...
nico@nctdevpuntnl (punt=.)
--------------------------------------------------------------

[John Devereux]

Phil Hobbs <pcdhSpamM...@electrooptical.net> writes:

> Fred Bartoli wrote:
>>
[...]

>>
>> Don't know for noise spec but I once encountered a low frequency issue
>> which was traced down to convection combining with the more often spec'd
>> LED output efficiency tempco. (the LEDs were driven withing their rated
>> specs)
>>
>> --
>> Thanks,
>> Fred.
>
> Drift has a characteristic 1/f**2 dependence, so it usually dominates at
> very low frequency.

That's the same as sqrt(t) right? Like the stability specifications you
sometimes get with references.

<http://www.linear.com/product/LTZ1000>

So the graph of typical long-term drift has a square root shape.

Hmm, is this related to the random walk sqrt(N)?

--

John Devereux

[John Larkin]

Given what you're trying to do, you really, really ought to get his
book.

John

[Phil Hobbs]

Not so--that's exactly what it's good for. You can't notch out the
modulation frequency before the demod, obviously, but you can widen the
lowpass on the output if you notch out the ripple components.

[Phil Hobbs]

Yes, it's random walk drift, sometimes called "Brown noise". Uniform
drift doesn't have a Fourier transform, strictly speaking, but you can
compute it formally by integrating a delta function twice. That gives
you 1/f**2 in amplitude, so 1/f**4 in power. The random version doesn't
always go the same direction, so if you do some stochastic Fourier
integral that I'd have to look up, it winds up being 1/f**2.

[George Herold]

On Jan 23, 8:05 pm, George Herold <gher...@teachspin.com> wrote:
> On Jan 23, 6:05 pm, spflanze <art...@wavenet.org> wrote:

>
>
>
>
>
> > On Jan 23, 11:59 am, Mikko OH2HVJ <oh2...@sral.fi> wrote:
>
> > > spflanze <art...@wavenet.org> writes:
> > > > We are looking for the smallest possible changes in transmissivity. So
> > > > I am looking to do state of the art in noise performance.
>
> > > What is your target on transmissivity change and available light power ?
>
> > > Can you take part of the light before the chemistry and use that as
> > > the reference for the intensity ? This with lock-in detection (done
> > > in HW or SW) has given good results for us in similar measurements.
>
> > > --
> > > Mikko OH2HVJ
>

> > The photodiode would look at the LED sideways, which would be outside
> > its lens's cone, but would still have plenty of light due to its
> > proximity. After Bill Sloman's response I am not sure about noise
> > suppression. I would expect a photodiode in such an arrangement to get
> > more light from some areas of the light emitting surface than others.
> > I have some lab experience where the feedback seemed to suppress noise
> > on the output of a Luxeon III. I do not know where the monitoring
> > photodiode was relative to this LED because both were in a sealed
> > unit. The two LEDs being considered now are:
>

> >http://partfinder.avagotech.com/Avago/Avago.jsp#!partSearch=&partno=a...

>
> > There is not yet a specific target.
>

> Dang optical units.  The Avago part has the following unit's
>
> 50mA currrent ~2V Vf
> 600nm wavlength, 8 degree view angle.
> and the 'winners'
> luminous efficacy 355 lm/W
> luminous efficiency 27 lm/W.
>
> So if I stuck a big photodiode in front of this at 50mA how many watts
> or amps of current will I measure?
>

> George H.- Hide quoted text -

>
> - Show quoted text -

OK to answer my own question... Reading the spec sheet in the light of
day I found the ouput 'power' of 22 cd. Which is 62mW/sr, and with
the 4 degree 1/2 angle about 0.1mW of power. So total power is
something like 0.2mW

George H.

[spflanze]

How many resistors would it take to get the equivalent quantization
noise of 16 bits_

[spflanze]

I am looking for low noise sine wave generation to drive an LED light
source. Frequencies will be somewhere between 100Hz and 1kHz. I will
set the frequencies in the part of the noise spectrum where noise is
lowest.

ADC does not apply. DAC and DDS does. But I have decided not to use
them for this for reasons I will give in another reply.

[John Larkin]

50 mA at 2 volts is 100 milliwatts, times 27 l/w gives 2.7 lumens.
That's roughly 4 milliwatts of light. I think. That would be around 4%
efficiency, which sounds possible.

A photodiode of, say, 0.5 A/W would give 2 mA, for a CTR of 4%, again
sort of reasonable.

John

[spflanze]

I have decided not to use a DDS. I recognize now that the same
quantization noise equation with its white noise approximation that
would apply to the 16 bit ADC would apply to the DDS. So if I use a 10
bit DDS I will get the quantization noise performance of 10 bits, not
16 bits. Since the white noise approximation applies there is no
filtering this out. The 14 Bit DDSs consume too much power. I am going
to input square waves to eight pole elliptical switched capacitor
filter chips. There will be a low level of switching noise that passes
through the LED driver to the photodetector, and through the anti-
aliasing filter. But if the switching frequencies are chosen right the
FFT will notch these frequencies out.

There has been mention in another reply of an LTZ1000 reference. I
have looked at the specs and it looks like a very good one. I plan to
use one them to voltage reference both the ADCs and the square wave
generators. Thanks for bringing this one up.

[spflanze]

On Jan 23, 9:46 pm, Phil Hobbs

If I understand this right the output of a monitoring photodiode is
scaled to be the same as the output from the measuring photodiode,
then subtracted from the measuring photodiode's output. When the
output of the measuring photodiode changes because an optrode or
whatever has detected something the laser noise will return and be
proportional to the signal, but be less than it would be without the
subtraction.

[George Herold]

On Jan 24, 1:01 pm, John Larkin

<jjlar...@highNOTlandTHIStechnologyPART.com> wrote:
> On Tue, 24 Jan 2012 09:14:07 -0800 (PST), George Herold
>
>
>
>
>

> John- Hide quoted text -

>
> - Show quoted text -

Oh yes the 22cd number is for 20mA of current... 50mA will give 2.5
times more.
so 0.5mW of light power. (Of course I could have made a factor of 2
or 4 math mistake too.)

George H.

[Mikko OH2HVJ]

George Herold <ghe...@teachspin.com> writes:

> So how 'good' can you do Mikko? I've looked at these really tiny
> changes in our optical pumping apparatus,
> (a straight DC light level measurement, no modulation or lockin) and
> seen changes at smaller than a part in 10^5. (That was averaging
> several sweeps though.. Signal to noise in one sweep is about 1 at
> changes of a part in 10^5.)

The first setup we built about 2 years ago got 0.00001 AU peak-to-peak noise
over 10 minute period with 1s sampling, about 400Hz simple LED source.

Our first try was a DC measurement and 1:10^5 sounds very good for DC!

Now we're doing a digital lock-in, but that is more due to low light available.
We get 0.0002 AU peak-to-peak noise over 100s with 800Hz LED, 200ms measurements.

--
Mikko

[Bill Sloman]

On Jan 24, 6:50 pm, spflanze <art...@wavenet.org> wrote:

> How many resistors would it take to get the equivalent quantization
> noise of 16 bits_

There's no quantisation noise. It's just filtering a square wave, and
the number of resistors/length of shift register determines where the
higher harmonics begin to come through. You normally shape the
frequency response of the summing amplifier to attenuate them.

--
Bill Sloman, Nijmegen

[Bill Sloman]

On Jan 24, 5:22 pm, Phil Hobbs

> Not so--that's exactly what it's good for.  You can't notch out the
> modulation frequency before the demod, obviously, but you can widen the
> lowpass on the output if you notch out the ripple components.

"Widen the lowpass"? You can certainly complicate the post-
demodulation low pass filter by adding a notch at the ripple frequency
- and an easily tunable notch like a bridge differentiator would be
the way to go - but a higher order low pass filter is usually a more
attractive option, since you don't need to tune it to match your
frequency source or worry about temperature drifts moving the notch
away from the ripple frequency.

--
Bill Sloman, Nijmegen

[Phil Hobbs]

Nonsense. Higher order filters make the phase shift worse. You're
moving the goal posts.

[Jamie]

spflanze wrote:

> On Jan 23, 4:21 pm, George Herold <gher...@teachspin.com> wrote:

>
>>On Jan 23, 5:48 pm, Bill Sloman <bill.slo...@ieee.org> wrote:
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>>On Jan 23, 10:09 pm, "tm" <No_one_h...@white-house.gov> wrote:
>>
>>>>"spflanze" <art...@wavenet.org> wrote in message
>>
>>>>news:e2aab768-3c58-4381...@t13g2000yqg.googlegroups.com...
>>
>>>>>If both the modulating and demodulating waveforms are square wave
>>>>>there will be contributions to the output at odd harmonics where the
>>>>>signal to noise ratios of the system is less. The square wave is the
>>>>>equivalent of doing both a carrier signal and Fourier Transform at the
>>>>>fundamental and again at all odd harmonics and summing the results.
>>
>>>>>The odd harmonics can be filtered. A filter with enough attenuation at
>>>>>the first of the harmonics and very little attenuation at the carrier
>>>>>frequency would require many poles, and would require many op amps.
>>
>>>>Use a notch (or several as needed) as well as the LPF to get rid of the
>>>>harmonics.
>>
>>>Notches aren't a good idea inside a lock-in amplifier set-up. Too much
>>>phase shift.
>>
>>>You start needing to add all-pass networks to your filter networks to
>>>keep the phase-shift low and stable.
>>

>>>--
>>>Bill Sloman, Nijmegen- Hide quoted text -

>>
>>>- Show quoted text -
>>

>>Nah, they're fine. You don't want to make the Q too high, but a nice
>>Q of ten can really help with the dynamic range. You, of course,
>>have to adjust everything, center the filter and adjust the phase.
>>It's tough to get down to the noise level without a bandpass somewhere
>>in the system.
>>
>>George H.
>
>
> I do not need any frequencies about the fundamental or carrier. Why
> put a notch at a harmonic when I could use a low pass to take out all
> the harmonics?
Guess if you're not worried about phase error, then go for it.

Jamie

[Bill Sloman]

On Jan 24, 9:05 pm, Phil Hobbs

> Nonsense.  Higher order filters make the phase shift worse.  You're
> moving the goal posts.

Actually, you are. Phase shift after demodulation doesn't usually
matter - the exception is when you are closing the loop around the
whole modulator-demodulator set up, and even then you've got to have
some kind of phase-lag budget to make a practicable system.

I know you feel the need to pose as infallible expert, in much the
same way the John Larkin feels the need to insist that all his
products represent "insanely good design", but trying too hard defeats
the purpose.

--
Bill Sloman, Nijmegen

[John Larkin]

Moron. I never said anything of the kind.

The real difference between the stuff I design, and the stuff you used
to design, is that most of my stuff sells. And, of course, I'm still
designing.

John

[Phil Hobbs]

You were the one who said that notches were a problem because of their
phase shift, and then you try fixing that by bringing in a higher order
filter, which makes the problem worse. Now all you have left is name
calling, and I'll leave that to you.

[spflanze]

On Jan 23, 3:05 pm, spflanze <art...@wavenet.org> wrote:
> On Jan 23, 11:59 am, Mikko OH2HVJ <oh2...@sral.fi> wrote:
>
> > spflanze <art...@wavenet.org> writes:
> > > We are looking for the smallest possible changes in transmissivity. So
> > > I am looking to do state of the art in noise performance.
>
> > What is your target on transmissivity change and available light power ?
>
> > Can you take part of the light before the chemistry and use that as
> > the reference for the intensity ? This with lock-in detection (done
> > in HW or SW) has given good results for us in similar measurements.
>
> > --
> > Mikko OH2HVJ
>
> The photodiode would look at the LED sideways, which would be outside
> its lens's cone, but would still have plenty of light due to its
> proximity. After Bill Sloman's response I am not sure about noise
> suppression. I would expect a photodiode in such an arrangement to get
> more light from some areas of the light emitting surface than others.
> I have some lab experience where the feedback seemed to suppress noise
> on the output of a Luxeon III. I do not know where the monitoring
> photodiode was relative to this LED because both were in a sealed
> unit. The two LEDs being considered now are:
>

> http://partfinder.avagotech.com/Avago/Avago.jsp#!partSearch=&partno=a...http://www.cree.com/products/pdf/LEDlamps/C503B%281079%29.pdf

>
> There is not yet a specific target.

I am going to build so the light coming out the sides is in a small
cavity with white walls for an integrating sphere effect. The
photodiode would be at the cavity wall. I expect this would make the
monitoring photodiode's current more representative of emissions from
the entire emitting surface. The size of the cavity is going to be
limited by available space.

Another idea I will look into is to have a surface mount led mounted
over a hole in the PCB, and on the other side of the hole have a
surface mount monitoring photodiode. What I am not sure about is how
much light comes out the underside of a surface mount LED. This also
means there will be no lens in the LED package, and I am not sure how
accepting the optics designer will be of that.

[George Herold]

> Bill Sloman, Nijmegen- Hide quoted text -
>
> - Show quoted text -

Hmm, in a control application you may want the wider bandwidth.

I've got a piezo + (spring and mass) that sings around 1kHz....
What happens if I notch it out of the control loop?

George H.

[George Herold]

On Jan 24, 1:40 pm, Mikko OH2HVJ <oh2...@sral.fi> wrote:

> George Herold <gher...@teachspin.com> writes:
> > So how 'good' can you do Mikko?  I've looked at these really tiny
> > changes in our optical pumping apparatus,
> > (a straight DC light level measurement, no modulation or lockin) and
> > seen changes at smaller than a part in 10^5.  (That was averaging
> > several sweeps though.. Signal to noise in one sweep is about 1 at
> > changes of a part in 10^5.)
>
> The first setup we built about 2 years ago got 0.00001 AU peak-to-peak noise
> over 10 minute period with 1s sampling, about 400Hz simple LED source.
>
> Our first try was a DC measurement and 1:10^5 sounds very good for DC!

Well there is a 'bit of DC drift.

I've got pictures! Source is a Rubidium lamp, through a Rb reference
cell*, a little less than 10uA of photocurrent. DC level is near 10V,
a gain of 1000 and a 100ms lowpass filter. This is a sweep through
zero magnetic field.
http://imageshack.us/photo/my-images/191/tek0009.png/
There's a little W shaped thing....About 50mV of signal in 50mV of
noise. The drift is due to the temperature drift of the Rb cell.
Here's a similar signal AC coupled and averaged.
http://imageshack.us/photo/my-images/593/tek0024.png/
I 'claim' the broad dip is an optical pumping absorption and the goes
uppy part in the center is CPT. Weird quantum stuff done without a
laser.
George H.

*there's also an interference filter (picks out one spectral line) and
a linear polarizer before the cell... optics details...

[Jamie]

Hopefully with new applications and not fixing exiting ones! :)

Jamie

[George Herold]

On Jan 24, 6:46 pm, spflanze <art...@wavenet.org> wrote:
> On Jan 23, 3:05 pm, spflanze <art...@wavenet.org> wrote:
>
>
>
>
>
> > On Jan 23, 11:59 am, Mikko OH2HVJ <oh2...@sral.fi> wrote:
>
> > > spflanze <art...@wavenet.org> writes:
> > > > We are looking for the smallest possible changes in transmissivity. So
> > > > I am looking to do state of the art in noise performance.
>
> > > What is your target on transmissivity change and available light power ?
>
> > > Can you take part of the light before the chemistry and use that as
> > > the reference for the intensity ? This with lock-in detection (done
> > > in HW or SW) has given good results for us in similar measurements.
>
> > > --
> > > Mikko OH2HVJ
>
> > The photodiode would look at the LED sideways, which would be outside
> > its lens's cone, but would still have plenty of light due to its
> > proximity. After Bill Sloman's response I am not sure about noise
> > suppression. I would expect a photodiode in such an arrangement to get
> > more light from some areas of the light emitting surface than others.
> > I have some lab experience where the feedback seemed to suppress noise
> > on the output of a Luxeon III. I do not know where the monitoring
> > photodiode was relative to this LED because both were in a sealed
> > unit. The two LEDs being considered now are:
>
> >http://partfinder.avagotech.com/Avago/Avago.jsp#!partSearch=&partno=a...
>

> > There is not yet a specific target.
>
> I am going to build so the light coming out the sides is in a small
> cavity with white walls for an integrating sphere effect. The
> photodiode would be at the cavity wall. I expect this would make the
> monitoring photodiode's current more representative of emissions from
> the entire emitting surface. The size of the cavity is going to be
> limited by available space.
>
> Another idea I will look into is to have a surface mount led mounted
> over a hole in the PCB, and on the other side of the hole have a
> surface mount monitoring photodiode. What I am not sure about is how
> much light comes out the underside of a surface mount LED. This also
> means there will be no lens in the LED package, and I am not sure how

> accepting the optics designer will be of that.- Hide quoted text -

>
> - Show quoted text -

Sounds like you need to cobble together some prototypes.

George H.

[John Larkin]

Well, I spent a bunch of today changing some code that I wrote in
1996. Some 20-bit serial DACs went obsolete, so we replaced them with
a 16-bit DAC and a CPLD to shuffle the bits. But there's a glitch, and
the easiest fix is to modify the 68K assembly code that drives the
DACs. Having a product line is like having kids or pets.

John

[George Herold]

On Jan 24, 9:08 pm, John Larkin

I always feel like I'm wheeling out a new baby, when some piece of
equipment, passes testing and goes to shipping.

George H.
>
> John- Hide quoted text -

[spflanze]

On Jan 23, 3:31 pm, spflanze <art...@wavenet.org> wrote:

> > --
> > Bill Sloman, Nijmegen

>
> It is the sawtooth that has the even harmonics. You are right that the
> triangular wave can be considered the result of a square wave with its
> higher harmonics being attenuated by an integrating filter.
>

The sawtooth waveform with its even harmonics is a component of a
staircase waveform:
http://www.analog.com/static/imported-files/tutorials/MT-001.pdf
This is very much like what the sine wave looks like on its way up and
down. The output of the DDS will have even harmonics.

[spflanze]

I was wrong about the white noise approximation applying to DDS
quantization noise. Because the quantization noise is synchronous with
the sine wave there is no psuedo randomization and so no white noise.
The harmonics will be clearly defined. My decision not to use the DDS
and instead input a square wave into a switched capacitor filter still
stands because I think I will still get a good result and it will be
easy to generate the square wave from the switching frequency I must
generate anyway. I think the DDS will be an additional part with
little benefit.

[John Larkin]

On Tue, 24 Jan 2012 20:08:19 -0800 (PST), spflanze

Switched-cap filters are noisy and nasty. They also alias everything
in sight, including their own power supply noise. And they make big
output spikes.

All you need is a modest analog lowpass filter ahead of the ADC, just
to keep the noise down.

John