Is DSP useful?

[JC1]

Came from this

JC,

I don't understand why all the discussion over DSP. We are not far enough along with this whole Design Group concept to even begin talking about
architectures. It's entirely possible that an eventual design team will decide that DSP is not the right approach. But it's pointless to make that decision now, or even debate it.

- Carl

Quote Sean_Goddard

One thing is for sure, we will need someone with more than a little understanding of DSP. Do we have anyone who fits this bill on the forum? Software too PC and micro controller if possible. Maybe ASIC design too.


My comments came from this statement

and my sensitivity to buzz words like

Artificial Intellegence, fuzzzzy loigic,

Digetital Sagnal Progcessing.

Rant rant rant.

[JC1]

Modulation

AH, Now I see JC1. Very good points raised there. I've read the book too, EXCELLENT!!

If super narrow DSP filters change the phase info so much, how come QAM64 used for digital television in the UK works so well? And the military radio I ws telking about demods the I and the Q of the signal as each contains separate data (NOT sure I should have said that on the forum).

Don't get me wrong, I'm NOT arguing with you as from your writings you obviously know what you are talking about.

I concur, the trick with DSP is WHAT to do to WHAT. BTW mark Rowan from Whites and I have virtually identical backgrounds, freaky eh?

BTW, I built a chirp detector aboout a year ago using a PIC and a DDS chip and a few other bits, it worked just fine, great sensitivity across the whole range. I've recently updated the design and it now has two Tx channels (four output DDS) one sweeps down, the other up, though I've found at the crossover you have to blank one channel or it overload the front end badly. The output is a sawtooth waveform from the demodulators, this is then nulled by an opposite polarity sawtooth generated by a trigger from the PIC and an analog circuit controlled by digipots and switchable capacitors. A target shows up as a "blip" or "pimple" of the positive going slope of the demod output, or if a large target, an overall increase in DC level on the I and Q channels. It's basically a normal VLF machine with a swept Tx signal. Works very well. I'll dig out the schematics when I get my lab built again after the recent move.

Interesting thing if I remember is that on the "Disc" channel, the amplitude and position of the pimple are dependent of the type of target material. Maybe someone could use this information further and investigate a new discriminate method. I guess it's because things like gold and silver max phase shift at different frequencies.

Hi Sean,

You sound like you understand this

very well. The digital TV pobably has

the carrier at a much higher frequency

then the modulation and the modulation

is not as effected by the phase distortions.

The phase on most sharp digital filters

is a ripple routine in the pass and stop bands

just like the amplitude. There are group delay

equilization routines for digital filters by the way

just for this problem. Of course there is more

space and power in TVs then metal detectors.

Now for a VLF the carrier is everything, and

the only real modulation is amplitude as you get

near far from metal, and the phase of the return

signal for discrimination.

And you sound like you know the difference

between DSP and just microprocessor coding.

DSP meaning Processing the Return Signal by

Digital methods.

Now apparently the Plan is to put together design

teams and then figure out what to do.

I always figured out

what needed to be done, and then put together the

teams. I Probably had it backwards.

[JC1]

Group delay

Digital variable group delay equalizer for a digital television receiver.

http://www.freepatentsonline.com/4531149.html

this is interesting, fixing the group delay in the analog

filters for the DACs for chroma, etc.

Have to pay attention to this stuff.

[JC1]

effects

Within the transmission of B,G/PAL, a precorrection of group delay is performed at the transmitter. The necessity of this group delay precorrection dates from the time when it was difficult and expensive to make receiver IF-filters with a flat group delay response characteristic. Nowadays, receiver IF-filters with a flat group delay response are available on the market at relatively low prices. Consequently, there is no longer any need for group delay precorrection.

There is a strong drive, coming from German broadcasters, to cancel the group delay precorrection process. They want to do this in two steps, viz. by halving the amount of group delay precorrection by 1994, and by cancelling the group delay precorrection entirely by 2001. However, cancelling the transmitter group delay precorrection causes a noticeable loss of sharpness and a chrominance/luminance delay difference in television receivers which still expect the television signal to have undergone the transmitter-sided precorrection.

http://www.patentstorm.us/patents/54...scription.html

[JC1]

Group delay

A system and method determines a group delay of a channel in a CATV ... [0037] In one commonly used digital modulation scheme, known as the QAM-64 scheme, ...

http://www.freepatentsonline.com/20020168001.html

scroll down.

[JC1]

Carrier Reconstruction

Hi Sean,

Still answering your question on QAM,

Not up to speed on UK TV but this is interesting

and not atypical of several systems.

They reconstruct the carrier, and correct

everything. Interesting but complex.

Real Carrier Frequency

A QAM signal is a suppressed carrier modulation, only information carrying sidebands are transmitted. Consequently it is difficult to measure the suppressed carrier. On the other hand, a carrier recovery mechanism is used for demodulation with a PLL, in analog form or in digital form. The nominal carrier frequency is a reference frequency for the PLL. The Loop correction value (together with the frequency/value relationship) is a measure of the real carrier, somewhat offset from nominal.

Equalizer Frequency and Delay Response

The equalizer is a time linear filter, using past and future samples (relative to the reference sample) that are added to the reference sample through complex multipliers. Normally the coefficients of the multipliers are adaptively determined to minimize the Inter-Symbol Interference. This effectively compensates for linear distortions of the signal such as frequency response, phase response and reflections.

The proposed method of the signal parameters measurement system comprises the reading of the complex coefficients of the equalizer, from these and using known techniques in the art (such as Fast Fourier Transforms (FFT) and the like), the method calculates the amplitude and phase vs. frequency responses of the filter and the group delay from the phase responses, to finally display on the screen of the instrument the amplitude and group delay for the operator to interpret channel signal parameters.


http://www.patentstorm.us/patents/71...scription.html

[JC1]

Chirped Waveform

AH, Now I see JC1. Very good points raised there. I've read the book too, EXCELLENT!!

Don't get me wrong, I'm NOT arguing with you as from your writings you obviously know what you are talking about.

I concur, the trick with DSP is WHAT to do to WHAT. BTW mark Rowan from Whites and I have virtually identical backgrounds, freaky eh?

BTW, I built a chirp detector aboout a year ago using a PIC and a DDS chip and a few other bits, it worked just fine, great sensitivity across the whole range. I've recently updated the design and it now has two Tx channels (four output DDS) one sweeps down, the other up, though I've found at the crossover you have to blank one channel or it overload the front end badly. The output is a sawtooth waveform from the demodulators, this is then nulled by an opposite polarity sawtooth generated by a trigger from the PIC and an analog circuit controlled by digipots and switchable capacitors. A target shows up as a "blip" or "pimple" of the positive going slope of the demod output, or if a large target, an overall increase in DC level on the I and Q channels. It's basically a normal VLF machine with a swept Tx signal. Works very well. I'll dig out the schematics when I get my lab built again after the recent move.

Interesting thing if I remember is that on the "Disc" channel, the amplitude and position of the pimple are dependent of the type of target material. Maybe someone could use this information further and investigate a new discriminate method. I guess it's because things like gold and silver max phase shift at different frequencies.

Ah yes this is better thanks for starting a side

thread so I won't bother anyone with my tech rants.

Hi Sean,

Interesting sweeping the freq on a VLF, that will make

the phase on Non-ferrous move around..

I was talking more of a real radar as a GPR.

With the Chirped Waveform as a Compression Radar

with the Chirp being compressed in a linear Phase filter

so that all the energy of the chirp comes out of the filter

at the same time. Maximum Range.

Here is a wiki link kinda describes it. I thought

maybe this is what you meant when you said GPR.



Synthetic aperture radar (SAR) is a form of radar in which sophisticated post-processing of radar data is used to produce a very narrow effective beam. It can only be used by moving instruments over relatively immobile targets, but it has seen wide applications in remote sensing and mapping.

Chirped (Pulse Compressed) Radars

A common technique for many RADAR systems (usually also found in SAR systems) is to "chirp" the signal. In a "chirped" radar, the pulse is allowed to be much longer. A longer pulse allows more energy to be emitted, and hence received, but usually hinders range resolution. But in a chirped radar, this longer pulse also has a frequency shift during the pulse (hence the chirp or frequency shift). When the "chirped" signal is returned, it must be correlated with the sent pulse. Classically, in analog systems, it is passed to a dispersive delay line (often a SAW device) that has the property of varying velocity of propagation based on frequency. This technique "compresses" the pulse in time - thus having the effect of a much shorter pulse (improved range resolution) while having the benefit of longer pulse length (much more signal returned). Newer systems use digital pulse correlation to find the pulse return in the signal.

http://en.wikipedia.org/wiki/Synthetic_aperture_radar

[JC1]

Patent Infractions

Hi Sean,

I had thought of sweeping a VLF in freq.

one time long ago more as way to get

around patents. Something new and

different enough to justify ownership.

the front end bandpass filters would

have to be wide enough to accomodate

the sweep range {more noise} or have

the filters adjustable in center freq,

multiple feedback, and track the sweep

{prefer}. If you find the schematics fine

otherwise I will still understand your methods

with enough explaination.

Still probably should do something

"new" or "different" to avoid the patents

and such. Even if it is trouble.

And don't get me wrong DSP may have its

place, certainly the processors are faster,

but the radio and TV routines are probably

a bit complex, expensive, and heavy.

[JC1]

Lock in Amplifiers

Oh and when I mentioned

Lock In Amplifiers here is what

I was talking about.

And some PC software.

Hey look software is done !!!

Now find one in a DSP chip

{some codes for filters and such are

already written} maybe for lock in amp.

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

http://www.bentham.co.uk/pdf/Lock-in...20tutorial.pdf

http://www.signalrecovery.com/lockinex.htm

DSP lock in

http://www.thinksrs.com/products/SR850.htm

http://www.boselec.com/products/siglim.html

http://www.boselec.com/products/siglimnoise.html

http://www.mrflip.com/papers/LIA/

software

http://www.wam.umd.edu/~toh/models/lockin.html


Now this things will pick signals out of noise.

And we already have the reference signal {Transmit}

[JC1]

Lock Er In

Doing the DSP fer real.

At least the lock in amplier.

http://scitation.aip.org/getpdf/serv...ps&prog=normal

http://www.analog.com/en/prod/0,2877...RDWARE,00.html

http://shoppingcart.analog.com/Shopp.../AddModel.aspx

[JC1]

Sweeping Freq

Hey Sean,

You and mark Rowan are the same background.

Interesting not sure what that means.

ok.

I'm almost done with my new DSP metal detector.

Add coils, speaker, meter, batteries, and a stick

and I am DONE !!!

See above post.

even can do sweeping detector.

how are you coming on yours?

[Delbert grady]

DickySPy

Thank you Wizard for posting code above

could someone explain about well known metal detector companies waisting millions on DSP.

[TheWizard]

This way that you produce is really very hard....

ADSP-21369 from Analog Device is wrong for IB/PI MD - you need ADC > 1 Mega samples
(ADSP-21369, 24-bit codec is good but sampling is very very slow only 96kS, very big price)

QAM and digital TV technology is for Analog IB but for Digital, the error is too big...
FFT(like Garret) is too slow and need very fast DSP for small errors...
"delays and the phase errors" is constants and you can correct digitally...

Chirped Waveform, Compression Radar, SARs... is for radars and not for MD, but you can get some stuff from Radars

PC softwares for Lock in Amp, good, but methods is wrong and slow and is for analog metal detectors....

"The SR850 is a digital lock-in amplifier based on an innovative DSP" - is precise device for accurate measuring,
for MD you need DETECTOR not meter....

PS: I say NOT DSP is best of the best - but is good way....
and sorry for my English......

[TheWizard]

could someone explain about well known metal detector companies waisting millions on DSP.

sorry but I can not give you answer in this question...

maybe is:
Analog Dinosaurs,
High spirits/self-confidence,
very conservative....

[JC1]

Millions

sorry but I can not give you answer in this question...

maybe is:
Analog Dinosaurs,
High spirits/self-confidence,
very conservative....

The companies that I know spent

millions on DSP metal detectors

some many years ago was Whites

never made a detector from this.

Bounty Hunter maybe didn't spend

millions but made a half baked effort.

Nothing came from this.

Garrett first made just a metal detector

prototype which they show at a Walk

through show which had real poor performance.

Later made the GTi series.

Which they still sale.

Almost everyone who was anyone at

least looked into it. Garrett was the only

one I know of who actually made something.

Most of this came from talking to the

companies at the shows.

Now there are those companies who have

a microprocessor to light the lights and

graphic display, but ARE NOT doing DSP.

Even though they say they are.