ADC sampling a decay pulse... / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

ADC sampling a decay pulse...

Move Thread

LAN_403

Nicko

Tue Aug 24 2010, 10:02AM

Registered Member #1334 Joined: Tue Feb 19 2008, 04:37PM
Location: Nr. London, UK
Posts: 615

Hi,

I want to look at a very small region of the decay voltage generated by pulsing a coil - I'm putting about a 50uS/12V pulse into a 300uH 2ohm coil, and looking at the decay - I want to look closely at a 20uS window about 20 - 50 uS after the end of the pulse, and again at about 100uS later. Pulse repetition is about 2kHz, and I will run a moving average over the samples to try to minimise noise. What I'm looking for is any movement in the averages....

Now, this can be done by analogue means, but for various reasons, mainly due to a lot a parameters I want to continually alter (pulse widths, delays etc.), I want to make as much as possible done in software (filters, averages etc.).

If I use an 8-bit ADC from a uP, then the quantisation distortion would be huge, so I really just want to use analogue techniques to grab my sample window, baseline it to 0V and amplify it to FSD for the ADC - I'm not really interested in the absolute value of the samples, just the delta. To use 10-bit or better, I would have to play tricks with the uP to reduce digital noise (quiescing parts of the uP etc.), and that would add complexity.

Thoughts on an approach to this problem would be appreciated as my analogue skills are rusty and I need the best possible resolution I can get with the minimum noise...

Cheers

GeordieBoy

Tue Aug 24 2010, 11:47AM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

Can you use a general purpose data aquisition card (either inside your PC or via USB) to gather the data to sufficient bit-depth... Then extract the relevant features from the recorded data by post-processing it in something like Excel, BASIC or Matlab?

There are a whole load of data acquisition cards of varying spec and pricetag from companies like National Instruments etc. You might even be able to use the soundcard input of your PC at a push depending on the bandwidth of the signals you are trying to measure, but i'd generally steer clear of this if you can afford a propper DAQ card/box.

If you are still at the experimenting stage in whatever you are trying to do, then it might make sense to use an off-the-shelf solution instead of trying to build your own data acquisition system at this point in time?

-Richie,

Proud Mary

Tue Aug 24 2010, 12:49PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

How about down conversion via a UWB mixer to an IF at which measurements can easily be made?

Avalanche

Tue Aug 24 2010, 03:03PM

Registered Member #103 Joined: Thu Feb 09 2006, 08:16PM
Location: Derby, UK
Posts: 845

Don't know if I've got the wrong end of the stick as this is a very different solution, but if you're only interested in the delta I would arrange to have a differential amplifier before the input to the ADC, and amplify the difference between the two measurements (with zero on the ADC being no difference, no need for a bias assuming the second measurement will always be less than the first).

To get the 2 measurements, make a 'sample and hold' using 2 sections of a 4066 quad bilateral switch, with each section triggered at the right moment by the microcontroller, and a precision rectifier on each section. The outputs would be connected to holding capacitors on the inputs to the diff amp, then you'd have to arrange for the software sample to occur at the right time. Should work with decent holding capacitors!

radiotech

Tue Aug 24 2010, 03:22PM

Registered Member #2463 Joined: Wed Nov 11 2009, 03:49AM
Location:
Posts: 1546

Nikko wrote ""want to look closely at a 20uS window about 20 - 50 uS after the end of the pulse, and again at about 100uS later.

These numbers may alias with a test method used in power maintenance work:

If this can be of use try digging up the patents-they have been around for years.

Steve Conner

Sat Aug 28 2010, 10:46AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

Hi Nick

You want some sort of sample and hold circuit that will snip out the part of the pulse you're interested in, and hold the voltage for digitizing.

When you say you're only interested in the delta, do you mean the time rate of change of the voltage within the S&H window?

Or do you mean the difference between some measurement that you make on this pulse, and the same measurement on previous ones? (You mentioned a moving average)

The reason why I ask: If the former, you can get your delta by just differentiating in your analog electronics. If the latter, you probably have to low-pass filter or maybe even integrate on the analog side, and compute the delta digitally.

Another technique of interest might be to use your uC's DAC to produce a voltage equal to the moving average your firmware is calculating. Then feed this to a diff amp, subtracting it from the output of your sample and hold block. The difference should hopefully be small, so you can amplify it a lot before digitizing it, and you can then use it to update the moving average. This really amounts to turning the whole mess into some sort of sigma-delta converter, with as many bits of resolution as your DAC (if you're naive) or your moving average register (if you're a DSP wizard

)

Are you making a metal detector?

radiotech

Sat Aug 28 2010, 06:32PM

Registered Member #2463 Joined: Wed Nov 11 2009, 03:49AM
Location:
Posts: 1546

In high voltage testing- tickling- the best info returns with a gentle touch you need to know what might happen without having it happen. ..the old tale-- is this working? no? did you touch it?..
It sure would be nice to know what is being built.

Proud Mary

Sat Aug 28 2010, 09:14PM

Registered Member #543 Joined: Tue Feb 20 2007, 04:26PM
Location: UK
Posts: 4992

The use of a radio frequency front end mixer, which I suggested above, is used in the paper below, though your subsequent signal processing will no doubt be different. An obvious advantage is that it avoids frighteningly expensive speedy ADCs. ADC12D1600, for example, a 3.2 GSPS 12-bit ADC, cost Â£3250 each.

Rev. Sci. Instrum. 72, 1837 (2001); doi:10.1063/1.1344177 (6 pages)
Pulse saturation recovery, pulse ELDOR, and free induction decay electron paramagnetic resonance detection using time-locked subsampling
W. Froncisz1, Theodore G. Camenisch2, Joseph J. Ratke2, and James S. Hyde2

1Institute of Molecular Biology, Jagiellonian University, KrakÃ³w, Poland Map This map
2Biophysics Research Institute, Medical College of Wisconsin, Milwaukee, Wisconsinâ€‰53226-0509 Map This map

(Received 5 September 2000; accepted 4 December 2000)

Time locked subsampling (TLSS) in electron paramagnetic resonance (EPR) involves the steps of (i) translation of the signal from a microwave carrier to an intermediate frequency (IF) carrier where the (IF) offset between the signal oscillator and local oscillator frequencies is synthesized, (ii) sampling the IF carrier four times in an odd number of cycles, say 4 in 3, where the analog-to-digital (A/D) converter is driven by a frequency synthesizer that has the same clock input as the IF synthesizer, (iii) signal averaging as required for adequate signal to noise, (iv) separating the even and odd digitized words into two separate signal channels, which correspond to signals in phase and in quadrature with respect to the IF carrier, i.e., I and Q, and (v) detecting the envelope of I and also of Q by changing the signs of alternate words in each of the two channels. TLSS detection has been demonstrated in three forms of pulse EPR spectroscopy at X band: saturation recovery, pulse electronâ€“electron double resonance, and free induction decay. The IF was 187.5 MHz, the A/D converter frequency was 250 MHz, the overall bandwidth was 125 MHz, and the bandwidths for the separate I and Q channels were each 62.5 MHz. Experiments were conducted on nitroxide radical spin labels. The work was directed towards development of methodology to monitor bimolecular collisions of oxygen with spin labels in a context of site-directed spin labeling. Â© 2001 American Institute of Physics.

Â© 2001 American Institute of Physics

radiotech

Sat Aug 28 2010, 11:09PM

Registered Member #2463 Joined: Wed Nov 11 2009, 03:49AM
Location:
Posts: 1546

want to look closely at a 20uS window about 20 - 50 uS after the end of the pulse, and again at about 100uS later.

In an oscillation. When the amplitude of an oscillation decreases in geometric progression , the Naperian logarithm of the maximum displacement of any oscilation to the maximum displacement in the same direction of the imediately preceeding oscillation is known as logarithmetic decrement.

If you want to see how the effect of the load on your coil is changing over time by observing damping differences this paper
may point out some path to take. This guy is using a 500uhy coil.

This about as analog as it gets.

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob