stretching a pulse

All circuit theory deals with unobtainables: ideal components, zero propagation delays, noiseless signals. Otherwise you couldn't get anything done.

An impulse, from which the term "network impulse response" derives, has zero width. And an ideal delay line and an ideal integrator each have infinite bandwidth. So the thing above has a rectangular pulse response to a unit impulse input. A negative rectangular pulse, since I got the sign backwards.

No, what I'm trying to do is stretch a skinny spike into a longer flat pulse, but preserve the amplitude information. The length of the delay line determines the output pulse width. In my case, the input is about

2 ns wide and I'd like, say, a 6 ns wide pulse, because that would be easier to process accurately.

My real-world 2 ns spike, passed through the network above, will make a rectangular pulse with finite rise and fall times but flat on top, still fine for my purposes.

John

Didn't the original PAL color TVs use a delay line for just that reason, to slice out the interleaved color subcarrier?

John

"John Larkin" wrote in message news: snipped-for-privacy@4ax.com...

The answer is "duh, you can't", courtesy of Heisenberg.

If you want to do otherwise, you need Star Trek's Heisenberg Compensator ;-)

If the phase is consistent, you can scope it. But you need a sample rate / bandwidth high enough, which is technically difficult (GSa / bits / price). Compare-trigger-and-integrate methods are plausible but accuracy suffers similarly (jitter, threshold, response), and doesn't integrate data like a good signal processor, at least without fancy work. Downconversion won't work very well if the difference is less than a cycle, i.e., the pulse looks like DC instead of cycles. Indeed, the range of frequencies for which it looks like DC corresponds to the uncertainty principle all the same. Yes, you can always run a curve fitting on the resulting "secant" of a sine wave, but that similarly becomes ill-formed.

But that's not the end of it. Assuming the oscillator has a rectangular window (or any other type), you can look at the frequency spectrum of the signal and determine the window type and center frequency. Data is data, so if you've got everything, you know everything about it, regardless of the domain you're viewing it in.

Tim

--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms

As I recall, the delay was an entire line (~60us, whatever PAL is exactly), so they mix the color information from the previous line with the current line, which alternates or something, to cure the notorious color drift of NTSC (never twice the same color).

Tim

-- Deep Friar: a very philosophical monk. Website:

"John Larkin" wrote in message news: snipped-for-privacy@4ax.com...

That "ill-posed" thing is all mathematical crepe-hanging. Deconvolution is only ill-posed if you expect it to work well in low SNR situations. I use it routinely without problems. The usual issue is that people want to do something stupid like turn a Gaussian into a square, i.e. to invert a filter way out into its stopband, in the presence of additive noise.

That's a good place for something nonlinear.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal
ElectroOptical Innovations
55 Orchard Rd
Briarcliff Manor NY 10510
845-480-2058
hobbs at electrooptical dot net
http://electrooptical.net

On a sunny day (Fri, 03 Sep 2010 05:24:29 -0700) it happened John Larkin wrote in :

'Original?" bull. PAL has two delay lines. One 64 uS one line delay to delay chroma info one line and then add it, so phase errors cancel, and only amplitude errors remain. The delay line Tim is talking about (approx. 450 ns) is in the Y (luminance) path, to compensate for the delay in the color processing (of R-Y, G-Y, and B-Y). The 64 us delay line is a glass delay line, usually a reflecting type, excited by piezzo transducers tuned to the 4.43 MHz chrom center frequency, the bandwidth is small (few hundred kHz).

I did build a system with 64 uS glass delay lines in the seventies that had a bigger bandwidth, and delayed R, G, B each some lines (delay lines in series) If I remember correctly that system was called TriPal (invented by Professor Bruch, the inventor of PAL). In that system a red, green, and blue line was sequentially send and then added again at the other side with the delay lines. Worked great on a VCR.

In this wikipedia link you see a picture of both delay lines:

The 64 us glass delay line:

The 450 nS delay line you can make yourself, as Tim described:

SECAM also used a 64 uS delay line, in SECAM the color difference sugnals were modulated FM, and send line sequential, the delay line made those then available at the same time in the receiver, by combining the one line delayed color data with the current incoming.

Nonsense. A reciprocal counter can easily measure the frequency of one cycle of a 1-Hz sine wave to 10 significant figures or more, if the SNR is high enough.

Even a linear system can do very well. As long as the transform of the pulse envelope rolls off far enough at DC, you can construct the analytic signal unambiguously, which makes the instantaneous frequency perfectly well defined with no Heisenberg worries.

Heisenberg is a statement about the time-bandwidth product of the pulse and its transform, not the uncertainty of the peak position itself.

Frequency is just the time derivative of phase, after all.

Locating the peak is a noisy thing to do, and with only a few peaks to work with, the relative phase of the window and the carrier matters a lot.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal
ElectroOptical Innovations
55 Orchard Rd
Briarcliff Manor NY 10510
845-480-2058
hobbs at electrooptical dot net
http://electrooptical.net

It's pretty simple. +Delta function -> sharp positive-going step -Delta function -> sharp negative-going step.

Result: rectangle function.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal
ElectroOptical Innovations
55 Orchard Rd
Briarcliff Manor NY 10510
845-480-2058
hobbs at electrooptical dot net
http://electrooptical.net

Aren't they all digital now?

John

It's not ill-posed to engineers, who only need good-enough. It seems to annoy mathematicians.

Did I ever show you my deconvolution evolver algorithm?

ftp://jjlarkin.lmi.net/Inv_Conv_demo.jpg

I did this to allow me to build an ugly TDR system (yellow trace maybe) and botox it to sellability (white.) I may still do that one of these days. This is fun to play with. As you say, add a little noise and ask it to speed things up unreasonably, and it explodes in interesting ways.

John

So, you don't know that the National Television Standards Committee predates color TV, or that VIR & VITS eliminated the phasing problems in the '70s?

--
Politicians should only get paid if the budget is balanced, and there is
enough left over to pay them.

On a sunny day (Fri, 03 Sep 2010 06:58:06 -0700) it happened John Larkin wrote in :

Funny you say that, I just got a big book with 100+ different security cameras for sale stuffed in the mail, each and every one of those PAL analog. PAL is very very much alive, I have 3 running 24/7. PAL analog is absolutely great as it manages real good speed with low bandwidth over a simple coax. No latency is also extremely important for those applications. And, today still, most digital cameras that record on FLASH also have a PAL output. If somebody comes up with a webcam with 25 fps with no latency, mm, now that would be cool. As to the delay line in the decoders, there are chips that implement that as bucket brigade on silicon. Look up: TDA4661.pdf

ns

pretty

and

intermediate

Cute. Of course that's more or less how fast digitizing scopes work--interleave digitizers to make something evil-looking but stable, and then calibrate the daylights out of it and display something apparently well-behaved.

That rug is big enough to sweep a *lot* of dirt under it. Your gizmo will probably be fine.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal
ElectroOptical Innovations
55 Orchard Rd
Briarcliff Manor NY 10510
845-480-2058
hobbs at electrooptical dot net
http://electrooptical.net

That was the posh later version. Early PAL colour TVs relied on the eyes limited chroma resolution to average the adjacent vertical lines out. This could lead to interesting bar artefacts especially on testcards and chequered suits when large phase errors were present.

Not on any US TVs that I ever saw. It was always highly amusing how the newscasters face fluctuated around ghastly green or purple flesh tones or still worse was clamped to an inanimate pale orange. That was at least until the mid 80's when the tint control finally became redundant.

The other unkind acronym for NTSC was No True Skin Colours.

I always assumed these phase related colour problems were intrinsic to NTSC broadcast modulation until I saw Japanese TV which is also a variety of NTSC but correctly implemented and capable of almost matching PAL or SECAM in colour fidelity.

Regards, Martin Brown

That might work. We'd have to squash the delay line between pcb planes, ground and a power pour, and dance around any vias. A few optional places to add shorts, to tune line length, might be prudent.

The shorted line + integrator thing is appealing, because the line is half as long.

We usually use 5 or 6 mil minimum traces. A zigzag delay line would need trace spacings somewhat wider than the plane-plane spacing, to mimimize sideways coupling that would add dispersion. There's no need to squeeze a delay line into a specific small area... it could meander all over the board.

Etching structures like this into boards can be dicey. I prefer using lumped parts when I can, to allow more ways to get out of trouble.

John

It's hard to do stable, precision, fast, cheap analog processing with tubes.

John

It might well. But a law of nature seems to be "there are no good delay lines", perhaps because delay lines inherently store information. Fiber optics is the best, but that's not feasible here.

John

Measure the period of "n" cycles and do the math?

tm

--- news://freenews.netfront.net/ - complaints: snipped-for-privacy@netfront.net ---

That would require picosecond period measurement, which is expensive. What might work is measuring a lot of small snippets of frequency and averaging. If the oscillator is, say, 50 MHz, and runs in bursts that last, say, half a microsecond, a number of classic gated frequency measurements will return a series of small integers, like 24 25 24

24 25 24... whose long-term average will tell us frequency to high accuracy. The problem becomes how to measure the number of ticks in each gated window to PPM statistical precision, which implies a gate width controlled to picoseconds, or maybe femtoseconds. Another interesting problem.

John