A lock in amplifier multiplies an AC signal + noise by a synchronous reference signal. If the signal is sinwt and the ref is sinwt, the output is sin^2(wt) which is always positive. Regardless of the particular shape it can always be smoothed to a DC output proportional to the signal -- with enough time.
Instead of multiplying it would be better to _divide_ the noisy AC signal by the ref.
One advantage over lock in is a simple capacitor in parallel with the output shorts out the AC component which is the noise.
A much greater advantage is when the noise drops, you aren't forced to waste a lot of time smoothing a squared signal to DC.
It's already there.
Quotient sync filtering is better than lock in in every respect.
You really need to go back to your trig tables. There are two components in the multiplied signal, one a 0 frequency and one at 2w. It is the
2w signal that is being filtered away. The whole point of a lockin amplifier is that the low pass filter after the multiplier sets the noise bandwidth and you can make that filter arbitrarily narrow. You cannot make a bandpass filter arbitrarily narrow.
Using the term "shorts out" shows that you have very little electronics knowledge.
The settling time is set by the filter bandwidth which is set by your requirements for noise bandwidth. The laws of physics tell you what the requirements are and they are not up for discussion.
Crazy. Near the signal zero crossings, you'd be applying an enormous amount of gain to what's left, namely the noise and offset. Multiplying weights the output by the best part of the signal; division selects the worst.
There's actual math behind those statements, but I'm suspecting that you're not a mathy type.
Some lockins, maybe most, multiply the input by a square wave, not a sine, because that's easier to do precisely.
Not an electronics type, either.
If it were, it would be in widespread use by now.
Amateurs often assume, from ignorance, that they can have eureka insights that generations of pros have missed. That's very, very unlikely.
Going in the "wrong direction" is SIP (Standard Inventing Procedure).
Is any expression more hackneyed than "thinking inside the cylinder?"
You actually have twice as many zero *touchings* -- both types of sync filters are always greater than or equal to zero so it's not really zero "crossings" -- with a lock in than quotient sync.
Any smoothing advantages provided by the 2X higher frequency look ridiculous compared to the quotient filter where only the noise needs to be smoothed.
For that to be true there would need to be some fundamental reason one type of sync filter is better than another.
Remember, both are, without question, sync filters.
In this case there is a fundamental difference: Smoothing is only necessary for the noise with the quotient filter.
Show us what you got.
I'm not holding anything back. This is full disclosure.
This math proof is high skool level.
You openly admit you think it's a big deal to smooth AC noise out of a DC signal?
You'ld be surprised as the $tuff you can find one step off trail.
If you were a hiker a snake would have been bitten you by now.
If you were a surfer a shark would have eaten you by now.
Many or most breakthroughs come shortly after switching fields.
Just yesterday Obama was talking about the economy running on innovation.
Obama needs to put his money where his mouth is.
He needs to pay engineers to go on sabbatical, industry hop, spend 6 months in another field asking "dumb" questions, start looking for problems to solve _outside_ their specialty . . .
_Any_thing to break up the sclerosis.
Tech incendiary needs to be a cabinet level position.
Bret Cahill
"Make things as simple as possible . . ."
-- Einstein
"The world revolves, not around the inventors of new noises, but around the inventors of new values. It revolves *inaudibly*.
Well it is ok to think of new ideas. It is quite another to make baseless claims like you are doing. But then you were trying to make up your own physics on an earlier thread.
You really need to look at fourier analysis and see what is going on.
Yep, you have never looked at this.
Do a simulation and look at the frequency output. You have not done that.
You really need to look at the math.
So you have not done the math.
So you have not done the math.
You are not familiar with filters either.
Yes, empty cans, cigarette butts etc.
If you were a student you would have failed.
Dream on that you have found anything.
No, Obama puts OUR money where his mouth is.
While you are waiting for that to happen, why not study some math and some electronics.
Ideally the unflitered / prefilter quotient output should _always_ be DC except for the noise.
Both the numerator and denominator go to zero at the same time so the noise will be large compared to the denominator when both the signal and ref both go to zero.
The solution would be to only sample outside a certain +/- angle when both the signal and ref are zero, say, say +/- pi/4 at 0 radians and
The best place to look for a signal is where you know it is. The worst place is where you know it isn't. Mathematically, we're looking for the signal that best cross-correlates to the input. For a sine wave, that turns out to be another sine wave. I sure ain't the reciprocal of a sine wave, which doesn't exist mathematically, much less physically.
Let's start with something simple:
{ Limit (N/X) as X==>0 } ==> infinity for nonzero N.
Division is messy, mathematically and in real electronics.
There are lots of books that cover "signals and systems." It's standard 2nd year EE stuff.
Only if the noise occupies a higher frequency band than the signal. If that's the case, your capacitor on its own will filter out the noise, and you don't need the lock-in arrangement at all.
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.