Hello,
Could someone explain the relationship between RC filter roll off steepness and overall response time? Lets say I have a signal at 10khz from a photodiode and I want to block stray 60hz light (not emi). If I add an RC high pass 6db/decade filter after the transimpedance amp, how do I figure out what the response time of the circuit is for the 10khz signal? In other words, if I pulse the 10khz off and on (100% AM) how can I predict how fast I can pulse it on and off?
Thanks,
Asa
Greets.
They are inversely related. The infinitely steep filter takes forever to produce a response.
That's a big subject for a newsgroup article. In general, you figure out the transfer function of your filter and compute its time domain response to allow you to apply your own criterion for what the "response time" will be considered to be.
I've never seen a 6dB/decade filter, so I cannot help you with that specifically.
A crude rule of thumb is: For each pole, determine its position on the real axis expressed in radians/Second. Compute the inverse of that value as Tp. Root-sum- square all the Tp values. Adjust according to whether you were looking for 50% risetime or 5% settling. Add salt to taste. For more accuracy, you have to do the transfer functions unless your filter has been taken from a cookbook or equivalent CAD tool.
Good luck.
-- --Larry Brasfield email: donotspam_larry_brasfield@hotmail.com Above views may belong only to me.
Fred Bloggs will probably have something to say about this - he has been critical of Larry Brasfield's responses in the past, and this should defintiely engage Fred's critical faculty.
Effectively, the original poster is proposing to amplitude modulate a
10kHz carrier, and wants to know what a high pass filter will do to the modulated signal.What the OP needs to know is that AM multiplies the carrier waveform by the modulating waveform, producing a waveform whose Fourier transform contain the carrier frequency and upper and lower sidebands separated from the carrier frequency by the modulating frequency.
If the modulating waveform is a pure sinusoid, there will be only one upper and one lower side band. If it is a square wave, there will be extra sidebands corresponding to the odd harmonics of the frequency of the square wave, but the amplitudes of the sidebands decline in proportion to the harmonic number.
IIRR 100% modulation suppresses the carrier frequency. There a way of suppressing either the upper or the lower side-bands - search on "single-side-band" - but that isn't relevant here.
So the OP should be able to modulate the 10kHz carrier at up to a few kHz before the sidebands get anywhere near the cut-off of a high-pass filter for blocking 60Hz.
----------- Bill Sloman, Nijmegen
If I Recall Right?
If you use AM terminology 100% modulation of a carrier means the negative modulation peaks go to zero volts and the positive modulation peaks go to 2x the unmodulated carrier voltage. With sine wave modulation this puts 1/2 the total power into the carrier and 1/4 in each of the two sidebands.
Suppressing the carrier requires that you multiply the carrier by the signal -- so if y is your output, c is your carrier and x is your signal then you have y(t) = c(t) * x(t). Traditional AM says that you limit |x| to be no greater than 1, then get y(t) = c(t) * (1 + x(t)).
-- Tim Wescott Wescott Design Services http://www.wescottdesign.com
When you use the filter(HP 10k 1st order) in the signal path, its rise and fall time will be identical to the original signal, but the flat top of the square input signal will be differentiated and becomes like needles with
*almost double* the amplitude. +------+ + | | | | | | | | | Input | | | | | | + +------+ || o-||----+------o + + ||C | | | .-. | | | | | | | |R |\\ | '-' | | | | | o-------+------o | \\ | | | | \\ | | \\ | | \\ | | \\+ | | | | Output + | /+ | / | / | / | | / | |/ | | +It might be that your receiving logic interprets this as a signal with double the frequency. You then have to lower the pole frequency to get more straight top lines. There will be no delay, unless you filter out the high frequencies with an additional lowpass. If your driving signal is a sine wave, it effectivly zero-crosses even 12us before the original, due to the differentiation!
-- ciao Ban Bordighera, Italy
6dB/octave ( which I assume you mean ) - 'first order' filters have a fixed damping / Q / roll-off and hence response.
The issue you mention of compromise is only relevant to higher order filters.
To examine the response time - you'll need to know if the source signal is indeed perfectly 'pulsed' implying an infinitely fast rise time - or indeed consider the natural rise time of the signal being filtered. It's not related directly to the 10kHz btw.
Graham
Perhaps if you were to state what was incorrect in my post, we could discuss the issue(s) intelligently.
The OP in fact claimed 100% amplitude modulation and wanted to know how fast he can detect that.
And what we need to know is what his detector looks like, including whether it has a fixed threshold or somehow adapts to the signal level when present.
If the threshold is fixed, and set near the lower limit of the detectable light signal, then for significantly higher signal levels, the transient response of the filter and how that interacts with the demodulator or detector will be of great interest to the OP. He mentioned "overall response time", but with too little information to know how a 1 pole HPF contributes to that.
In all likelihood, the modulation is effected by an object which occludes a light beam. The OP mentioned that his signal comes from a photodiode.
That would engage critical faculty if not already well covered by Mr. Wescott. The carrier is not suppressed.
Correct.
You have too little information to proclaim that. What you do not know is how much rejection the OP needs at 60 Hz in comparison to the passband gain. For the presumably 1 pole highpass mentioned by the OP, its rejection will depend on where the corner is set. For example, if the corner is set at 6 KHz, he can expect
40 dB of rejection at 60 Hz. If he needs 45 dB of rejection, response to a few KHz of modulation is clearly going to be affected. In modulated photo- current detection schemes, (which are used to get greatly improved sensitivity and LF rejection), it is easy to believe the OP may need more than 50 dB of 60 Hz rejection.From the OP's evident concern with "roll off steepness", it seems he understood these issues. That is why I made an effort to answer the question he actually asked and brought out the vagueness of his "response time".
Of course, I could have engaged in more guesswork about his circuit and requirements, then responded as if I already knew them. I prefer to let OPs get their real problem better established first, usually with some questions or by identifying open issues. Some folks here take that as reason enough to add some heat, but I have a hard time caring about that.
-- --Larry Brasfield email: donotspam_larry_brasfield@hotmail.com Above views may belong only to me.
"Fred Bloggs" wrote in message news: snipped-for-privacy@nospam.com...
[The topic is, roughly, about rejecting "60hz light" in some kind of system using 10 KHz carrier with AM.]
interference powered by a 60Hz source will actually be at
Actually, I spent 2 years designing systems that used modulated light to detect beam interruptions of various kinds. Because they had to be rejected, I've meaasured the (interfering) signals present in ambient light from quite a few sources and designed a modulation scheme to avoid their spectra. Informing the OP about the 120 Hz to be seen in many countries while ignoring the higher harmonics produced by fluorescent lights would be more misleading than leaving the subject alone.
While it is true that 120 Hz interference is likely present, many other harmonics of 60 Hz, including odd ones, are present in the real world and must be dealt with in such a system designed for high sensitivity.
-- --Larry Brasfield email: donotspam_larry_brasfield@hotmail.com Above views may belong only to me.
news:gcz9e.92$ snipped-for-privacy@news.uswest.net...
I'm not going to make your arguments or decide what your issues are. Your beef is essentially that I did not read enough information into the OP's question to allow his real questions to be answered. There is no practical limit to where that could go if we are willing to make enough assumptions about what he wants to achieve.
function of the filter. Depending on the - unspecified -
the output of the detector.
That looks like noisy agreement to me.
(Assuming you mean "kernel of the problem"), I think it has everything to do with the OP's response time concern. If he has a simple detector, say one that responds to an LPF'ed absolute value of its input, the transient response of the highpass filter will be crucial. If he has some kind of demodulator, reduction of the lower sideband could also be important.
just a capacitor with a loading resistor to ground.
That is a 1 pole HPF. Is your problem that I did not mention the zero at 0?
sideband if the 10kHz is modulated at 2kHz.
That depends on where the pole is. You again assume that it is set below the lower modulation sideband.
chopping disk, which produces a square wave modulation
transition times in most practical applications, so there
Other, equally plausible modulations are step or impulse beam breaks or beam appearances. I see no reason to assume away the OP's issue. (For all we know, he has built something that is too slow and merely suspects the HPF with or without adequate basis.)
current drive.
Not if the information you wanted was beam occlusion.
[SSB stuff cut.]have to do quite a lot of work before this is something
That, too, is obvious. But I think you are missing my point about where the passband of the HPF begins. That is going to affect the detector response near the limits of its speed.
the frequency domain, which happens to be crucial
So, passing up an opportunity to expound upon a topic is evidence of ignorance. With that kind of logic, just about anything can be known.
talking about a single-pole RC filter understands
Agreed.
did, many years ago - you might have posted less about
I also have tackled that problem and similar ones. That was about 25 years ago, but if I had wanted to post what I know of the subject, there would have been a lot more. As for guessing what the OP needs to know, I generally try to avoid going overboard there. Over and over here, I see long forays into aspects of assumed problems which have no apparent interest for the OP. It reminds me of clucking in a henhouse long after a noise has set off the process and ceased.
-- --Larry Brasfield email: donotspam_larry_brasfield@hotmail.com Above views may belong only to me.
It doesn't suck light in on the negative half-cycles?
Best regards, Spehro Pefhany
-- "it\'s the network..." "The Journey is the reward" speff@interlog.com Info for manufacturers: http://www.trexon.com Embedded software/hardware/analog Info for designers: http://www.speff.com
Speff, (may I call you "Speff"?) - you know better than that. For that, you need to put a Dark Emitting Diode in antiparallel.
;-P
-- Cheers! Rich ------ "clitoris, n: A haired trigger."
Nah - that's what drunk driving is for.
-- Pig Bladder on a Stick
"Fred Bloggs" wrote >> "Fred Bloggs" wrote in
interference powered by a 60Hz source will actually be at
That is provably false and your willingness to state such falsehoods shows how little regard you have for truth.
And I will correct your errors and misleading posts when I see fit.
[derf]-- --Larry Brasfield email: donotspam_larry_brasfield@hotmail.com Above views may belong only to me.
And the idiot, Brasfield, apparently misses the fact that illumination interference powered by a 60Hz source will actually be at 120Hz and not
60Hz."Larry Brasfield" schreef in bericht news:gcz9e.92$ snipped-for-privacy@news.uswest.net...
If you can't work it out from what I did post, this seesm unlikely.
The depth of the modulation doesn't make any difference to the transfer function of the filter. Depending on the - unspecified - nature of the detection system, the depth of the modulation may well affect the output of the detector.
All this would indeed help, but doesn't get to the kern of the problem.
The simplest high pass filter - which seems to be what the OP had in mind - is just a capacitor with a loading resistor to ground.
A 10kHz signal will go straight through it, as will the - say - 8kHz lower sideband if the 10kHz is modulated at 2kHz.
When I used to do this sort of stuff, the standard modulator was a rotating chopping disk, which produces a square wave modulation if the blade cuts the beam at a focus (rare) and something with slower transition times in most practical applications, so there isn't too much energy in the higher harmonic side bands.
Today, I'd expect to use an LED or laser diode source and modulate via the current drive.
So much is obvious. Granting the level of the question, the OP is going to have to do quite a lot of work before this is something that can be appreciated.
Missing the point that the OP didn't understand what modulation looks like in the frequency domain, which happens to be crucial here. You don't seem to understand it too well either. It is also questionable whether anybody who mentions "roll-off steepness" when talking about a single-pole RC filter understands much about filter design.
If you'd had to tackle the OP's problem yourself, as a rank beginner - as I did, many years ago - you might have posted less about what you thought you understood and more about what the OP needed to know.
---------- Bill Sloman, Nijmegen
There ya' go! Dark Emitting Antiparallel Diodes.
-- Keith
"Fred Bloggs" a écrit dans le message de news: snipped-for-privacy@nospam.com...
Nope. He was of course speaking of polarized light :-)
-- Thanks, Fred.
"Spehro Pefhany" a écrit dans le message de news: snipped-for-privacy@4ax.com...
Like an oscillating black hole?
It would be a cool gadget to have exposed, for those liking Chineese puzzles:
Your curious friends just grasp hold of it and then are sucked into. The purpose of the game is to escape out of it. Of course you'll have to show at least some glimmer of cleverness.
The nice side effect is that you'll definitly get rid of the stupid ones.
-- Thanks, Fred.
interference powered by a 60Hz source will actually be at
You're a crock of shyte- you never designed anything professionally in your life. You are an ignorant muddle-headed jackass who screws everything up you get your hands on. And *I* will decide what *I* tell the OP- not you. You have no credibility whatsoever- you are an ignorant and pretentious pseudo-intellectual p.o.s.- totally worthless by any reasonable measure.
You can go ahead and try- but all you have done so far is make an ass of yourself- you are inferior garbage it turns out.
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.