So far I haven't seen that advantage that much. Very fast chips are usually very expensive, meaning they are out of budget for many projects. On the discrete front there are lots of blazingly fast SiGe transistors. Even some really old regular ones can still run circles around chips. An example is ye olde RF work horse, the BFS17A. It can ramp several milliamps in under a nanosecond.
But there is often a case for chips. I must admit that I have used the uA733 a lot in the old days. That was because volumes of it were built into VCRs so it became a cheap device. Whenever I use an IC I try my utmost to avoid a single sourced part. This is because I have seen too many people including clients get zinged by that. Heck, I even flung a SMPS design to semi-discrete once and even after a decade in production no PWM chip has come close in cost.
In ultrasound apps I'd say that all but two of my front-end designs were discrete. The two that weren't used the AD603 which is a marvelous chip but at around $5 it is too expensive for some designs.
Please! These words are not fair to Larry. His advices have help me last night.
I still have one thing not clear: why the signal from first two stages is so clean, but after the third one, it suddenly becomes so bad. The noise is not oscillation, nor interference. The high spikes of the noise keep giving false alarms to the detection circuit.
That's decent of you, Zhi. I urge you to apply a filter to Fred's remarks. Once in awhile, some facts can be found in his posts, and often those are correct, and somewhat less often they are also relevant. As for his comments about me, quite a few are provably false, so I give the whole lot no heed.
The upper band edge of your amp string is near
68 KHz. (4 MHz op-amp GBW, closed-loop gain of 50, 3 like stages) With your inter-stage coupling caps, the lower band edge is near 19 KHz. This leaves about 49 KHz of bandwidth in the -3 dB sense. The noise bandwidth is slightly different, but I'm too lazy to calculate that. (Maybe Fred will take on that challenge, since it is manly work.) Near your center frequency, the response will be about 14 dB down with respect to the ideal value of 50^3 you would get with higher loop gains and the HPF's not cutting in from the low end.
If you were to short the input, then, due to the thermal noise of the input resistor, you would have an equivalent input noise density of about sqrt(4 k T R) = sqrt(4 * 1.38e-23 * 300 * 1e4) or 12.9 nV/sqrt(Hz). Adding to that the input voltage noise of your op-amp, 16 nV/sqrt(Hz) typically, (and adding RSS-wise), you should expect input noise of about 20.5 nV/sqrt(Hz). If we were going for accuracy, a smidgen could be added from the feedback resistor, but its contribution would be lost in the errors already in this calculation. (Fred might be willing to enumerate or even quantify them for you.)
Taking the gain of your amp chain, which is about 24.7e3 in the middle of its passband, the input noise results in about 506 uV/sqrt(Hz) noise density at the output. Multiplying by the square root of bandwidth yields 112 mV of RMS noise. This last step is not quite correct because the bandwidth that should be used (the "noise bandwidth") is a little different, but it should be close enough to help you see whether your result is worse than should be expected.
With your transducer as the input, the noise could go up or down depending on its source impedance. That is why you would have to measure it or take it off the datasheet to see whether the noise you see with it connected is what it should be.
Once the input is better characterized, it will probably be easy to revise the first stage a little bit to get lower noise.
As for why you do not see the noise on the earlier stages, that would depend on your instrument. With a typical o'scope, not being able to see the 4 mV to be expected at the
2nd stage output is to be expected. And of course, it's even harder at the 1st stage.
If you peruse the rest of this thread, (with filters in place, I advise), you will find some comments about your op-amp choice that you may find useful. And if you care about input noise and power consumption, take suggestions about using CMOS gates as amplifiers with a degree of skepticism.
I have no idea what that is. Since, (apparently), you are reaching into the noise floor for signal, it may be worth your while to post the detector and solicit ideas for improving it. A tighter bandpass filter could do some good as well.
[Brasfield's AACircuit directions and link cut.]
[Fred's invective regarding such "advice" cut.] [Fred's unsubstantiated ad-hominem rant cut.] [Reply to noise filter output:] Nothing left!
--
--Larry Brasfield
email: donotspam_larry_brasfield@hotmail.com
Above views may belong only to me.
that is a tough one, have you closely examine the signal from the second stage to make sure there are no narrow spikes that are not being noticed? also,with that kind of gain and the TL082 having a BW of around 10 Mhz (last time i checked the specs, could be wrong), you maybe getting local AM or low band broadcast in the signal. have you decouple the second stage with a small Cap? have you used a cap on the 200K loop back bias R to make sure only aprox 100Khz of bw is being past ? etc. things to look over. the other problem could be board construction, you may have some kind of layout that could be causing oscillation and make sure you have a Cap at the voltage divider for the + inputs. and using a by pass cap (disc/no L type) at the Vcc of the chips to remove stray noise in the traces.
"Larry Brasfield" wrote in message news:m8k_d.27$ snipped-for-privacy@news.uswest.net... ....
I should have mentioned, to the extent the noise does not go up under this condition, your input is seriously mismatched to its source for purposes of SNR maximization.
That's understated. It's a near certainty, IMHO.
....
--
--Larry Brasfield
email: donotspam_larry_brasfield@hotmail.com
Above views may belong only to me.
He can use a dog-simple hex inverter out of that 4000 series you like so much- simple discrete FET source follower Sallen-Key 40kHz bandpass Q=10 ac-coupled into input chain and logic compatible square wave burst output.
Why don't you take your own advice and do the same , loudmouth. You can score this thread another zero contribution interference, with all your generalistic pseudo-theoretical bs and "I assume"s, the same old crap. You're nothing but a bit-head programmer ( not very good either) and bs-artist.
You are not educated enough to recognize that Larry is an incompetent and waste-of-time, but this is more than apparent to the rest of us.
Is the third stage in the package with the unused opamp? If so, this may be your problem, take that unused amplifier and configure it as follower with OUT to IN(-), then tie IN(+) to the 4.5V reference or "pseudo-gnd".
My bet is that Zhi is educated enough, in matters both technical and interpersonal, to correctly form his own judgements on my competence.
That suggestion may save a little bit of power. It does nothing for the noise phenomenon Zhi sees.
That was his idea. If you want to understand what is happening, study my post of 22:38 today and compare the predicted output noise with what you observe. If you use an o'scope for that, estimate RMS by putting cursors (mental or actual) at the levels that encompass about
2/3 of the fuzz, "measured" by brightness, (and record 1/2 the difference, of course.)
Nice, and thanks.
--
--Larry Brasfield
email: donotspam_larry_brasfield@hotmail.com
Above views may belong only to me.
I read in sci.electronics.design that for_idea wrote (in ) about 'amplify 40kHz audio signal using TL082: first two stages are fine, but high noise from the third stage', on Thu, 17 Mar 2005:
What is the peak-to-peak output voltage of the second stage? Remember, your third stage is trying to amplify it 20 times. So if it's more than about 200 mV, your third stage is overloaded, and if it's very much overloaded it may well produce spikes.
--
Regards, John Woodgate, OOO - Own Opinions Only.
The good news is that nothing is compulsory.
The bad news is that everything is prohibited.
http://www.jmwa.demon.co.uk Also see http://www.isce.org.uk
Actually you can do better than 10- I have some neat circuits with follower feedback into a parallel LC tap on input that can do much, much better than that- from the days of discrete precision tone encoders/decoders for telecom.
The only thing Zhi Yang has to keep in mind is a modest or at least predictable battery drain. If you bias at VCC/2 the current will vary quite a bit between different batches.
Agree. Same for building a discrete circuit with transistors. Before attempting any of that Zhi Yang would really have to be sure that he knows all the tricks in those areas. Or has a colleague who does.
To the OP and others looking for technical content: There is but one strictly technical fact here. See "1.".
"Fred Bloggs" wrote [in response to Fred's pontification to the OP: "You are not educated enough to recognize that Larry is an incompetent ..."]
Fred, it is your certainty about falsehoods that proves you are at least a very disturbed, neurotic individual, if not outright psychotic. Your obsession with me and my posts only confirms this. You need professional help.
out of the profession.
Out of three factual clauses, only one with any truth, and that available from what I have already posted. So it would appear the rest is baseless conjecture. How in the world, Fred, do you become so certain of "facts" you conjure out of this air? That tendency marks you as either delusional or a liar, or both.
pretense with little bits and pieces of elementary
the thread.
More bare assertion without substantiation. For those following this branch for amusement, the OP has asked about a final stage in an amplifier chain with enough gain to make its input noise visible on an oscilloscope, wondering why it is "significantly corrupted by noise". Fred is now upset because I have explained it (or pretended to explain it in his delusional perspective) when his own far-fetched diagnosis has fallen flat. His blathering about inapposite "theoretical factoids" serves to reveal the depth of his ignorance and delusion. If he was truly as knowledgable as he pretends, he would be capable of recognizing the simple validity and relevance of my post today explaining how the OP's noise observations have come to be. But he is not that knowledgable, so he is able to imagine that simply because he fails to comprehend my analysis and holds derogatory opinions of me, it must be irrelevant and silly.
The only problem with my analysis is that it explains all the OP's reported facts, relying only on well settled and accepted theory together with facts that were either posted by the OP or are publicly available (except for one [1]), *and* was not contemplated by Fred, the master of all things electronic.
[1. There was an unstated assumption about the ambient temperature. I offer this as bait for certain idiot(s) to grasp as grist for their continuing process of self-embarassment. ]
I note that Fred has not (yet) shown enough courage of conviction to actually attack my analysis on its technical merits, point by point. Somewhere, deep within his disturbed thought processes, he surely must realize that he cannot do that without showing himself to be even less competent in one realm of expertise than myself, a person he has obsessively denigrated as some kind of totally incompetent fraud. So, here we see the result of some very hard choices for him, a conundrum of his own making.
More false conjecture. I never attended EET. (Until I looked it up just now, I was unfamiliar with the acronym.) And I never flunked anything. Whatever led you to "believe" such nonsense? Delusional.
Wrong again. Amazing how you conflate the growth of some fantasy in your fevered mind with an ability to divine certainties. Delusional.
I could mention here many of the indicators of a reasonably well conducted and satisfying life, but pity prevents it. I have my own ideas about what Fred's life must be like, and if they are true, he could only become more disappointed with his existence upon apprehending its contrast with mine.
Fred, I wish you could comprehend how satisfied I am with your respect and affection. If you were able to understand that, there might be hope for you.
--
--Larry Brasfield
email: donotspam_larry_brasfield@hotmail.com
Above views may belong only to me.
Well here's something about which there is no uncertainty. You're a minor punk programmer and former electrical technician who was fired out of the profession. This explains your constant misunderstanding of the OP's question and this pretense with little bits and pieces of elementary theoretical factoids that you blow out of proportion and have no bearing on the thread. So, loser- tell us how you flunked EET and then transferred to some quick IT school and progressed along with your miserable and superficial life. You're the usual third rate scum I have come to expect from newsgroup...
As someone already mentioned its normal the more stages you add the more noise you get. More current noise from bias currents into the opamp flowing into the output resistors, voltage noise in the devices in the opamp etc... . Each of these noise contributions gets amplified in each successive stage and the noise sum also gets added on until your circuit starts giving off false alarms.
Your other question was how to improve the 3rd stage. How well are your resistors scaled against noise? Plug into the total rms output noise formulae the values for the corner frequency for voltage noise(from opamp datasheet) , corner frequency for current noise , the feedback factor, corner frequency for the closed loop gain function, plus the thermal noise resistor values. Calculate each noise component. if the resistor noise component is much greater then the other 2 then you need to scale your resistences. a rule of the thumb used for low noise design is Eno(V)^2 + Eni(I)^2 = 1/5*Eni(thermal)^2 I'm sure you can find these formulae on the web or in an electronics book. In case you can't let me know and I'll try and dig them up. It also helps looking at them too because you'll know what is actually increasing your noise and what to look for in a low noise Opamp.
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.