Differential amplifier basic operation and negative feedback
ฝัง
- เผยแพร่เมื่อ 4 ส.ค. 2024
- Today we take a look at the differential amplifier or long tail pair as it is known in this version. It is commonly used and not well understood in its basic operation.
- วิทยาศาสตร์และเทคโนโลยี
The notch representation was a great visual aid. Thats what I've been missing and i feel like i finally understand this topic. Thank you!
WOW! Many hours were spent on circuit analysis trying to understand this first stage of the audio amp, primarily for the troubleshooting of dc offset problems in vintage Kenwood KR stereo receivers that do not contain DC offset cal pots. The old Kenwood’s relied on a closely matched differential pair (hfe gain matched) and associated calculated components (caps/resistors) for the biasing of the pair. A drift from design specs in the bias circuit caused by components degrading over time very often results in a hi (100mv or so) DC level (DC offset) ending up at the speaker terminals and sending the amp into protection. Very often the transistors start to leak beyond the original design, unbalancing the complementary pair and making feedback compensation non-linear. The concept of this video is a necessity for the understanding and repair of any of the GOLDEN AGE (77'-82' Marantz, Pioneer, Sansui, Kenwood) receiver amp linearity problems. Thanks a million for making this video!!
Great work sir, your feedback explanation is so clear and easy to undetstand. please keep up the great work.
i started watching your videos on audio application related issues, but then also theory topics like this one to refresh some knowledge. i had this explained several times in tech high school and engineering college. i guess your explanation was as accurate as the college ones, and far easier to understand than the high school ones. congrats!
Very nice seeing a Snickers flashback, RIP.
Good info John.
Excellent, at 11:30 and on is what I was searching for an could not find anywhere else - explanation on how differential amps and feedback works. Thanks!
Thanks, this has shed so much light on something that was a complete mystery to me! I have built amplifiers from schematics that I found, but never really understood how the differential input worked or why they would use it.
CoolDudeClem Hi Clem. I'm glad this has helped you out. I watch your videos as well. Waiting for the next "Tube Time"!
Hey, Clem, man!
Wow, this last example with notch is very good :D. Thanks man
Your Explainations are easy to understand and i am glad to be a Subscriber. Thank you for your good work! :-)
Elnufo Many thanks for the kind words.
Not even halfway through it made so much sense lol. Thank you!
Great explanation, Thank you!
Understood this amplifier topology well, I always just called it "totem pole" but I enjoyed watching you tell us about it.
The long tailed pair is one of the more interesting electronics building blocks I feel too. It has both AC applications, but very intriguingly also DC applications with instrumentation and control circuits.
Extremely useful circuit topology
th-cam.com/video/63pGZRff_lU/w-d-xo.html
Such a perfect explanation. Thank you.
Hi John,
Yes you did explain it very well and 100% accurately. Nice to be the first to comment, Might be nice to show positive feedback, nothing like a screamer!! Reminds me of my valve days, trying to find which way round to connect the feedback from the output transformer. Should be 50% chance of getting it right but the result was usually positive feedback. Good work I always enjoy for vids.
Michael Beeny Hi Michael, I'm sure I'll be dealing with phase shift and oscillation when I try to get that amplifier going as a finished project.
this is how it should be explained. very good, thanks.
Ok I finally understand how negative feedback can correct distortion. Thanks.
Wow.. u demystefied this for me.. lots of thnks
It really helped. Thanks..from india
Very nice! Curious to see you replace the two resistors in the Long Tailed Pair with a Current Mirror... and see that on dual trace scope...
Thank you so much
I love your videos. What's the difference between a current source and a current sink?
wow, learning every day ta
As of the feedback, is this a series - shunt configuration ?
Hi John,
Your video was very clear and made me want to try building a similar circuit. I'm new to audio circuitry. Can you reccomend any good books on the basics of amplification?
Thanks.
Adam Morris "All About Circuits" is a very good site with a chapter like layout. Rod Elliot's ESP site is mainly audio based but may be a little deep for the beginner. I think he has an intro section. As far as actual books, your local library usually will have a few titles on the subject.
I'll check out both. I've heard of all about circuits. It's supposed to be a great site. One more question. I have a usb powered oscilloscope/logic analyzer I use for low voltage microcontroller stuff. (I'm working on a degree in computer engineering). It says the device is good up to 30V. Do you think it would be safe to use for looking at amplifier circuits? It's a digilent analog discovery.
Thanks again for getting back to me.
Adam Morris You should be fine with the analyzer on most solid state lower powered amplifiers. Higher powered amps can have voltages above 30v. Tube amps can have several hundred volts on their anodes, so I would not use it there.
First thx a lot. It is very well illustrated and demonstrated. One suggestion could be to indicate eventually with a textual comment that the "noise" event is not conceptually a isolated event but a kind of "reccurent" event (I imagine that this is the case if we talk about distorsion that adds up to the signal) . So the diff will compensate not just ONE event but a sequences of an event (I.e an undesirable frequency or recurring event). Is my suggestion making some sense in essence ? Do you envisage to enrich when necessary your videos with text. It is sometime very good tool to vehiculate a more complete/correct info. Brgds.
I was very confused on this point actually. I think that would be a good addition to the video, which is great by the way. How does one make sure that their negative feedback is in phase with the noise in the signal? How is it that the true signal isnt negatively fed back as well?
Very good vid. I have seen these amplifiers, many times, on the input stage of push pull amp and tried to understand what they do. Your vid explains very well to me. I am guessing that you have to end up with some error? if you completely fix the error there is no error fed back to the differential amp, so no corrected wave form will be sent to the power stages, so your error will return. I guess you would get the oscillation that you mentioned.
Am I right or talking rubbish?
richardcloudbase The idea is to design the amp to have as low as distortion as possible before relying on negative feedback to correct. NF does have limitations in how much the distortion can be corrected.
Is that an Elenco FG500 you're using?...If so how good is it?
ric2play It is okay for generating waveforms from 2Hz to 100KHz, but distortion is too high for any serious measurements.
Thanks for replying.....
I had a question about your explanation of negative feedback at around 15:00. Wouldnt the signal being fed to the right most transistor be passing through the speaker at the exact same time? In this case, how is the inverted version of the signal correcting anything? Wouldnt it be too late for the signal to be corrected by the time its fed into the transistor? Any information would be a great help thanks.
Also how is it that the negative feed back doesn't also cancel out the true signal?
Amplifiers work very fast so the signal from the feedback arrives at the inverting input at the same time to be in phase with the input signal (the action of the inverting input makes it 180 deg. out of phase). Nothing is instantly fast in the real world so a phase shift develops at high frequency between the input and feedback signals. If this phase shift is large enough, the amplifier may oscillate. Something called frequency (or phase) compensation is used to reduce this phase margin and keep the amplifier stable.
Yes, the feedback cancels out much of the gain. A resistor divider network is used to reduce the feedback signal so that the amplifier has some gain yet yields much of the "error correction" that feedback offers. If the the full output was sent back to the inverting input, all the amplifier's gain is canceled out.
You are right - this youngster did not understand complitly what Feedback is about. I have made a new commnt that corrects.
He's 48, mate.
Electricity in circuits travel to the speed of light which is much faster than any electronic audio signal ( wavelength ), so there is no delay in that sense
wouldnt you need to delay the initial signal before a second parallel amping stage that would be the one corrected... if not it mean you correct a future waveform with a past one.. lots of engineer praise single ended with non negative feedback..
That is best Radio Shack Breadboards you have.
Advice: 1. Lock down your camera. 2. Use a pencil to point at things, not your fingers. 3. Scratch kitty's head, that's what (s)he wants.
while the concept of negative feedback is correct, the phase of the "notch " is not consistent with the phase of a sine wave at the collector of Qin, input transistor ( connected to the summing junction).
Perhaps showing at Qin - E*sin_wt caused by the input signal plus:
+-beta*E*sin_wt caused by the feedback is closer to the truth .
He did say he was going to explain it as simply as possible so any one could understand. Using all that engineering mathematics would over complicate it.
I never understood how a distortion that has been presented to the speaker in the first place, can be sent back and get corrected without the speaker 'noticing' the distortion.
The correction happens instantly as far as audio frequencies are concerned so the wave form coming out of the amplifier is already corrected.
Think of it like this: TV/movies the way the frame rate works is based on the persistence of vision so we don't notice that they're a bunch of still images made to move with a shutter (we don't notice the flicker on film any way). Hearing is kind of the same. We don't tend to perceive minor distortion (as it's being corrected). However, if there was no negative feedback, we would indeed hear it, just like at Brian May from Queen, who used a Vox amp most of the time for his solos with no negative feedback. It gives a kind of warm early clipping.
the feedback does not correct the waveform only the amplification level. Notch example 13:34 wrong ! The feedback will correct changes caused NOT by the input waveform itself (till a certain limit). Yes it will reduce the NOTCH but it will not make it disappear. The feedback will keep Amplification Constant by the price of reducing total ampliffcation(ß)
Understood. NF can only correct errors inside the feedback loop and the correction is not 100%. The less feedback provided results in higher gain but less error correction. This is a basic explanation.
Yes , you got me right. Here is some more that will explain the whole story. Take as an example simple circuit like transistor in common emitter connection with an emitter resistor. When transistor gets warmer base emitter will conduct more and will cause higher Ic (collector current) . The effect will be higher voltage drop on the Re resistor and that will cause that the Vbe will get smaller ( More voltage drop on Re= less voltage difference between base and Emmiter) and that is the self correction to the effect of HEAT. But if the Signal will have the notch you talk about it will still be amplified as the rest of the signal. Now to something else - in your video you show the Common tailed pair (differential amplifier) these are used at the input stage of any op amp. one input is (-)inverting and other (+) when SAME signal is applied to both inputs they will cancel each others affect on the Re resistor. Let's say a cable (coax) runs from one city to another with a signal. On the way it collects noise (from other electric machines). When it arrive the second city and it is being connected the inner cable signal + noise to one transistor and the shielding of the cable (outer part) to the second transistor. the two common noise parts will cancel each other (on Re) what will remain will be the original signal that was sent from city 1 . That is probably what you've red about "negative Feedback fixing the signal". Wish you fun with electronics.
@@yoramstein now you're talking about CMR (common-mode rejection), which is a subject more to do with reducing noise in balanced systems. This really isn't relevant to an LTP being used as a front end to an audio power amplifier.
least conceps covered explain more math.
You realise that there is a LOT of math involved in this deceptively "simple" circuit, right? It over complicates explaining it's operation "simply".
@@AstrosElectronicsLab you require it when you work with it.
@@lakshmitejas1685 only if you're an engineer. Building them from kits or just repairing them, you only need a basic understanding of how they work - which is what this video is aiming at.
@@AstrosElectronicsLab ok sir thank you