Differential Amplifier Experiments
Вставка
- Опубліковано 2 лют 2022
- This is a follow-up to the Marantz 2240 Video. I didn't really do a very good job explaining the differential amplifier section and why I made some changes, including the addition of a constant Current Source. Hopefully this will give a demonstration of the diff amp and how it works with a resistor "tail" vs a CCS "tail". We also touch on how the feedback signal works in an amplifier.
Remember, there are many different ways to configure a differential amplifier. Additionally, there are many things that affect the performance of a diff pair including the characteristics of each transistor, temperature of the transistors, component accuracy, noise ingress/shielding and design variables.
In case you are interested, here are the actual values of some of the components used in the experiment:
Transistor 1: 2N2222, Vbe 626mV, Hfe 191
Transistor 2: 2N2222, Vbe 624mV, Hfe 193
Transistors were tested using one of those low cost component testers. I didn't do a proper curve trace or use my matching fixture.
Collector resistor 1: 10k ohms (actual resistance was 9755 ohms)
Collector resistor 2: 10k ohms (actual resistance was 9892 ohms)
Tail resistor: 5.6k (actual resistance was 5570 ohms)
CCS was a 2.5mA CCS - Розваги
The lightbulb went on when you said the second signal was from the power amp and thus the global negative feedback!!! I have watched so many videos and yours finally stated that and cleared it up. Thank you!
I love the Theory with schematics. It definitely builds my understanding of these types of circuits. Having the Demonstration with the oscilloscope really drives home.Please keep it up.
Just remember a lot of the components in the 741 are there to compensate for the fact that making high value resistors in silicon is hard, and also makes them very large, but you can make transistors do the same job, with better stability as well, because silicon resistors are really bad at absolute value, and drift badly with temperature, but really good at matching. Thus all those transistors are there to make them behave like large value resistors, and also make the circuit not vary with supply voltage, temperature and time.
Great video, I've been struggling with how this type of circuit works and you have helped greatly!
Appreciate your hard work to simplify. Thank you very much .
Fantastic video. Been staring at an SU-7700 schematic for days and this cleared some of it up for me.
You are a very conscientious teacher Tony, thanks for sharing.
I suspect looking over your shoulder at work would be very interesting too, but I understand the nature of the proprietary work you do and why you can't share that.
Great verbal explanation of an abstract yet useful circuit. I had to fiddle around with diff amps in a simulator for a while before I really understood the importance of a good CCS circuit. Basically I did what you did on the test board, but using TINA (circuit simulator) software.
There are two inputs, V+ and V-. However, we find it more useful to consider the combination signals V_D = V+ - V- (difference) and V_cm = (V+ + V-)/2 (common mode). Since V+ = (2 V_cm + V_D) /2 and V- = (2 V_cm - V_D) /2, these are entirely equivalent perspectives; however, the constant tail current makes the response to the common mode signal minimal so that the amplifier responds almost entirely to the difference.
Great video! Always appreciate your presentation method. Very helpful and enlightening. Your room full of equipment is nothing short of incredible and causes no end to salivation every time I tune in ... :-) Don't know why but constant current sources have always baffled me, ie., how they work and how to design them. Mostly how to recognize them when I come across them. Any thoughts of doing a video on CCS as a main subject?
Very interesting. It's also worth noting that if you apply the signal to one input and ground the other, the first transistor will work as a common emitter amplifier stage, and the other one as a common base stage fed from the emitter. That's how long tailed pair phase inverters in tube amps work too, except they typically use a pair of triodes.
sweet 😃
This was great a fascinating tutorial...cheers.
Great video as always. Thank you
this topic is i want all about. thank you for your video.
Thanks, Tony! Much appreciated.
Great work. Just over my Head.
Your the master
Great video. Nice demo of the differential stage of an amp. Next step would be to demonstrate how to remove the distortion :-)
Hope the snowstorm does not cut of your electricity...
Awesome tutorial...cheers.
no need to apologise thanks for the edutainment!
Thank you sir... For good content
Big big thanks again Tony, very appreciate ,if you have that kind off explanation for current mirror it could be cool.
Merci
R
Thanks from Columbus Ohio USA
this reminds me of that strange stereo tube amplifier you were doing a video on a while back
Great job explaining, thank you. I want a low noise, diff in, diff out OpenPlay that will work with a dual 15 V supply. All I find are dual five Volt. Any suggestions please?
Hello. Very interesting demo / idea of a diff amp in an amplifier or a receiver. However may I ask if these circuits can then used in Muting circuits of an amplifier ?
Thank you
Hi, Very clear presentation of the differential amplifier circuit. However, you never really said why would I have it in a high-quality stereo circuit? It would seem that a single transistor is all that you need on the front end and then more powerful transistors (able to handle high current loads) leading up to the speakers are all that is needed to generate sound. Why the feedback for this circuit? That part was not clear. Thanks for any help!
I think the 741 made people want to use current sources and current mirrors automatically. In an ic Op amp they reduce the need for bulky components. I gave the example of the Hitachi design. You might give it a current mirror to the input long tail pair. That might cause new problems requiring capacitors added to the second stage. A lot of these ideas evolve without real need. The Hitachi had fantastic sound. I added a current source to the Hitachi input long tail pair. No obvious difference. 47K 55V. 2SA 872A. 2SD756 second pair. 2SB716 and 1N4007 hybrid current mirror.
Thank you!
I remember the 741 back when they came out in the 70s .. they are now “classic/vintage” .. they have great gain, but as an audio amp, you could hear hiss …
A quad version of 741 was used in a Soundcraftsmen EQ. They are NOT very "musical". I replaced them with an LM837.
2:59 - So, if I want to use an LM334 instead of R703 (6.8k) at the top, will it give more linearity to the pair?
And could I use another LM334 in lieu of R704 (1k) at the bottom, but not for R705 otherwise they will cancel each other out?
Actually, a video explaining the use of this small but handy CCS device in front-end amps would be nice. 😀
Edit: just saw the previous video of Marantz 2240. Thank you.
wow, cool oscope , my 1st oscilloscope was a Leader gee are they still making test equipment? 🤔 thanks for the interesting amp stuff 🥳🥧☕
What is seldom said is using a long tail pair differential amplifier for the inclusion of an extra transistor makes design of an amplifier vastly easier. If matched to a further long tail pair as in the 1979 Hitachi mosfet design a very high degree of performance is achieved. Hitachi use a current source on the second stage. 7 active devices 100 watts 8 ohms. 2 % current error maximum. In an amplifier the input long tail pair won't swing much voltage , mostly current. Simple resistors do surprisingly well. The long tail pair works almost like a bonus computer maintaining DC conditions. The Hitachi had extremely low distortion and behaved much like an Op amp.
just wondered what the benefit if any would be if you split the tail into two going back to the power source? I could see a benefit if you had two transformers but is there a benefit using the one?
Where does the negative feedback come from (Marantz diagram)? What determines the correct amount?
Can it be said: the amplification is in differential mode, and the feedback is common mode. Or is it, you use a differential amplifier with common mode feedback, and ccs.
Hi Tony I'm having difficulty with a Kenwood KR-5010 and I think it's not working because I can't source a new TA-1001. I got two from ebay and both have the same problem. Have you ever worked on this model? I already changed both STK-0050 Darlington ICs.The old ones were shorted. The original TA-1001 voltage amplifier looks a little different but has the same number. I took one that didn't work from ebay and pried it open and nothing looked burned and the transistors inside do not measure the same way that a typical bipolar transistor does. The speaker relay does not engage because there is -10 volts on the output pins on both STK-0050
I haven't serviced a KR-5010, but I would first make sure that the STK-0050's are working properly. If you are sure they are good, check pins 5,6,10 & 11 of the voltage amp IC. they should have around 1- 1.5 volts on each pin with reference to ground. If they are off, then check pins 2 & 14 of the voltage amp chip. This is the offset adjust for each channel. I'm not sure what they should be, but you should have voltage on each of them and it should be adjustable. You may have a problem with the voltage supplying the offset. You could also have an issue with one of the power supply rails. Hope that helps.
@@xraytonyb thanks Tony
Hello Tony how are you. I ordered an STK-3042 and replaced the TA-1001 that dalbani claims to be original but I think they're Chinese knock offs. Dalbani sells them on eBay. I emailed them about the problem and they sent me another one but I didn't bother installing it I already know it won't work. The pins are shorter and thicker than the original so I know they are not OEM. The STK-3042 came in and that one works. The output of the STK-0050 will now respond to the offset adjustment like it supposed to and the protect relay now can engage the speakers and works very good 👍 instead of being stuck on -10 volts and no response from the offset adjustment with the dalbani TA-1001. The STK-3042 also has the same looking pins that are small like the OEM. This was an expensive learning experience. The dalbani IC cost $32.00 and the STK-3042 was only $10.00 so if you ever work on one of these beware. I like the earlier descrete transistor receivers much better. I don't care for IC type they always challenge me more than I like because of the OEM parts no longer available. I searched on Google about this receiver and finally saw a UA-cam video of a technician who repaired a Kenwood and shows it working with the STK-3042 so I gave it a try.
Where was this 4 months ago when I was noodling on fixing an old KR-9050? This exact issue was confounding me. Granted, probably should have worked on something simpler to start learning.
Where does the negative feedback come from (Marantz diagram)? What determines the correct amount?
Years ago I played around with differential amps and I remember buying a device that was two transistors in the one case; i.e..there were six legs to it and the idea was that the package offered thermal stability between the two transistors. I can't remember who made them now.
They are still made. They are known as a "matched pair" of transistors. Sometimes they come in a DIP package with several matched pairs in the package.
@@richardgray8593 Thanks for the info, Richard.It s been some decades since I've played around with them.
Has anyone tried implementing a differential amplifier in DSP or on the Raspberry Pi IV using a package called Pure Data.
Hi, Can you recommend me a good book on differntial Amp and op Amp, pls?
Book: Op Amps for Everyone - Mancini. Published by Texas Inst.
Thank you very much....
Sir, you are a genius I really appreciate your explaination, sir I have a SAE 2500 ch A blown could you please give me your valuable advice how to fix, pl email.
is it not better to add two more transistors and make it a capacitor multiplier? smooth out the ripple.
This is why God invented Op-Amps 😁
This could be done much better with an Opamp.
Wow as soon as I finished typing that, you brought out the Opamp circuit.
which one would you use on this application?
@@HipocratesAG TL072 is good choice.
@@sr.padilla1633 Cool I think I have some of those on stock to experiment a little bit
Actually the whole power amp is a discrete op amp. The negative feedback will compensate for some of the faults of the simple diff amp without the current source. Thank for taking the time to build and demonstrate the circuit.
Sorry Tony, but you may be mistaken. Current sources are used for the different purpose. They are used to reduce 120Hz noise from the power source.
I continue to witness witless "differential" amp circuits with a Buffer or Inverter for the UnBalanced Input directly feeding their circuit "differential" Output's + (or -, depending) also feeding an Inverter which directly combines with the former. et VIOLA!
This is certainly so for Polarity, but not in the Time Domain, correct, Maestro?
wtf?
I'm so dumb that I don't know, or at least don't remember, why you want that feedback -- what its purpose is. A word about that for the truly naive would have made the whole thing more meaningful to this software implementor type, no friend at all to the electron.
In this case - and this is just my assumption - the current from the final transistors will be fed as a negative feedback. So, the higher the current, the lower the signal will get. It could also compensate for thermal runaway. In the same manner, it could compensate for other factors so that the volume stays similar regardless of what happens (for example supply voltage fluctuations, temperature). It's basically used for protection or stability.
When incorporated into an audio amplifier, any difference between input and output signals (i.e. distortion) is canceled out. So you get much lower distortion than if you didn't have the long tailed pair.
Good question!
As the input signal passes through each stage of the amplifier, it can be affected by the components of that stage. Any change in the signal, other than its amplitude, results in distortion. That means that the signal coming out of the amplifier does not perfectly match the signal going into the amplifier, hence distortion. As we saw in our little demo, when we apply a signal into the second transistor in the diff pair, it can either add to or subtract from the signal at the first transistor. This is all dependent upon the amplitude and phase of the feedback signal. Due to the differences between the distorted output signal and the undistorted input signal, the differences would cancel out. While this doesn't totally remove all distortion, it will greatly reduce it, although the amplifier as a whole will have less total gain. (Gain of the amp without feedback is higher than the gain of the amp with feedback installed. This is why feedback-based designs always have a higher open-loop gain). One thing this won't fix is crossover distortion on the outputs, because the transistors can't pass a signal until they begin conducting at a voltage above their Vbe drop. This is why the amp has to have a bias current, meaning that the transistors never turn fully off and stay above that Vbe drop. Hope that helps.
@@xraytonyb I don't think you could make it any clearer. Your prior video caused no confusion to me with what you just reinforced with a visural demonstration. Thanks again for some of the best audio advice found on ut.
Thanks, Tony, and others, for the clear explanations. I greatly appreciate it.