The unbiased audio amplifier output stage

[@hectorchavez1807]

That cat looks unbiased too!

[@FelixMartin-rk6eq]

i dont mean to be so offtopic but does anybody know of a method to log back into an Instagram account?? I was stupid lost the password. I would love any tricks you can offer me

[@FelixMartin-rk6eq]

@Stephen Marvin i really appreciate your reply. I got to the site through google and im trying it out now. I see it takes quite some time so I will get back to you later with my results.

[@FelixMartin-rk6eq]

@Stephen Marvin it worked and I now got access to my account again. I am so happy! Thank you so much you saved my account!

[@stephenmarvin5416]

@Felix Martin no problem :)

[@marttiinnanen4911]

Wow this guy is legit. An amp enthusiast who acknowledges that listening room, speakers and source audio are the main typical problematic things in the audio chain. Thumbs up!

[@fredfabris7187]

Very cool video, I like how you were able to clearly demonstrate the amplifier operation and how feedback is a powerful tool. Thanks

[@dantheman1998]

i love these experimental amps you build. keep it up!

[@isettech]

The correct name for that distortion is called crossover distortion. With a high frequency dither circuit on the input and an inductor based low pass filter on the output, you can transition into class D operation when you pick up the feedback after the low pass filter it will saturate the transistors so they don't run in the heat generating linear curve while still reproducing low distortion analog after the low pass filter.

[@salec7592]

In place of an op-amp, you could use a small power differential audio amp, e.g. LM386, or TDA2002, or whatever is hip today. Or, if using integrated PA defeats the purpose, make two stages: first stage biased, AB class amplifier, second stage your B class amplifier with current dump provided by first, AB stage. But I am probably stating the obvious and describing ubiquitous designs from decades ago.

[@MrBanzoid]

Damn! I was going to suggest the same thing. You beat me to it.

[@johnyang799]

That's a good idea. I thought about 1 month ago when i saw this comment. At that time I thought it's not that good because you need to add extra resistor and the slew rate isn't all that great with those. And they don't have superb distortion figure in the first place. However, I thought about more, this circuit has a lot of potential. Basically I am able to design a very low distortion high speed high slew rate high voltage amplifier (only 1A needed) instead of high current as well. But how can we close the loop? I can use composite amplifier topology, we add another stage in the front. This stage has to be only very very low distortion, and high open loop gain. So basically we can get best of three worlds this way by using opa1656 + lm3886 + 0.1ohm resistor + 2-4pairs of transistors. Also this way I can divide the current/thermal without complicating the circuit too much. Really a lot of potential.

[@johnyang799]

or opa1656 + tpa6120a2 + 0.1ohm resistor + 2-4 pairs of transistors for lower impedance. This way the performance can be even better.

[@davebutler3905]

Thanks for the excellent video. I've been playing with op amp based audio amps. Very interesting explanation of Quad 405 current dumping.

[@demoncloud6147]

Just tested this with LM358 (20mA) , BD135 npn, BD136 pnp, 2 Lipo (+/- 3.7 V rails) - it worked, about 1/4-1/2 watts power (guesstimate), low frequency makes more distortion, mid-high frequency (1kHz +) works fine which I experimented with equalizer while playing 'Soul to Squeeze'. It was fun

[@matthewday7565]

Now I'm recalling something that did the rounds once, the "non switching amplifier", the idea being that it was a class AB that ensured the opposite transistor always had a minimum bias, as even with a biased class B, as the drive increases at high frequency, the dead zone reappears due to the inherent delay of bringing the opposite transistor back into conduction. One approach I recall, used faster transistors and diodes to "switch" a sustain bias, but some others used a more elegant solution.

[@Theineluctable_SOME_CANT]

I have a simple circuit I devised, based on National Semiconductor application note, that can elegantly do that for you. Let me know if you want to see it and I will do a video on my channel for you.

[@HillsWorkbench]

Glad you re-visited this.

[@anindyamitra5091]

I have had seen this scenario in Medhi Sadhagdhar's Electroboom channel, at that time I wondered, how did he set up the unbiased push pull for such clear sound.........now its clear to me, thanks Mr. John.

[@clee2423]

which video is that ?

[@clee2423]

nvm, found it

[@JohnAudioTech]

I watch his channel but haven't seen that one yet. I'll have to check it out.

[@clee2423]

@@JohnAudioTech here you go kzbin.info/www/bejne/ZmfFpZqgdrtmfs0

[@anindyamitra5091]

@@clee2423, exactly, I was talking about this one, thank you.

[@woosix7735]

For the old video’s piano distortion, it sounds more like there is no dither, but quantization distortion instead

[@dhpbear2]

I believe the high offset voltage is cause by a mismatch in the input resistances. To make them match, replace the 22k with a 470k. Replace the 1k with the 22k. Now each input 'sees' ~22k.

[@Markoul11]

470k input resistor will pick up too much noise

[@nocturnal0072]

I've never seen anyone go beyond 100k on the input for consumer equipment. I have no clue about pro equipment. Also most run of the mill designs will use a coupling capacitor on the input. It could be DC bias from the source.

[@peteb2]

My career began the late 1970s at a State run national networked radio station where I started out to become a tech. By mid 80s all the tube gear was gone , I was now officially a qualified studio tech. Head office always supplied the technology from parts to whole pieces of gear & it was the same for the new general monitoring amps which were based on Quad 405 Current Dump. I dunno but because the Govt was paying for everything, the designers had a ball and these amps were the most complex things things I ever worked on. The design was over-engineered to the nth degree, called for 0.5% tolerance resistors everywhere, matched gain opamps, matched transistors infact the unit’s were a dog of a thing to repair but they sure were great performance-wise & left alone in a studio control room were reliable. Then along came the high power mosfet for audio devices and the current dump amps were DUMPED overnight in favour of a much ‘easier’ less costly approach. I must see if a can’t find any info out about this particular past C-D amp though for nostalgic reasons. They were hated at the time & with no Internet back then the data as in schematics etc on them wasn’t kept...

[@martinda7446]

There are thousands of 405s still out there, still supported by QUAD, and all info is available, it is after all from the most famous hifi company in the world (probably). (o;

[@radioengineering8247]

I have three of those BCNZ CD amps on my shelf, and I have the documentation for them as well. I just took a look at the schematic. I see what you mean about over-engineered!

[@peteb2]

@@radioengineering8247 WOW! any chance i could get a hold of a copy etc of that service data please. (Total nostalgia to me)!

[@radioengineering8247]

@@peteb2 Absolutely. Just need a way to get in touch with you.

[@peteb2]

@@radioengineering8247 i've looked every which way for something that can done in You Tube to turn on a valid email so you can see it. Google has it locked down it seems being as YT is theirs. Their email system is gmail.com so if you put (now edited out)@ infront then that'd work but then i'd need to come back a few hours later and edit this if you get my drift.... The things we do for our electronics hobby huh?

[@Luckless_Pedestrian]

Nice demo... informative. Thanks and keep up the good work....

[@JerryEricsson]

Sounded good to me. I was thinking the other day as I watched a tech rebuild a reel to reel machine. Back in 1970 I bought a nice Teac reel to reel, with a receiver and a couple of speakers and set them up in my bunker over in Vietnam. I had tunes that could make a jukebox blush. Well come the day when orders came down with my name on them, I had 3 hours to pack up and move north to a new unit. No way I could box up my stereo system and ship it home and no way in hell I could carry it all aboard the C-130 that would transport me North. The unit I was leaving was being deactivated, so I couldn't just leave it and hope to reclaim, so I had a quick bunker sale and got most of my money back out of the system. When I got up North, I never got another setup, just a small cassette/AM-FM radio and a portable TV so I could watch "The Captain and the Worm" on AFVN network. Well I was wondering why I never got another good set when I got back to the world, and it dawned on me, my damn hearing had been screwed up in Nam, in fact I went to the doctor about it after a few months back in the States and they found much of my high frequency hearing was flat. Well today I have hearing aides, thanks to the VA, and I can pipe music through them but the music just isn't there. So I guess my take-away is enjoy the music while you can because one day something may well take away your precious hearing and leave all those wonderful tunes as flat as the center of that amp you built. Thanks for the video, I did enjoy the learning experience.

[@adamp9553]

As far as I remember, additional resistors are added at the bases to prevent either transistor from collapsing at zero. A simplified class B at the beginning...

[@biologicalagent]

Adam P that’s biasing.

[@nocturnal0072]

Neither of you are wrong. Also most designs use an input capacitor to avoid/prevent DC bias. I'm know this circuit is far from its final form. Edit: noticed the capacitor was later added to the schematic, wasn't there at the start of the video. 😂

[@alexeyzaplavnov747]

Bless KZbin algorithms for finally bringing this onto my sight. How I didn't found you before! Now you've got a new subscriber :)

[@rich1051414]

For feedback circuits using global feedback, this is why I like using mosfets for the transistors in the push pull power stage. The impedance on the gate is so high the op amp can slew fast as hell. However, the quality of the output is then totally dependent on the op amp itself, but I like that. It minimizes the variables. Also, since the voltage and current are not linearly correlated, you HAVE to use global feedback if using a mosfet. The signal sent to the mosfets' gates will be highly distorted compared to the mosfets' output.

[@JohnAudioTech]

I've seen versions of this using mosfets. More complicated but it works.

[@rich1051414]

@@JohnAudioTech By 'this circuit', do you mean without feedback? I would really like the see a video on that if that's what you mean. Would have to be complicated unless you knew exactly what your load was.

[@KarlAdamsAudio]

Excellent and most enjoyable video - scoping the op-amp output reminded me of what happens in the precision rectifier circuit I used in my VU meter driver. Very impressive just how reasonable an unbiased output stage can sound. I wonder if the go-to op amps for driving headphones (like the venerable 4556) might be the go - or maybe several of them in parallel (or perhaps even a BUF634 between op-amp and output stage?).

[@molepistol]

i thought of exactly same circuit myself... but instead of a 5534, i used mcp 6022, and a triple darlington .... then i half biased the darlington, so there was still almost no current flowing through the main loop... now when given a signal, the high slew rate of the model used, handles all the distortion.

[@tubeDude48]

NCN & NCS is FREE music, and quite nice.

[@metallitech]

Thanks for sharing, this is great.

[@davebutler3905]

I agree with you about piano music. It is the most revealing source for reproduction limitations.

[@Starbuckin]

If you run the Op Amp and push pull transistors on single supply and bring the inverting input feedback resistor strait off the output to speaker (The Emitters) this circuit works just fine. Why people run op amps with dual supply I don't know, I never run them that way. I use TL 081 high speed/high slew rate Op Amp with PN2222A/2N3906 for the push pull at the end. The simplest way is simply use TL 081 followed by LM 386N-4 strait to 8 ohm speaker, it sounds great!

[@audiotechlabs4650]

Love ya John! You keep us up on SS. I have not designed and built a +\- 5 watt cigar box guitar amp that runs on a 9 volt battery( 2-9 volt batteries, i.e. 18 volts dc)or a 12 volt dc wall wort. Transistor v IC? I will go back and take a look at your previous videos and see if you have something useful for my amp design. You Always have something useful! Love that Snickers! Thanxz

[@Bishka100]

You could use germanium O/P transistors for lower crossover distortion.

[@jasoneyes01]

John, I think you invented a You Tube music paralasys demonitizing protector. Until you added the bypass resistor!

[@xanataph]

Just needs to be made as plug-in that you can use in your favourite video editing program!

[@bjtaudio]

Its just as easy to use a LM1875, the idea of buffering the output of an op-amp to drive more current, it works but not that well, there was a headphone amp by silicon chip designed this way. Another crazy idea is to get like 100 op-amps and wire the outputs together for more current via a small resistors, however you will find the sound quality of the LM1875 better. When going discrete best to go all the way, and match the output device to get a very linear output.

[@eugenepohjola258]

Using a fast op. amp. may be one way to decrease the notches. However. It is likely that charging the base-emitter capacitances of the BJT.s is another aspect. The op. amp. output should be able to supply current fast at switch-over. Plastic capacitors of say 100 nF should be used in the supply lines of the op. amp. They should be as close as possible to the + and - pins of the op. amp. The cap.s will supply current fast to the op. amp.

[@TheTrueVoiceOfReason]

An update on the old Class B video! Couldn't help but think that to solve your last little problem of not enough current to push through the feed forward resistor, set up a small Class A non-inverting amp just after the op-amp to feed the output pair. Even a 2N3904 should have a more than high enough band width to work with the slew rate of the op-amp. Almost reminds me of the Losmandy amp. A bit more complicated, but much more stable. There's a link here: forum.nutsvolts.com/viewtopic.php?f=45&t=14664&p=71993&hilit=Losmandy#p71993 (yes, that's my thread over there) You could try one of these on for size some day. If you really want to blow people's minds, republish your LM317 amp video. Long live Vegmatic1966! :D

[@Dazzwidd]

A high slew rate op amp like a NE5532 and negative feedback can give good results

[@Elitemonkeys32]

I Love Your Channel ! Keep up the good work !!

[@researchandbuild1751]

I built this "basic version" but the top waveform cuts off waaaay before the bottom. Probably the NPN needs bootstrapping but I don't know how to fix that up without a bias circuit.. Doesn't help that my transistor betas don't match exactly either though.

[@russellhltn1396]

Seems like this amplifier would be quite efficient. Zero idle current in the output stage, just whatever the op amp needed.

[@laharl2k]

In that case just use a D type amplifier, also the mosfets are way mor efficient and the mass of th specker does the low pass filter part for free.

[@PeterMilanovski]

@@laharl2k no, not even if class D amplifiers can somehow power themselves, I wouldn't use them for audio reproduction. Just because you can use them for audio, doesn't mean that you should! You will understand what I mean when you have been using class D amplifiers for awhile and then by chance you get to experience an older amplifier that is either class A or class B... There's something that the class D amplifier seems to leave out of the music but I just can't put my finger on what it is! I first experienced it when I purchased a class B amplifier that was built in the early 70's, I was going to gut it and fit a high powered class D amplifier inside it. I made the mistake of plugging it in and having a listen, I still have that amplifier and I still can't explain why it sounds so good! It's only rated at 13 watts per channel but it's plenty of power to fill the house although I am running 15" three way speakers! The Sony ES amplifier that used to be connected to those speakers didn't sound as good even though it was rated at 100 watts! But it is a class D amplifier! I have noticed the same thing with my car amplifiers too! The older amplifiers run hotter but they definitely sound better which I can clearly hear...

[@johnyang799]

I think this circuit has a lot of potential. I can design a ultra low distortion(0.0000x%) 1A output current high speed high slew rate amplifier in the place of the opamp and use two or four high current high speed transistor to dump the current. It not only increase the current output, but also guarantee the output distortion to be low from low amplitude to high amplitude and they share the current so they reduce some of the thermal. Plus, due to the clean class B configuration, I can parallel output transistors as well as I don't need to add output resistors at emitter So I can have very very low physical output impedance. Also it's possible to use lm3886 composite amplifier in the front and use more transistors to dump current to lower distortion at high output.

[@Theineluctable_SOME_CANT]

Love it.

[@rohitb1869]

I was thinking about your cat and voila it was there, assisting you with your project.

[@kelvinjinxd]

For a proper current dumping, the inductor is required! Otherwise, it is just a distortion generator. Please check some papers regarding current dumping.

[@mouseminer2978]

Awesome RadioShack breadboard. Anyway great video. Thanks

[@STRIKEcorperation]

If the purpose of the 22ohm BE bypass resistor is to attenuate the signal propagation delay from the op-amp to the load, why don't you use a capacitor?

[@Trev9]

man. some people are just dangerous

[@BrianG61UK]

Why 22R for the dumping resistor? If you work out 0.6V over 22R you get 27mA which leaves nothing to turn the transistors on if the OpAmp can output output 20mA as you said.

[@tenkowal]

Yeah, I'd love more explanation on how the value was chosen.

[@Trev9]

@@tenkowal If a guy says 'sodder" instead of 'Solder', then you know he had not read and probably has no formal electronic theoretical training and therefore a poor grasp of Ohms Laws

[@michaels3288]

THAT IS COMPLETELY RIDICULOUS. I SAY "SODDER" AND HAVE PAID FOR MORE FORMAL TRAINING THAN I CARE TO ADMIT. IT HAS MORE TO DO WITH WHAT GEOGRAPHIC REGION YOU GREW UP IN THAN ANYTHING.

[@Trev9]

@@michaels3288 Why would you follow? Don't you read?

[@michaels3288]

I read just fine, but don't understand your incorrect use of the word I guess.

[@Darieee]

beautiful video !!

[@Markoul11]

use the dual 5532 its laser trimed and does not need any dc voltage offset. 5532 is a hi fi opamp, one of the best.

[@nocturnal0072]

I think the njm4556 would be good for this amp design. It handles more current and is just as cheap. An lm4562 would be better, but I would be worried about oscillation. I think the ne5532 would struggle with the load, but it is an amazing opamp.

[@anindyamitra5091]

Can you explain the functionality and significance of the zobel network in audio circuits, in one of the next videos ?

[@HillsWorkbench]

Maybe John will do a good clear video on this. I fumbled around with this topic last year, a video and a corrective video. kzbin.info/www/bejne/o6bLZo2nq7OngtE

[@HillsWorkbench]

And kzbin.info/www/bejne/Z3qzk6qVqZ5no80

[@anindyamitra5091]

@@HillsWorkbench, thank you!

[@urugulu1656]

this is actually very significant. i build an audio amp a while back around the lm386 and just thought i could give a fuck on reading the recommended application circuits and build the thing without the cap of the zobel. guess what that poor ten ohm resistor (quarter watt at the time) just got enormously hot. as did the ic. what did i do against it? well just use a five watt ten ohms resistor to keep that resistor from burning. until i realised that using that cap there makes all the heat problems gone at once. very easy and effective. so again its significant

[@vladan_kovacevic]

@@urugulu1656 Are you aware that you totally missed the point with ''Zobel'' network? You actually connected 10 ohm resistor in parallel with your speaker. If you were using 8 ohm speaker, your IC was effectively driving almost 4 ohm load. You were unnecessarily wasting almost half the power your IC was putting out, on that tiny resistor. You could totally eliminate 10 ohm resistor from your design if you wanted. "Zobel" network (more accurately - Boucherot cell) is used for stability reasons, but is not strictly necessary.

[@frankreiserm.s.8039]

I like your home-made heat sink. Frank Frank Reiser Video/Audio Service

[@laharl2k]

its better to just buy the cheap chinese cpu heatsinks, use or get rid of the crappy fan and you have a big heatsink for way cheaper than at the electronics store.

[@harviecz]

Are there some chances of both transistors being accidentaly open at the same time (half-bridge cross conduction)? That might be quite damaging...

[@harviecz]

Ok. It's made clear at 1:00

[@vincentrobinette1507]

Absolutely not, with the bases and emitters tied together. It guarantees that one transistor cannot turn on, until the other is completely off. Even at high frequencies, you have to clear the base charge of which ever transistor is on, before the op amp can swing over and turn on the opposite transistor. It sounds awful, but it's completely protected.

[@BruceNitroxpro]

Seems like a warmed over version of a modern (??) class AB amplifier. And the audio sounded it.

[@nocturnal0072]

New (real) class A amps are rare and/or expensive today. Most are class D with a fancy label. In the age of Bluetooth portable speakers and sound bars, it's efficiency over quality. :smh: Not that class D can't be awesome and sound great. Just not gonna happen without proper engineering and manufacturing.

[@johnyang799]

Also what would happen if you have the feedforward resistor but with biased output stage?

[@AstrosElectronicsLab]

Negligible improvement/non-improvement as the OS is now biased, it's not needed.

[@ranbymonkeys2384]

Have you ever blew up a capacitor with the cat sleeping there, haha

[@naderhumood]

Great apparatus. .....good job

[@martinda7446]

This was a brilliant video. Peter Walker of Acoustical Mfg was the man behind the QUAD amplifiers and speakers. He was amazing. But I hated the 405 it was a horrible sounding amp. Again great video.

[@joohop]

Hi Ya Bud Youve Got Me Thinking Is There A Way To Amplify A Crystal Set Radio Using Ambient Air / static Wave Energy Via High Impedence Semiconductos You Are A Master @ Your Work ! Well Done Fella ATB

[@andymouse]

Did I hear ' Van de Graaff Generator ' please ?.......Great video !...cheers.

[@SianaGearz]

So i was thinking of building this circuit into a headphone amplifier. Because ideally i'd like more current handling headroom than NJM4556. Not sure what for, but eh. I punched the circuit (including global feedback) into LTSpice and i am running a distortion analysis, and it's not doing very well. Even into 16-32 Ohm loads it's got like -65db second harmonic, give or take. I suppose with speakers or the like it must be worse. This is not quite Hi-Fi enough IMO. I tried adding small value resistors collector and emitter side, it makes the takeover smoother, but doesn't solve the issue, i get maybe 6db advantage at best. I am simulating against NE5532 or against NJM4556, of course it makes a difference what value current dumping resistor one can get away with, but ultimately not well enough. In contrast your headphone amplifier you showed, or the very similar Project 113 from ESP, they simulate with like -120db second harmonic, this is perfect. This is better than perfect. I wonder to what extent these are simulation only issues or whether they represent an actual issue. I mean there exists a possibility that discontinuities that are present in the simulator model, really are somewhat smoother in the hardware. Or maybe i should ignore, it's only 0.05% distortion or so? Then again, the proven problem free design is just a couple components heavier. I'm really spending hours in order to save myself a couple seconds of construction effort for a one-off amplifier? :D

[@nconrad4504]

Thank you!

[@diabolicalartificer]

Thanks, interesting video. could you go into more detail of the Zobel network/Boucherot cell please? Cheers, DA.

[@AstrosElectronicsLab]

It prevents, or at least attempts to prevent the amplifier going in to high frequency oscillation when connected to inductive loads over a reasonable cable distance (say 20')

[@johnjohns9572]

The op amps used in the Objective 2 may prove useful for your circuit. and... putting them in parallel just like the O2 -> 140mA output. (more than the 80mA you wanted). and also, you can bias the bjt's in that circuit as long as Ib remains small. check on the circuit in the link. sound.whsites.net/project113.htm . in short, the voltage drop across the diodes mimics the voltage needed to get "Vbe" on for the BJT's. the resistors set the current through the diodes which lowers or raises the voltage across the diodes. 2mA is a good point for headphones but you may need it more like 4mA or much higher to power speakers. Much larger caps could help as well at lower hertz. using BJT's with high current gain helps too. Removing R7L and R8L allows the BJT's to idle dumping more current - probably too much if you remove it completely. my guess the best combo is 4mA to 16mA through the biasing diodes with perhaps higher R7 and R8 until it idles reasonably (but this seriously limits the power going to the speakers). the cost of raising R7 and R8 i think is mostly a smaller max voltage swing. (note: the 16 mA isn't to turn on the BTJ's more.... it's to have much larger Ib's before running into trouble )

[@omarcusihuaman4261]

hello, i know that transistor audio amplifiers indeed amplifies current, if i want them to amplify voltage i will need an audio transformer? if so, how can i determinate the primary impedace?

[@xanataph]

I thought maybe it could be upscaled a little using something like an LM380 and a 2N3055/MJ2955 combo for the transistors, that's of course a -ve feedback loop can be used with such a chip?

[@TheTrueVoiceOfReason]

Since Paul Boucherot was French, the way to pronounce his name is Boo-sher-o (that's a long o sound at the end). Below is a website link showing how to pronounce Boucherot-cell in French. That's pretty much how I remember such words sounding during my short stint in French class, and how some friends north of the border pronounce similar words. www.howtopronounce.com/french/boucherot-cell/

[@siggi51991]

What is the campasitor .1 and 10 ohm call butcheros cell and what dose it do? and how?

[@stephanc7192]

Great video

[@researchandbuild1751]

I think you lose some additional power out because of the Vbe voltage drop. So you cant get quite as much power out as you could with some biasing i think, that seems to be what my experiments are finding. With 12volts and a 4 ohm load i can get max of 512mw or so. That seems low, i lose some voltage swing in the opamp but also lose some voltage on the Vbe junction too. I am using a single supply setup though which also may not be able to swing as large

[@DAVIDGREGORYKERR]

Can you not use two BD131 transistors on the output stage with the OP-Amp driving the bottom transistor google for the stage that I am talking about search for 5W Class A power amp is only thee transistors but is possible to lose the PNP transistor and replace it with the OP-Amp.

[@dhpbear2]

10:30 - The notch would probably be audible if you turned the volume down.

[@PeterPan-ob8lo]

Also class B amplifiers you can not listen very long time due to distortions (the notch) which you can not hear but you subconsciousness can.

[@Krmpfpks]

Super interesting.

[@eugenepohjola258]

Nice. The notches could possibly be made smaller using fast op. amp.s. LF356 has a Unity Gain bandwidth of 5 MHz. AD 847 is a RF grade op. amp. with a Unity Gain bandwidth of 50 MHz.

[@nocturnal0072]

Fast opamps would require a more stable (complex) circuit. At that point he could just bias the output transistors. It looked like he might have had some oscillation issues as is.

[@eugenepohjola258]

You are, of course, 101 % correct.

[@micomrkaic]

John, great video, as always. I'd like to mention three things. 1. Distortion at 10 kHz hardly matters in practice. The second harmonic is at 20 kHz, which almost nobody can hear anyway. So this amp is really good for practical purposes. It cannon thermally run away, which is a huge bonus. 2. Did you look into the frequency response? I've read somewhere that this type of amplifiers does not have a very flat frequency response. 3. NJM4556 would be very good for this application -- it should easily put out about 80 mA.

[@EJP286CRSKW]

Mico Mrkaic You read where? The feedback equation guarantees flat frequency response within the passband.

[@micomrkaic]

@@EJP286CRSKW Horowitz, Art of Electronics.

[@EJP286CRSKW]

Mico Mrkaic Can you provide the complete quotation? The 405 is certainly deliberately band-limited to 15-55KHz, and doesn't work well as a medium-wave transmitter, but then it wasn't designed to.

[@micomrkaic]

@@EJP286CRSKW The quotation is complete. Get the Horowitz book and read it, it is a classic. You'll find the reference to the frequency response (and not of the 405, but of the power booster of the kind John started the video with) in the chapter on BJTs or op-amps, don't know exactly. Also, the "feedback equation" does not guarantee anything if the open loop gain is insufficient.

[@EJP286CRSKW]

Mico Mrkaic There is no quotation here. The question is still what Horowitz actually says, and you still haven't answered it. And, for the third time, I can do without the smart-alec remarks. They add nothing. So far I've been polite about this when you've had second thoughts, but enough is enough.

[@uni-byte]

Use an op-amp with a faster rise time?

[@ceilingfanmusic6597]

You could fix the dead zone with a pre driver stage between the op amp an the out put stage

[@psionl0]

I was thinking of something similar. Replace the 22 ohm resistor with an emitter followr directly coupled between the bases and emitters of the output transistors?

[@Starbuckin]

All you have to do is connect the 22 K resistor feeding the inverting input of the Op Amp to where it's supposed to be, the OUTPUT, not the bases of the transistor push pull current booster, THEN the Op Amp will compensate for the crossover distortion when using a single supply to feed the circuit. That exact circuit was in my Malvino basic electronics book when I went to engineering school.

[@ceilingfanmusic6597]

@@Starbuckin the output of the amp?

[@EngineeringEssentials]

The video I was waiting for, Thank you first of all. The same thing I build with tda7294 instead of opamp and 2sc5200 pair for transistor, used for subwoofer. But it was disappointing, subwoofer didn't played punchy instead played boomy. What could be the reason? Low slew rate ?

[@nocturnal0072]

If you feel the slew is slow try an lm4562 or opa2132. The problem with using high slew rate opamps, in general, is the possible oscillations. They could kill your sub.

[@nocturnal0072]

Oh also, the input caps could completely change the sound. Try some C0J/NP0 or polymer caps if not already. You could have accidentally made a high-pass filter making your sub sound boomy and not getting that deep bass.

[@EngineeringEssentials]

What if I replaced transistor with mosfet

[@shamalperera4875]

Good sir...explain good..im srilanka...i very like..hi fi amp and speaker system...

[@zodak9999b]

This may be a dumb idea, but could you put an unbiased 2n3904/3906 pair as a driver between the op-amp and the output transistors to, but still pick up the feedback from the same place?

[@urugulu1656]

if i understood that right i guess that you just drive a small signal version of the power output stage from the opamp and use that to drive the final stage. i can not figure why this should make matters any better.

[@snaprollinpitts]

Hi John, it sounds very good, but, okay you knew there was a but coming. the op amp is great for simplicity sake, but wouldn't a small class A transistor amplifier be better is this case. there wouldn't be any crossover distortion. let me know . mike

[@JohnAudioTech]

Yes, a small fast class A amp that can deliver more current to the output could make this design work better.

[@johnyang799]

There is fundamental limitation with this design. (could be improved with more complex design) The reason why this circuit works is because the current output of first stage is limited and creates a voltage drop via the resistor. When the voltage drop is 0.5V or whatever to turn the transistors on the current will flow. If the first stage is ample and with low resistor value the transistor may never turn on. This should be improvable in more complex circuit. But the current simple circuit, a better first stage wouldn't help.

[@dantheman1998]

so basically the op-amp is driving the speakers during crossover?

[@JohnAudioTech]

Yes, but at such a low voltage, the current required is low.

[@ngneer999]

I'm building one of these amps for subwoofer only.

[@slap_my_hand]

Where is all the noise comming from in the demo at the end of the video?

[@anotherdave5107]

maybe you could drive your output into another op amp to act as an hi impedance to low impedance translation and get rid of your 10 ohm resistive output load

[@nocturnal0072]

It would be easier to just wire in 2 diodes, 2 resistors and 2 caps to bias the transistors imo. The absolute best an opamp could ever do is about 250mW momentarily. Most do around 30-50mW.

[@SianaGearz]

I think this calls for NJM4556.

[@miseryparty3726]

song?

[@vincentrobinette1507]

At the very end, did you have the 4 or 8 Ohm load on it? if not, the signal will pass freely through the dump resistor, keeping the transistors completely out of play. That would make the amplifier entirely distortion free. The input of your computer is 10K, perhaps even 20K Ohms.

[@shamalperera4875]

What is better sound ? Bjt and mosfet ?

[@SianaGearz]

Neither! The characteristics of the amplifier will be determined by the topology and component selection rather than component family.

[@ford1546]

Have you thought about the following? You have 2 amplifier board (stereo) or one stereo amplifier. You connect 1k resistor over R. GND and one over L. GND. What happens then? yes you get 2k in between R. and L. and you get more mono sound! ?

[@thisisyourcaptainspeaking2259]

Beats a TI Class D output stage all to heck! :)

[@userPrehistoricman]

Really? Did we find that out on this channel?

[@nocturnal0072]

Bold claim.

[@nocturnal0072]

The ta2020 would like a word with you.

[@thisisyourcaptainspeaking2259]

@@nocturnal0072 lol, definitely! Especially from the perspective of heat sink dimensions.

[@johnsenchakinternetsecurit8935]

We need more Snickers

[@urugulu1656]

at 23:14 didnt you forget to draw the input resistor connected to the ground. its flapping around in the breeze atm

[@AfdhalAtiffTan]

Is it possible to use independant opamp for the npn and pnp transistor?

[@nocturnal0072]

Something like that could be done for a headphone amplifier. Like the O2 kits. Opamps would get murdered past say 250mW or so on the high end.

[@NebukedNezzer]

motorola used this design on there hand held radios. both germanium and silicon transistors. the germanium sounds not as bad but they both are poor.

[@AstrosElectronicsLab]

For hand held radios (I'm assuming you mean broadcast radios and not two-way, either way makes no difference) as their output frequency response would be limited to around 600Hz - 5kHz, the distortion would be unnoticeable over the crap sounding speaker :P

[@urugulu1656]

is that ne5562 an actual audio amplifier? if not using one that is meant for audio purposes will certainly improve matters quite a bit.

[@urugulu1656]

well then. i just havent heard of it. i tried to build a similar thing around an lm386 and as of today it somewhat works but still has major quirks

[@nocturnal0072]

It's an ne5532 derivative if I recall.

[@rlwings]

Marantz advertises that it uses current feedback in it's amplifier sections as a very good thing! ... They also claim that there are no OP amps anywhere to be found in their designs. Instead they use their proprietary HDAMs. (Hyper Dynamic Amplification Modules) for a much smoother and superior sound. OP amps aren't really designed for audio.

[@blogtodeath4736]

That's strange as you will struggle to find professional/studio audio equipment that doesn't use op-amps!

[@JohnAudioTech]

"Op amps aren't really designed for audio" is the biggest bunch of nonsense I've ever heard.

[@nocturnal0072]

The ne5532 says otherwise. Also one could say the only reason the LM4562 exist is audio.

[@hexhere3851]

Hello John! Can you do a video about a stereo to mono circuit( summ) kinda like a subwoofer but with full range , thank you.

[@Auberge79]

Just 2 resistors from each input and third to the ground? No?

[@AstrosElectronicsLab]

@@Auberge79 pretty much.

[@AstrosElectronicsLab]

All you need is 2 splitter resistors (one for Left and Right) tied together at one end, fed in to a normal opamp input stage, boom, done. The value of the resistors is up to you but usually around 10 - 100k is sufficient.

[@LegoBeto]

If you ever need some music to test your circuits on your videos, you are welcomed to use my bands music any time. PM me for more info.

[@blueconcretezebra]

Current dumpling as per Quad 405.

[@JohnAudioTech]

Apple dumplings taste better.

[@EJP286CRSKW]

Not really. He doesn't have the feed forward/feedback bridge that makes the 405 work.

[@Theineluctable_SOME_CANT]

Nice video...

[@johnyang799]

why not try opa2156

[@martinda7446]

You can 'stack' a few 5532s in parallel to improve current and other parameters. I've never done it but I don't see immediately why you cant?

[@BrianG61UK]

You can't just parallel up OpAmps. It would only take a tiny difference in their offset voltages and large currents would flow out some of the OpAmps straight back in to the other OpAmps.

[@martinda7446]

@@BrianG61UK No, that is not the case.

[@martinda7446]

@@BrianG61UK Think about it...

[@beanMosheen]

Do you ever worry about solder residue and the cat bathing itself?

[@nsfeliz7825]

yeah noticed that.

[@AK2I47]

How is a single op amp so loud?

[@AstrosElectronicsLab]

Let's look at how amplifiers work. They are what are known as current to current converters. In order to drive a loudspeaker, you need reasonably high current to get decent levels. The opamp alone cannot (well at least the NE5534 cannot) produce enough current (John said its output current was in the order of 80mA). For listening levels you need at least 200mA or more, so that's what the two transistors do. They allow a much higher current swing to produce the "so loud". In terms of wattage output, that's the amount of voltage swing and current swing the output can produce in to a given load (not just one or the other).

[@friedmule5403]

You have almost everything ready, a CAT-ode but where are the lade called An? :-)