Discrete audio amplifier project PT8.1 EF output stage test

[@andymouse]

Very well described, I'm slowly familiarizing with the stages and terminology and function, long way to go but its fun you showing why things are the way the are...cheers.

[@rrangana11]

Thanks John... really enjoying this project, progressing very well.

[@iblesbosuok]

Keep educating viewers, sir. Cheers from Indonesia

[@michaelzoppo534]

Great demo of crossover distortion!

[@davekazoroski6548]

Most important thing for reliable operation is DC stability, so whichever hookup is the most DC stable is the way to go. The miller cap on the VAS will most likely eliminate all oscillation. As far as distortion goes, I defy anyone to say they can hear less than 1% in speech or music (with the exception of some professional musicians)

[@davekazoroski6548]

Rod Elliot has a good article on Darlington vs Compound. He also has a very simple cure for the negative half cycle oscillation.

[@terencekaye9948]

Getting super Exciting John! :)

[@marijntopgear]

John! You rock!

[@christophschuermann6512]

The stability is a keyfactor for any amp design! If you dont load your amp to a minimum current at the output, you run into several problems due to the slow "off switchingtime" of the BJT's .

[@HillsWorkbench]

Yes, the larger caps didn't cause oscillation (w/CFP) because they were starting to become a significant load.

[@JohnAudioTech]

That was the EF stage that was more stable with the larger caps. I'd guess the larger caps have a different phase relationship with the output at those oscillation frequencies. The boucherot cell would present a resistive load to the output at those very high frequencies where a direct cap is reactive. I would think the non inductive resistors I use as a load would be resistive too. Of course there is parasitic reactance all over the circuit.

[@christophschuermann6512]

@@JohnAudioTech you have to be allways careful with parasitics of the breadboard and the wrong impression due to interaction of different sources of errors. I have doing ampdesign for over 30 years and won over 70 awards worldwide with my amps etc. , and yes the world is very analog and komplex if you did deep enough....

[@JohnAudioTech]

Hi Christoph. I hope you stick around for advice. If you are available, I can be contacted at my [e-mail] account.

[@1pcfred]

I think we knew all along what you were going to do.

[@tuanluong3627]

Thanks for sharing, John. Could we use an Opamp for input stage?

[@MrAlFuture]

Verrrrry interesting!

[@johnyang799]

What about diamond buffer? Also could you explain about error feedforward correction? Thanks.

[@akkudakkupl]

Needs incredibely well matched transistors, viable only in ICs really...

[@pliedtka]

@@akkudakkupl Thanks for the advice, means to make something like Cordell does I need a lot of pieces to match them = lot of $

[@AboubakrA]

thank you for sharing the filter of the Sziklai stage .. however I had a question, how can the emitter follower become unstable with a 2 ohm//10 nF load !!? there no local nor global loops, I think it might be the preamp that is not isolated enough from the output and somehow becomes unstable ?

[@anindyamitra5091]

can I use half darlington pair and half sziklai pair configuration as the output stage? I mean I want to use npn-npn(2n3773, where its complement pnp isn't available) output stage, with npn-pnp driving transistors. Does that have any issue?

[@Markoul11]

John, Don't do it too much with the global feedback. Although it is good for linearity, BW and stabilization it is catastrophic for the transient distortion of the amplifier. This is the ablility of the amplifier to respond quickly to the transients of music thus very abrupt and large variation in amplitude of audio frequencies and differs from the dynamic response of the amplifier which is really its S/N ratio. A very low transient distortion amplifier reproduces sound very natural and is an important part of hi fi audio. This is achieved with very fast and very small rise and fall time transistors and of course supported by a strong and fast power supply. Transient response is many times neglected by hi fi enthusiasts. There many designs which are using HF transistors instead of AF transistors for that reason. MOSFETS designs are considered slower than bipolar transistors. Unfortunately large global negative feedback is a killer for transient response of an audio amplifier. In my designs I don't use more than 3% global feedback and prefer to use large local stage feedback emitter resistors a.c coupled. Also by reducing the amount of global feedback used you don't need that large open loop gains on your amp stages which is known to cause many linearity problems and and loss of stability. I prefer the two stage approach whenever I want a large gain instead of getting it from a single stage, which will reduce the BW. Also remember a 6X2=12 gain dual stage, has less BW and is less linear than a 4X3=12 dual stage. Finally, I use in all of my voltage stages IC regulated voltage supply for increased power noise rejection and stability except the output current amp stages where I use unregulated power supply.

[@pliedtka]

Hi, I remember Non global feedback amp from LC Audio in Denmark. To keep the gain low the emitter resistors on differential input stage was in 1-2k value. My friend designs hi-end spkrs drivers - he does plannars as well as regular midwoofer, woofers some of them are the fastest, cleanest I have heard but to make them takes a lot of $ for small production. He favours DSP crossovers since the flexibility in creating additional EQ points and adding delay to much acoustic centers or desireable freq resp is just heaven send comparing to passive crossovers. So yes I would agree with you that impulse response is very important and makes for much more transparent, high resolution amp. What in your opinion in terms of V/usec rise time or slew rate make a good amp. Also what is your opinion on CFT amps.

[@iceberg789]

isn't 2SD882 better choice than those BD139 ? because more higher & flatter gain curve over wide collector current range ? other than that, this configuration sounded more punchy than cfp..... donno, may be because of the music itself.

[@JohnAudioTech]

It is not usable in my circuit due to the low Vce of 30 volts. The drivers may see 60+ volts across them during full output. The On semi datasheet (BD139) shows a reasonably flat gain in the region of current that I will use them.

[@iceberg789]

@@JohnAudioTech i see. thanks.

[@GnuReligion]

You may benefit from soldering up screw terminals and round-hole connectors to a PCB in the configuration for Quasi Pair Complementary, Darlington Pair Complementary, Compound Pair Complementary, etc. This way you can test high currents on various output stages while retaining the ability to switch out caps, resistors and power transistors. You seem to spend a lot of time bread-boarding output stages for tests. 0.1" and 0.2" screw terminals exist: www.pololu.com/category/177/0.1-2.54-mm-screw-terminal-blocks Looking forward to building up your final amp!

[@DAVIDGREGORYKERR]

What about an all NPN transistor amp with 3x BC182L transistors driving 2x BD131 power transistors to produce a wide band Class A amp with an output of 5-7watts like that used as the color gun driver amp with a voltage of 13.8V DC.

[@harri9300]

At 14:25 I was waiting you to start rapping about amplifier design.

[@johnsenchakinternetsecurit8935]

BOTTOM SIDE FURRY = Snickers

[@snaprollinpitts]

hey John, if you need a board made up, let me know. I have boards at the ready, and chemicals ripe for the dipping. I will need the full schematic, with or without the preamp. so leme know. mike call or email, or message me and I'll call you.

[@JohnAudioTech]

Hi Mike. I'll see how it goes. I'm now putting it on a solder type bread board so I can continue with the design. I'm getting closer to the first prototype. I still need to get back to you with a good transformer selection. Haven't forgot!

[@snaprollinpitts]

thanks John, mike they call that board perf board. but you probably already knew that.

[@CoolDudeClem]

They're both pretty good but I preferred the first output stage from the previous video, this one just doesn't sound as "full" to me. I wouldn't even been able to tell when the sound was distorted on the "youtube safe music" since. all "modern" music sounds like distorted crap to me anyway.

[@rrangana11]

I could hear the crossover distortion when the bias was reduced...

[@pliedtka]

The modern distortions or Remastered disaster - if you clip the DAC it creates distortions similar to badly designed AB class amp with underbiased output stage. Since a lot people listen to portable devices with limited voltage swing, a lot of music is compressed with reduced dynamics so there's enough volume from crappy headphones or small speakers. Before it was done to make less noisy recordings with tape recorder, but the way tape recorder compress and distortions it creates is much different and not so offensive to the ear, think of it as progressive saturation of magnetic core in tube amp output transformer. The way it is done today with plugins in DAW is very aggressive, often the final result is what you describe - Junk-but-loud, JBL. Funny because there are some so called audiophile recordings that are very compressed and some, say from MA Recordings recorded with very small level, yet the noise is not an issue. On average these days the level is 15dB higher than 25-30 years ago when most of CDs was mastered not to clip DAC and usually a bit lower than 0dB. Welcome to the new deaf world. BTW, I went to Cinema a month ago and I thought my ear drums are going to explode, so I spent the whole movie with earbuds in my ears - that f.. loud, at least 10dB too loud.

[@joohop]

Play This I Made It

[@joohop]

kzbin.info/www/bejne/iounXpuoq8p7qsk&lc=z22ozb4oqordgbm5j04t1aokgsqi2osbmfkxoag0gm13bk0h00410