Discrete audio amplifier project PT15 square wave response test

[@isoguy.]

Really enjoying this series, thank you for sharing

[@amitanaudiophile]

You have done great design class AB hi fi amplifier. Square wave response too shows it will sound very transparent. Nice upload sir.

[@demoncloud6147]

Watched all the episodes in 2 days, glad I didn't have to wait 1 year !

[@JohnAudioTech]

Thanks for watching. More to come...

[@1pcfred]

Everything is better with some big Nichicon capacitors.

[@rolfts5762]

As always, enjoy watching&listening the JAT video's ..The Cat was really thirsty : - ) //Thanks for sharing John

[@theoldbigmoose]

Great series, and a good design with great explanations. Much appreciated

[@elgringoec]

It's great to see people still getting into this! Kudos for studying and being methodical. You're about where I was forty years ago. These tiny amps are great learning tools but as you probably know or will discover at some point you need a couple kW per side for concert-approaching realism. Keep on rockin'!

[@amitanaudiophile]

Yes but there is another approach also to go with high sensitivity full range speakers and transparent sounding low power tube amp.

[@elgringoec]

@@amitanaudiophile Been there, done that. Unfortunately though, neither of those two things exist.

[@amitanaudiophile]

@@elgringoec oh

[@elgringoec]

@@amitanaudiophile Everything's relative, right? Like what's "full range"? And "transparent"? They're subjective and based on comparisons. There's no objective criteria defining those descriptions. They can be whatever one decides they are. And our own criteria refines with experience and exposure. I encourage people to enjoy where they are along their journey of discovering. Ultimately it's a psychological experience that we make for ourselves.

[@amitanaudiophile]

@@elgringoec ya I agreed

[@anindyamitra5091]

I am loving it! Ok, so you have cleared up all of my confusions. Thank you very much!! cheers!!

[@steveswan5714]

Hi John you make what was once very boring to me 😴is now very interesting🤔 😊👍

[@Dazzwidd]

Hey, I remember saying that the oscillation was caused by supply rail decoupling (bypassing). Seems my gut instinct from experience in practical electronics experimentation is right 😁

[@SuburbanDon]

Excellent. Closer to my dream. By the way i dont think you should have done anything different than you did.

[@rickloresto9647]

you are my idol in designing ...building...audio amplifier....

[@stephanc7192]

Well done! I enjoyed!

[@pliedtka]

Even when I do some 'spicing' of amps I include this kind test. Also to check damping factor, I simulate load very much resembling actual speaker impedance, and then add capacitive load in parallel - interesting how some amps behave. If it's a piece of ... the simulation will show it. I think Prof Leach amp was pretty well behaving, despite the fact that it is quite number of years when it was published. Still, actual test won't replace simulations 👍

[@martinda7446]

The clipping variation was droop on supply cables. If you put those big reservoir caps nearer the board you may even get a little more. Well I think that's what it must have been considering your regulated supply.

[@humpadinger]

Hi John, could you let me know how you applied the capacitive loads during testing? Did you apply the caps directly to the board terminations or across the dummy load? If you connected across the dummy load there will be some series resistance that tends to reduce the ringing seen right? How should we do this to formulate the test in the correct way, i.e. provide scenarios that the amp sees in real life use as opposed to a really arduous test that will never be replicated in normal use. Thanks, keep up the great work!

[@bashcoder]

Great stuff - when your circuit is done how will you measure THD?

[@diedraak1]

Good video. Quick question. Would you say its normal for most bench power supplies to be affected by transients when powering amplifier circuits. My bench supply is rated at 4A and has no problem with weird loads. BUT when I power my amplifier projects without adding extra filtering, it cant keep up and goes out of regulation.

[@garyhunkin]

Do a dynamic power pulse test with the big supply caps. It makes a difference with bigger supply caps. Doing a class H output stage with higher additional supply voltage and big supply caps really delivers dynamic power. Makes a big difference with dynamic music. Even at reasonable volume levels.

[@n.shiina8798]

class H only helps heat efficiency. the power still limited by both its HV and supply current capability

[@badscrew4023]

1 microfarad or even more can easily be found across the output if a tweeter or a mid driver shorts out for example!

[@johnyang799]

Would you use two pole compensation in the output network? Higher stability, higher feedback.

[@johnyang799]

Also output inclusive/T compensation.

[@JohnAudioTech]

I'm using miller compensation. It is simple, yet at high frequencies feeds back some signal to the transistor base which provides correction at high frequencies. Not to be confused with the miller capacitance internal to the transistor which is voltage dependent and will increase distortion. More advanced compensation techniques are better but I want to keep this amp fairly simple. If I find the high frequency THD figures are not meeting my goals, I may look to something more involved.

[@n.shiina8798]

definitely wouldn't be as simple as the current design but it's worth to try if you want to achieve better performance

[@johnyang799]

@@n.shiina8798 Well, it's as simple as a modification can get. 1 extra capacitor and 1 extra resistor. It's true that it's not necessary. But basically by doing that, you can either enhance the performance or enhance the stability, or kind of both. With the aid of the output zoble network. It can achieve terrific stability with lots of feedback. And well, it's best and easier if it's doing in simulation. Guess we'll wait and see how current one performs and let John decide eventually.

[@n.shiina8798]

@@johnyang799 yup. i've tried TMC, TPC and OITPC last year. not easy for me but they could give more either performance or stability for sure. i was helped by simulation model. otherwise, i might blow a lot more devices lol

[@MrDunk66]

John On the LM1875 board your signal input capacitor is about half the size of the one I have used because I could not find anything smaller. (Panasonic 10μF Polyester Capacitor PET 100V dc ±5% ECQE(F)) What did you use? Thanks.

[@JohnAudioTech]

I used a 1uf cap (!) as the pole frequency is plenty low enough for audio (~8Hz). A 2.2uf would move the pole down under 4Hz so that the frequency response down to 20Hz would be very flat. Going higher doesn't make any sense to me.

[@MrDunk66]

Thanks John. I found some 1uf polypropylenes. Much smaller. I presume this is the relevant calculation? 1/(2 × π × 22,000 ohm × 8hz)=900nF

[@JohnAudioTech]

Yes, that is correct. Sorry I missed this, it was in the spam folder for some reason.

[@DAVIDGREGORYKERR]

Can you mount those monster transistor direct on the heatsink and use screened cable to connect the transistors to the circuit board.

[@montynorth3009]

I was wondering if shorting out the 4.7 ohm resistors would affect the square wave performance.

[@n.shiina8798]

it would if the value was too high but 4.7Ohm seems reasonable and wouldn't affect too much

[@johnneybot9542]

SNICKERS NEEDS WATER !

[@BenState]

liquid cooled audio amp - GO!

[@lordzamasu6220]

Hey how much will it cost?

[@johnyang799]

I think 50 dollars with pcb board and components. No power supply, no heat sink etc.

[@JohnAudioTech]

I hope to keep it under $50 for the complete kit (board and quality sourced components).

[@lordzamasu6220]

Oh nice, will you send it to Lithuania?

[@johnyang799]

@@lordzamasu6220 I think when it's out. You can get BOM and pcb gerber. You may then buy parts from mouser, digikey, and order pcb from jlcpcb, pcbway. It won't cost too much. Both mouser and digikey offer free shipping when the parts are over 50 dollars. And pcb + shipping cost around 20 dollars with dhl.