Discrete audio amplifier project PT6 power supply design 

It's main purpose is to shunt away the switching noise before it gets back onto your mains, making other electronics in your house very unhappy.

Good quality switchmode supplies can certainly be used but aren't that cheap. With their fairly good regulation, they make the dynamic power close to the continuous output power. Dynamic headroom is always welcome.

For the size yes. For the price? Idk man.

I'll check that out. I've seen Avel and Hammond transformers so far.

this is the website. www DOT antekinc DOT com

I have a question . Can I use 18V 250VA toroidal transformer with Vishay 400V 35A diode bridge ? I want to power 5 x LM1875 & 1 X LM3886. I am building a 5.1 hi-fi system. Will 250VA be enough?

Yes, chokes are the way to go. Very common on low current HV tube amp supplies but almost never seen on LV high current solid state amplifier supplies.

If cost is no object, you can use a larger transformer. Most speakers are a pretty easy load to drive since there is a series DC resistance that helps to swamp the effects of reactance. Of course, there are some speakers that can be hard to drive.

JohnAudioTech we should develop a new power standard that makes sense for music and the actual speaker load. Like ‘dynamic power’ that would automatically take into account a load range - like maybe 8 to 2 ohm, so it goes unsaid and to be expected. Much like we expect the frequency bandwidth to be 20-20 kHz. Then we would need a test pattern that would verify the dynamic power is clean using maybe a THD analyzer.

Just thinking that we could come up with a better way to quantify power ratings. Instead of audio being so subjective...

@@KissAnalog You know salesmen always want to make some thing huge out of nothing. Which is the reason we got into that mess. And salesmen never mentions the bad sides of the story. I remember a time the watts always was in 8 ohm. When that was gone it went wild the watts sometimes was meant in 2 ohm, although it could not managed that. Salesmen......

Tbonyandsteak You are so right - marketing loves numbers and keeping the focus on what they want.

Just make it yourself. Making a transformer is pretty interesting experience. You only need to find a core. Maybe from some broken variac, or an old TV. Of course you can buy a new core, but this may be even harder than find a ready-made one. Come to us in Russia, I know a few manufacturers who sell transformers and cores ;)

Believe me, I have already studied this area, make self. The toroid winders are far too expensive to buy. They can be home made though, but only someone like me, who has a machine shop could make one. Making an older style laminated core type has it's problems. Mainly producing the laminations. And even if you can, I could if I wanted, getting the right quality consistency of the raw material is an issue. Then another issue, joining the laminations so they do not vibrate, or come apart and vibrate. That definitely takes experienced personal, and good methods. A problem though, making one self, is that unless you make it as well as a major manufacturer, and that is most unlikely, you will end up with the transformer producing self noise at levels that are unacceptable. This noise may not matter if one is making a door bell, which was the reason I wanted a transformer a year ago, but in electronic applications, noise can make many circuits non functional. I have considered manufacturing transformers if they actually are difficult to come by.

Agree

Last I looked, there was a good range of toroidal types available new. Other than that, shop the surplus market though pickings are slim.

@@BiddieTube if you need only one or two transformes you don't need an expensive winding machine, you can do it with your hands. This requires some practice though. Concerning the quality, I must say that all transformers that you can find in modern electronic equipment are pretty crappy. The main goal of any manufacturer is to make the product as cheap as possible. They usually pick a smaller core and thinner wire, wind less turns, don't use a shielding winding (you rarely need it though), etc. You definitely can find a good transformer, but it will be SO expensive! I just want to say that it is not a rocket science. Do some simple math to get the core size and the number of turns. Take a core made of steel with high flux density. Wind the wire neatly (it may take a few days). Voilà! You have a transformer which is better than 90% of the commercially available rubbish.

Hi Mike. Thanks for the offer! Let me poke around the interwebs to see what is available and perhaps fine tune what I should get.

cool

Thanks for making the list! YT held it for review. I also have a playlist for this project.

That is a class D amplifier, and the spec does say 420 Watts into 8 ohms with a bridge output or two channels at 210 Watts into 4 ohms. The voltage at the rails need to be +/- 41 VDC.

I have no idea not knowing the circuit. I would error on the side of caution. You may find the supply voltage will drop with just idle current draw. You can put a few silicon rectifier diodes in series with the supply since they have a voltage drop of about .6 volts and check what the idle voltage is. If it is too low, just remove one until you get around 17v.

@@JohnAudioTech thanks man.

eeeh careful now you connect + and - to make the dual 24 - 0 - 24, dual is a bit tricky. it has to be the correct + and - Blue and red Find the schematics. You better check the voltage out put when you do what you do. It is like you make a single 48 volt output but uses the connection of the 2 lines as ground. If you only get 24 volt you have done it wrong, you got to get 48 volt That will be your 24 - 0 - 24 if you connect the negative as ground the 2 lines will be in the same phase, they have to be the opposite phase of each other. The problem is if they are in a same phase you have doubled the current. The circuits may not work as intended.

The transformer is way overrated for a stereo 50w amp, but can be used. The load would be light for it and load regulation good. I think you meant 26a at 24 v as you would parallel the two secondaries. Transformers don't have "negatives" but you can join the two secondaries together so they are series for 24-0-24.

Here is all of the information that they offer, thanks John 625VA Toroidal Transformer 117/234V to 2 x 24V or 48V at 13 Amps www.surplussales.com/item/_TP/aa56252-024.html www.surplussales.com/Transformers/Pdfs/Actuant_Amveco_1-14.pdf www.surplussales.com/Transformers/Pdfs/Actuant_Amveco_18.pdf • Primary: 117/234 volts 50/60 Hz • Secondary: 24/48 volts • (two 24V @ 13A or one 24V @ 26A or one 48V @ 13A) • Apparent Power: 625 VA • Weight: 9.5 lbs • Dimensions: 5.5"DIA x 3.2"H • Factory New Surplus • Manufactured by Amveco Magnetics • Alt. P/N: AA56252

Tbonyandsteak+ Thanks, I told them I'm new at this and all I need to know is what 2 wires to tie together to get a 24-0-24 and they could have just said to tie a neg to the other pos or tie the 2 negatives, but the company wouldn't explain it and told me to look at the PDF's. I left all of the information on this transformer in the comments from the company. Could someone please look at this, I would greatly appreciate this, thanks I'm happy I asked, I was under the influence that AC didn't matter about pos and neg and thought that they only wrap one wire bent over then cut it when they're done @ 24v then have it again to get another 24v and they color code the wires so you don't have to test the continuity and make sure you don't have the same wire

That IS a beefy toroid. You're right about there being no + or - and about the center tap winding being one wire, double wound and cut to make the 2 secondaries. The issue (as Tbony mentioned) is getting the phasing right. This is determined by the direction of the windings. Same direction=same phase=double voltage (48-0-48@13A). Opposite direction=180phased=double current (or 24-0-24@26A) A quick power up trial with your multi-meter should set you straight. Twist 2 different coloured wires together and probe across the untwisted wires. If it's 48VAC, switch ONE of the twisted wires for the same colour untwisted one and you should see 24VAC.

Biggest difference is the quality of the core and winding, so you can have good EI. With toroids you can have lower stray inductance. Generally toroids are build to better quality. My Pilitron audio toroids are quieter, run cooler and are more efficient than similar EI used in Far East made equipment. Tim de Paravicini of EAR, an electrical engineer by education uses 1000W core for 700-800W transformer in his designs. All my Japanese 'made' equipment have transformers which are average in quality - they run very hot comparing to more Hi-end pieces and usually are undersized.

Thanks for the support! I will consider Antek when I get serious about purchasing.

I've only seen heavy filtering and common mode chokes on switching supplies that must keep the noise out of the mains. The typical receiver or power amp that uses a standard line frequency transformer often has simple filtering such as an across the line cap.

You're right. I made an LT SPICE model using your schematic and sized the transformer to put out 420VA with +/-25V voltage rails. There is a small amount of distortion of the input current from the LISN, but it starts falling off at 40 dB per octave at 1 kHz and isn't even close to what you would see from a solar inverter. I guess the just the leakage inductance from the transformer (I assumed K would be between 0.97 and 0.99) in conjunction with the capacitance across the line is more than sufficient to resolve any issues above 150 kHz where conducted emissions are tested...

ats89117 the bridge rectifier is the source of the noise - the Harmonic content is based off the 60 Hz, and well within our audible range - but yes well below the 150kHz corner of the fcc EMI specs, but are we not concerned about what might be offensive to our ears? What did you measure as Harmonic content above 60 Hz?

@@KissAnalog I had a PC with cheap power supply that would sent back into AC lines 120Hz - if I connect the frequency counter it read 120Hz. Didn't have FFT analyzer back then. So, yes there's a lot of junk flowing around in the mains. As for the switching noise from diodes in regular bridge rectifier - that is interesting to explore. Is this the reason for Schottky diodes in power supply.

pliedtka Schottky diodes have a lower voltage drop so are more efficient. Capacitors that are too big can cause high peak currents and if they are too small they can cause high ripple voltage.

Not overall, there will probably be some regulation onboard for the front of the amp.

Line voltage stepped down, rectified and filtered is regulated enough. You can consider choosing the right voltage output transformer the regulation step.

Regulation means you would have to pick a transformer with a higher voltage at the same current for the regulator dropout headroom which means the transformer would be a higher VA rating. That means a larger and more expensive transformer. You also have to design a regulator that can handle the load from the amplifier(s) and dissipate heat. The continuous power would be the same as the dynamic power. I'd rather it not be regulated and have the extra dynamic power since music is not high a level continuous signal (unless you like the bass type stuff).

i see. so, if we are into bass, then it will be helpful to have regulation a bit.

Where would a regulator get additional current from? If I wanted more current I'd use a bigger transformer and more filter caps.