need power supply advice for future power amp project.

Hello all.

In a couple of weeks im a be starting an amplifier build, which will consist of 5 LME49810 and one LM4702 Chips, ill be looking to get around 80W~100W RMS @ 8ohms and capable of pushing a 4ohms load. per chip.
the first problem i have out of the many to come while working on this project is selecting the power supply transformer, I was planing on getting this Avel toroidal transformer from parts express which seems like a reasonable price @ $85usd, specs are: Load Capacity: 800 VA Secondary Voltage: 60V + 60V , Primaries: 115V + 115V. frequency: 50/60 Hz, or one of these from eBay Link here
im just not sure if i need 60v per rail rather then 35v per rail, also wondering if 800va would be enough.since its going to be running all 6 chips and around 700wrms of total power at full blast.
also would 80v caps at 2200uf be enough for the filtering stage of the power supply at 2 per rail?
i have done lots of research about all this, but im very new at building and designing amps and power supply, so i don't trust my final decisions as of yet,
it would be greatly appreciated if you guys would lend me a hand at it. thanks in advance.
a BJT output stage would need about +-50Vdc supplies to generate ~100W into 8r0. This is generally provided by a transformer that is 35-0-35Vac or 0-35 & 0-35Vac
If you want 4ohm capability from all the amps then that equates to 200W into 4r0.
If all six amps draw power simultaneously then you need a VA rating ~1.5*(total output power) ~=1800VA.
Are you prepared to relax the 4ohm capability requirement?

With UK tolerances on mains voltage supply and the typical regulation of power amp transformers you will find that 63Vdc capacitors suit upto 40Vac.
allright so for [email protected] and [email protected], i can do with 35v per rail, no need for 60V per rail thats only if i need more power right. and what i have to worry is on the va capability of the transformer. i guess the general rule is 200Va per 100watts? so ill need a transformer capable of pushing 1.5kva. would it be better to have two smaller transformer like the avel brand from part express rather then one large 1.5kva transformer?

thanks for the help so far andrewT.

Ps: i was planing on using mosfets on the output stage, any thoughts on that and power supply requirements.
It depends on the chosen topology but the more usual source follower will need an extra 4 or 5V to get the same power as a BJT output stage.
So 100W FET needs about +-54Vdc from a 38Vac transformer.

The usual range for the transformer is VA rating of one to two times the maximum power & most find that 1.5*power works quite well.
Transformer power ratings need some thought. The power they dissipate is highly dependent on source material. Pop music typically has a peak to average ratio of ten to one (classical is closer to 20 to 1). This means that if you are not clipping, a 100 watt amp averages 10 watts at "full" volume. By using larger capacitors in the regulators you can control how long a high power burst can be sustained. Since transformers heat very slowly and heat is the primary limit on transformer size the 700 VA transformer you propose is probably adequate.

Two exceptions; class A designs dissipate pretty much full rated output power at all volume levels and music amplifiers (such as guitar amps) should be designed for continuous full load.

There is nothing wrong with oversizing the transformer and a design that supports full power continuously on all channels would be called conservative, it's just expensive.

Commercial receiver power conditioning tests are typically done at 33% of rated output power since that is the power level that dissipates the most heat in a class AB output stage, if not the most in the transformer.
well i figured id buy the 50vac 0 50vac @ 800va toroid transformer to power 4 LME49810 chips @ 100w @ 8 ohms each, with 4 ohms capability and use one of the 35v 0 35v transformers i already have to power the other 2 chips.

can i use 70v capacitor on the 50vac 0 50vac transformer?

i have a couple 70v 4700uf and 80v 4700uf caps that i would like to use on this project. i figure i could parallel the caps till i get the proper total capacitance per rail. then do the same with 50v and 63v volts capacitors that i have at hand for the 35vac 0 35vac transformer.

let me know if that will work.

thanks for the help so far guys.
Aluminum electrolytic capacitors should be run at around 75% of rated voltage, so a 100V will be about right.

Electrolytic capacitors have a finite life (if you plan on 10 years maybe as long as 20 you won't be far wrong) Too low or too high a voltage will shorten their life.

Film capacitors can usually be run as low as you like, but never over their voltage rating, films generally have no wear out mechanism.
thanks for the reply.

these are the measurements i got from the transformer i currently have.
its a dual 45v, so ~44.2Vac 0 ~44.2Vac , thats what it measured with the DMM with out a rectifier.
with the rectifier installed the dc voltage measured at: +39.7Vdc 0 -39.7Vdc , it lost 4.7Vdc because of the diodes.
with the filter capacitors installed it measured: +58.2Vdc 0 -58.2Vdc.

I also measured the voltage with a 4ohm resistor load wire in series to five 12v light bulbs in parallel to each other, just to see how much the voltage dropped, it dropped down to 44vdc and with the amperage meter hocked up in series with the load it read 4.5amp draw and the transformer didn't sound like it was hurting so it may have some more juice left in it.

would you say its safe to use a 60v filtering capacitor or should i go with the 70v?

To pick a capacitor for your transformer, you need to know the power line voltage. Good practice allows for a variation from nominal of +/-10% although +/-5% is more typical.

So to answer your question you need to know if the power line was high, low or average at the time you made the measurements.

It used to be that designers used 115VAC as the most likely nominal, over the years that has increased to 117VAC and now days 120VAC isn't uncommon. The highest I've personally measured was 126VAC.

It is safer to err with capacitors rated higher than the voltage you expect, getting it wrong that way has a tiny impact on life. Using an electrolytic rated too low will result in gases escaping the aluminum can, sometimes explosively. Even if they don't explode, they leave a messy caustic goo all over your circuit

I'd go with 70 or 75V
thanks to you hermanv and AndrewT for your explanations. now i pretty much understand how to select the proper capacitance voltage rating and more.

some of the instructional files that i found on the internet about power supply design, although very informative, they lacked this necessary pieces of information.
thanks you guy for showing me the missing parts that i needed to know to finish this section of the project.
even though it was a simple thing, using a too low of a voltage or using a capacitor rated way too high, was only going to cost me more and more money. so thanks for saving me some cash. lol

tomorrow ill measure the input voltage to see if it was at low 110v or at high 120v and ill adjust from there. im sure ill end up using the 70v capacitor, but ill check just to make sure.

cant wait to get the power supply built, so i can start designing and building the LME49810 chip amp.

thanks again for all the help.
talk to you guys laters