Help with first Power Supply for LM1875

[danielwritesbac]

Originally Posted by pacificblue

In a centre-tapped transformer current will only flow in one half of the winding during one half-cycle. If you use a dual secondary tranformer with full-wave rectification, i. e. one rectifier bridge per [each] secondary, current will flow in each secondary during both half-cycles. That leads to a higher efficiency for the dual secondary version.

And lower voltage. With dual secondaries you can decrease the transformer amperage rating by the same proportion of the voltage drop. That's not relevant to efficiency on an amplifier of this scale.

Steady tone figures are either irrelevant or the steady tone is an accident, in which case the device is morbidly inefficient.

That doesn't relate to a design goal.

LM1875's have a too-small tolerance for current, thus limiting the importance of current by that amount. With its a too-small resource and too-large job, it must get by on charge. So, what is the most efficient power supply for charge?

Consider the amperage limit. Consider the voltage limit of 26.5+26.5vdc (long term testing cannot pass this figure). Consider that the LM1875 can vary between 28 watts and 89 watts reliant on charge, you could say that the charge is more important the the current by that extent. Perhaps it could perform better with a relevant power supply?

[danielwritesbac]

Originally Posted by pacificblue

. . .
Midpoint connection: Pt / Pdc = 1,5
Bridge rectification: Pt / Pdc = 1,23
. . .

After rectifier, these figures are in proportion to the voltage output (albeit slightly above average diodes they would be).

The smaller amperage dual secondaries making less voltage or the larger amperage center tap making more voltage. . .

I can't identify a description of efficiency with that.

[pacificblue]

Hey, long time no see.

Originally Posted by danielwritesbac

And lower voltage. With dual secondaries you can decrease the transformer amperage rating by the same proportion of the voltage drop. That's not relevant to efficiency on an amplifier of this scale.

Right. The current decrease that corresponds to the additional ~1 V voltage drop is pretty negligible.

Originally Posted by danielwritesbac

Steady tone figures are either irrelevant or the steady tone is an accident, in which case the device is morbidly inefficient.

Did you ever hear the piece 'Thus spoke Zarathustra' by Richard Strauss live? It starts with a 20 s long continuous sub-contra C from an organ. You have to hear it live, because most speakers will not reveal that tone at all. It has a frequency of 16,35 Hz and even 12" speakers have a hard time to replace enough air to make it as much as audible, even less reproduce it at realistic levels. Not an accident at all, but the composer's choice to represent nature in musical form.

Originally Posted by danielwritesbac

With its a too-small resource and too-large job, it must get by on charge.

I don't understand the meaning of this. You have to put in more, than what you want to come out. The transformer must supply more power than the amplifier consumes as an average. Even big capacitors can only buffer signals that are above the transformers power rating only for a very short time.

Originally Posted by danielwritesbac

So, what is the most efficient power supply for charge?

Power supply efficiency does not depend on what it supplies power to. For each application you need to find the right compromise. A power supply that supplies exactly the necessary amount of power is efficient in that every machine has its highest efficiency at its nominal power. E. g. a transformer may have 70 % efficiency at its nominal power, but only 50 % efficiency at 70 % of its power. Don't nail me on those numbers, they are only there to explain the principle. A big power supply is efficient in that bigger transformers generally have a higher efficiency than smaller ones. E. g. a small transformer may have 70 % efficiency at nominal power, while a bigger transformer may be 80 % efficient. If you want to improve the efficiency, you will not get around calculating, whether a specific application is better off with a fit-to-size transformer or an oversized one.

Originally Posted by danielwritesbac

Consider the amperage limit. Consider the voltage limit of 26.5+26.5vdc (long term testing cannot pass this figure). Consider that the LM1875 can vary between 28 watts and 89 watts reliant on charge, you could say that the charge is more important the the current by that extent. Perhaps it could perform better with a relevant power supply?

Power supplies do not have a sweet spot in the sense that too much is bad for the sonic performance. If you want a certain amount of output power, there is a minimum size of power supply you need. Let me put it in an exaggerated way and ignoring other aspects that also determine transformer and capacitor sizes. If you listen to music with little dynamics, use a big transformer and small capacitors. If you listen to very dynamic music, choose a small transformer and big capacitors. For an amplifier that works well with any type of music, use a big transformer and big capacitors.

Originally Posted by danielwritesbac

I can't identify a description of efficiency with that.

Efficiency is the relation of power input to power output. If you need to draw 150 W from the grid to get 60 W from the amplifier that is less efficient (eta = 0,4 or 40 %) than if you only draw 123 W from the grid and also get 60 W from the amplifier (eta = 0,49 or 49 %).

[danielwritesbac]

Originally Posted by pacificblue

Hey, long time no see.

Hi man! Thanks for the reply!

. . .
Did you ever hear the piece 'Thus spoke Zarathustra' by Richard Strauss live? It starts with a 20 s long continuous sub-contra C from an organ. You have to hear it live, because most speakers will not reveal that tone at all. It has a frequency of 16,35 Hz and even 12" speakers have a hard time to replace enough air to make it as much as audible, even less reproduce it at realistic levels. Not an accident at all, but the composer's choice to represent nature in musical form.

Yes, that does take current. Charge would run out during organ bass. There's almost no way to cheat when the bass notes are so big and so long.

. . . You have to put in more, than what you want to come out. The transformer must supply more power than the amplifier consumes as an average. Even big capacitors can only buffer signals that are above the transformers power rating only for a very short time.

I cannot figure out how to explain without getting terribly mixed up. All that I can say is that two near identical amplifiers with the same current, same wattage rating, same power supply board, same transformer, can differ on pulse power output by about 300% because of charge and other efficiencies. I'd love to know why but I probably couldn't understand the math.

Power supply efficiency does not depend on what it supplies power to. For each application you need to find the right compromise. A power supply that supplies exactly the necessary amount of power is efficient in that every machine has its highest efficiency at its nominal power. E. g. a transformer may have 70 % efficiency at its nominal power, but only 50 % efficiency at 70 % of its power. Don't nail me on those numbers, they are only there to explain the principle. A big power supply is efficient in that bigger transformers generally have a higher efficiency than smaller ones.

Although most documents say that a linear power supply is only efficient at maximum. . . instead, I believe you are correct.

I think that you said its efficient up until the make/break point of additional noise.

E. g. a small transformer may have 70 % efficiency at nominal power, while a bigger transformer may be 80 % efficient. If you want to improve the efficiency, you will not get around calculating, whether a specific application is better off with a fit-to-size transformer or an oversized one.

Desired output at the speaker jack comes into play. For full linear audio amp, chances are good that the larger transformer will be more efficient at facilitating the necessary tasks.

Power supplies do not have a sweet spot in the sense that too much is bad for the sonic performance. If you want a certain amount of output power, there is a minimum size of power supply you need. Let me put it in an exaggerated way and ignoring other aspects that also determine transformer and capacitor sizes. If you listen to music with little dynamics, use a big transformer and small capacitors. If you listen to very dynamic music, choose a small transformer and big capacitors. For an amplifier that works well with any type of music, use a big transformer and big capacitors.

Hmm. There so much to this that I don't know. But, results support your statement quite well.

Um, but a Far too large transformer will need a large RC network, and that could reduce efficiency a bit.

Efficiency is the relation of power input to power output. If you need to draw 150 W from the grid to get 60 W from the amplifier that is less efficient (eta = 0,4 or 40 %) than if you only draw 123 W from the grid and also get 60 W from the amplifier (eta = 0,49 or 49 %).

We almost agree, but the scope is different. . . I think that Efficiency relates to power in at the plug spent towards desirable output at the speaker cone. And, that power really should have a purpose.
This applies nicely to this thread because chipamps are most suited to active speakers, since the chipamp design is sealed shut inside the chip, thus whatever workarounds aren't with the power circuit are with the speaker itself.

Thank you for the cohesive reply. It makes sense and some of it relates directly to my question on charge. Its the dynamics! When the transients on LM1875 can shoot up to 89 watts, that's a heck of a lot of charge. I'm not so familar with that topic, but it seems important for an amplifier with such limited resources to have every resource you can give it and also important to make the difference between an exciting "its like live" presentation versus boredom listening "to" "canned" audio of a speaker. Can you expand your comments on power supply construction and its effect on dynamics?

[AndrewT]

Originally Posted by danielwritesbac

And lower voltage. With dual secondaries you can decrease the transformer amperage rating by the same proportion of the voltage drop. That's not relevant to efficiency on an amplifier of this scale.

Steady tone figures are either irrelevant or the steady tone is an accident, in which case the device is morbidly inefficient.

That doesn't relate to a design goal.

LM1875's have a too-small tolerance for current, thus limiting the importance of current by that amount. With its a too-small resource and too-large job, it must get by on charge. So, what is the most efficient power supply for charge?

Consider the amperage limit. Consider the voltage limit of 26.5+26.5vdc (long term testing cannot pass this figure). Consider that the LM1875 can vary between 28 watts and 89 watts reliant on charge, you could say that the charge is more important the the current by that extent. Perhaps it could perform better with a relevant power supply?

does this mean anything to anyone else?
Or is this another example of Daniel's version of the physics and science behind electron flow?

[Speedskater]

It's just "Daniel's World" engineering.

[pacificblue]

Originally Posted by danielwritesbac

two near identical amplifiers with the same current, same wattage rating, same power supply board, same transformer, can differ on pulse power output by about 300% because of charge and other efficiencies. I'd love to know why but I probably couldn't understand the math.

That sounds very improbable. What exactly are the differences between those amplifiers.

Try to explain, what you think 'charge' is. From a technical point of view it is the energy that is stored in the capacitors, and if you have the same power supply board and same transformer, the charge will be the same.

Originally Posted by danielwritesbac

Although most documents say that a linear power supply is only efficient at maximum.

Not only. It is most efficient at maximum.

Originally Posted by danielwritesbac

I think that you said its efficient up until the make/break point of additional noise.

Efficiency is not a digital concept that is either on or off, but an analog concept that is higher or lower, but rarely zero and never 100 % as far as human knowledge goes. Depending on the actually used output power there are situations, where the smaller transformer has a higher efficiency, and others, where the bigger one is more efficient. For an amplifier design you need to assess, in which situation the amplifier will mostly be used. Or you can simply assume that the differences are not big enough to worry much about. The difference in efficiency from one transformer to another is far from the gain in efficiency you can get from an SMPS. But even SMPS have their best efficiency at their nominal rating and many fall short, when you use them at very low powers, like the usual listening level of an amplifier at home.

Originally Posted by danielwritesbac

Desired output at the speaker jack comes into play. For full linear audio amp, chances are good that the larger transformer will be more efficient at facilitating the necessary tasks.

Transformer size limits the achievable output power.

Originally Posted by danielwritesbac

Um, but a Far too large transformer will need a large RC network, and that could reduce efficiency a bit.

Only the R part of the RC network would reduce efficiency, but transformer size and the need for an RC network are not directly related.

Originally Posted by danielwritesbac

Can you expand your comments on power supply construction and its effect on dynamics?

Dynamics refer to the difference between the lowest and the loudest sound. The question you have to answer for yourself is, are those changes fast or slow and do they happen often or seldom in the music you listen to. When you listen to organ concerts you have very loud passages that remain loud for a long time. When you listen to a drum solo, you get lots of short peaks. To remain with my simplified comments from above, you need a big transformer and small capacitors for the organ and a small transformer with big capacitors for the drums. Use a big transformer and big capacitors, if you want to listen to either instrument every now and then.

If transformer and capacitors have a reasonable size you won't gain much by stepping them up further. Once the transformer rating is three times the amplifier output power and the capacitors are in the tens of thousands of µF range, you will in most cases have achieved what dynamics you can get out of an amplifier. It is the usual game of diminishing returns.

The LM1875 like all chipamps has a voltage rating that limits, what dynamics you can achieve and a current limiter that also restricts, how much dynamics you can achieve. If that IC cannot fulfill your expectations, you need a different amplifier with higher voltage rating and a current limiter with higher threshold or no current limiter at all. Or more efficient speakers.

[danielwritesbac]

Originally Posted by pacificblue

That sounds very improbable. What exactly are the differences between those amplifiers.

I was in error--The power supply board is different in order to make factor 3x difference in dynamics between two similar amplifiers.

. . .But even SMPS have their best efficiency at their nominal rating and many fall short, when you use them at very low powers, like the usual listening level of an amplifier at home.

I seem to have SMPS fall short in a different way.

. . .The question you have to answer for yourself is, are those changes fast or slow and do they happen often or seldom in the music you listen to. When you listen to organ concerts you have very loud passages that remain loud for a long time. When you listen to a drum solo, you get lots of short peaks. To remain with my simplified comments from above, you need a big transformer and small capacitors for the organ and a small transformer with big capacitors for the drums. Use a big transformer and big capacitors, if you want to listen to either instrument every now and then.
. . .

An excellent clue! Thank you very much!!!

Perhaps: bridge rectifier > big caps > slight resistance > small caps > amp ?

The LM1875 like all chipamps has a voltage rating that limits, what dynamics you can achieve and a current limiter that also restricts, how much dynamics you can achieve. If that IC cannot fulfill your expectations, you need a different amplifier with higher voltage rating and a current limiter with higher threshold or no current limiter at all. Or more efficient speakers.

Or a fet?

[danielwritesbac]

Originally Posted by Speedskater

It's just "Daniel's World" engineering.

My hope is that its not boring.

[pacificblue]

Originally Posted by danielwritesbac

Perhaps: bridge rectifier > big caps > slight resistance > small caps > amp ?

That reduces ripple and rail voltage. You will have slightly less dynamics, but may still get better sound thanks to reduced power supply noise.


Or a fet?[/QUOTE]

Which makes it a different amplifier.