Help with first Power Supply for LM1875

[danielwritesbac]

Originally Posted by pacificblue

Which makes it a different amplifier.

Question: Since Lm1875 is an opamp, can it be used in a circuit like this, but yet scaled up to drive a speaker?: HDOA - High Dynamics Op Amp

[Juergen Knoop]

yes, of course. It's a standard circuit you find in textbooks. Mostly done as power amp with TDA2030 which is almost equal to the LM1875.

Just an OT question Daniel, you are writing or have written your bac paper/thesis? What's your discipline?
Regards

[danielwritesbac]

Originally Posted by pacificblue

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

So, this example does simply more powerful audio via avoiding amplification of power supply noise (its finite resources doing more audio, less noise). Its a basic increase of efficiency?

Okay, so how do we do both "slightly more dynamics" and also less noise at the same time?

[danielwritesbac]

Originally Posted by Juergen Knoop

yes, of course. It's a standard circuit you find in textbooks. Mostly done as power amp with TDA2030 which is almost equal to the LM1875.

Just an OT question Daniel, you are writing or have written your bac paper/thesis? What's your discipline?
Regards

Many papers written in my youth, when I was in college, and absolutely none of them were of any personal interest.

Discipline is computer network engineer. As a theme, I'll either push a problem until it breaks or is repaired and either way gets repaired--half measures grieve me greatly.

The reason that I would obsess over LM1875 is because it doesn't fall short on anything except for output power. And, except for output power, its the chipamp that doesn't compromise fidelity. Usually, the harmonic output of LM1875 amplifiers is just the sort that you'd like to increase by bridging, but the current rating is too little unless you have 16 ohm speakers.

[pacificblue]

Originally Posted by danielwritesbac

Okay, so how do we do both "slightly more dynamics" and also less noise at the same time?

More dynamics
-> higher rail voltage in all load situations = big transformer and big low ESR capacitors.
-> efficient speakers.

Less noise
-> big capacitors or
-> cascaded RC circuits as per your post #88 or
-> LC circuits as recommended by Nelson pass or
-> regulated power supplies with noise at different frequencies, where it is less audible or easier to reduce.

[Juergen Knoop]

Originally Posted by danielwritesbac

The reason that I would obsess over LM1875 is because it doesn't fall short on anything except for output power. And, except for output power, its the chipamp that doesn't compromise fidelity. Usually, the harmonic output of LM1875 amplifiers is just the sort that you'd like to increase by bridging, but the current rating is too little unless you have 16 ohm speakers.

I have a pcb somewhere for bridged and transistor boosted TDA2030s good for 200W with a 4 ohm load . Bought it from a kit supplier.
But I don't think, it would preserve the 'harmonic output'.
Regards

[danielwritesbac]

Originally Posted by Juergen Knoop

I have a pcb somewhere for bridged and transistor boosted TDA2030s good for 200W with a 4 ohm load. Bought it from a kit supplier. But I don't think, it would preserve the 'harmonic output'.
Regards

Is that the circuit that has no thermal compensation?
If so, that's the TDA2030A (unity stable version) and does about 120 watts of music. Perhaps that schematic could be revised to use LM1875's on regulated power and perhaps parallel output devices, maybe fets? I don't know.

What I do know is this: Once the new owner of an LM1875 amp hears it, he or she will want to crank it up. . . and then there's a problem unless the speakers are very efficient indeed. A possible alternative is an array speaker with as many LM1875's as there are speakers.

[danielwritesbac]

Originally Posted by simonlarusso

Also the IC isolator which came with the kit is a rubber pad about 2-3 mm thick. . . .

I think that you've described the QK50 "K50" LM1875 kit.

The following post contains comments about K50 and refers only to K50.

A to220 size mica with thermal compound should work a bit better. This also needs a plastic "shoulder washer" to insulate the screw.

The K50 has typically a feedback resistor (R5) clear up at 180k, which is a bit bizzare. A value that's not so hindersome is 120k.

You can set the gain with its partner, R4, whereby a smaller value at R4 will increase the gain or a larger value at R4 will decrease the gain.

R4 on the K50 kit is AC coupled for safety. Unfortunately its coupled to a very low quality capacitor, C3, of a too-small size. A better quality cap of 47uF or more is recommendable than a poor quality cap of 22uF. However, you may choose to leave the factory-provided cap in place and parallel (add to it) a tiny value polyester/mylar capacitor to level out its frequency response. Choices on the value of this cap affect the mid-bass versus low-bass contour and thus the choice of size is speaker-reliant.

C1, the input filter cap at 1uF is somewhat undersize unless your woofers are tiny indeed.
Choose this cap size based on your speaker's abilities as there is no need to waste power for what a given speaker cannot produce.
To maintain the classic sound, you can use an electrolytic if you wish; however, classic hifi with electrolytic input caps did, historically, also use a tiny value polyester/mylar "bypass cap". One possible improvement is 3.3uF electrolytic with 10nF (0.01uF) polyester added. There are many, many options for C1, but a low quality 1uF isn't optimal.

The factory standard K50 LM1875 kits have a nice frequency response but a "pancake flat" soundstage. The above mods are for the purpose of alleviating the soundstage issue so that you can enjoy both mono and stereo in a hi-fi presentation.

In a nutshell, I recommended un-doing the factory treble boost by swapping their 180k for a more reasonable value of 120k (afterwards correct the gain with its partner resistor, if necessary). Then, putting the treble boost back into place via adding tiny value polyester/mylar capacitors to both the NFB cap and the input cap to complete their bandwidth. You should be able to "break even" on the frequency response and the LM1875 will now operate within tolerances, better able to obey the signal at its input. In that case, each speaker will be able to project in 3d, instead of flat up against the speaker driver.

Materials to fix K50 kit:
Variety pack of resistors, including a 120k
Better cap for the nfb cap
Better cap for the input cap
Cheap polyester bubble/dip caps in the range of 4.7nF, 10nF and 22nF for the purpose of paralleling (adding) to input caps and NFB caps.
Mica of TO220 size
plastic shoulder washers
thermal compound

There are many other options available, but LM1875's are so very good that this simple touch-up can give it a chance to show off.

[danielwritesbac]

Just starting out? Perhaps this chart is helpful.

[Mazeppa]

Originally Posted by danielwritesbac

Just starting out? Perhaps this chart is helpful.

Yes I am and yes it is. The bar markings were previously undecipherable to me.
Thank you.