Power amplifier on NE5532 - ty.ytka

katakrowa wrote:

To make it modern, it begs to be built on op amps in class D

What to do, it's best to buy any Chinese D-class module for PLN 8.88 and off you go.

austriackimalarz wrote:

I would suggest converting some polyphase converter from the GPU power supply to the amplifier

Today, in fact, the effect would probably be similar, but as you can see, the heat emitted by the WO is a big problem, as evidenced by this whole topic.
I bet that within 10 years there will be such class D WOs that will be indistinguishable from analog ones in the audio band.

katakrowa wrote:

I bet that within 10 years there will be such WOs in class D, which in the audio band will be indistinguishable from analog ones.

They are like that today. Interestingly, these are still "single-phase" models, and they do not switch at megahertz, but closer to the standard 300-500kHz.

austriackimalarz wrote:

They are like that today. Interestingly, these are still "single-phase" models, and they do not switch at megahertz, but closer to the standard 300-500kHz.

Can you provide some model?

yogi009 wrote:

I wouldn't get into quad dice for several reasons. Cooling, crosstalk, the ability to replace the system with another

In theory you're right, but:
1. The sum of the power dissipated will be rather similar in each case, so the surface area of the heatsink will also be. In the absence of a heatsink, double, and even better single, will do better.
2. Crosstalk... But in each opamp will be the same...
3. I was thinking about the SMT version, so no one will play with resoldering ICs. At most, it would be possible to use the design of the board.
In general, I treat the project as a curiosity, with the option of making two or more versions, e.g. "high end" on NE5532 and better components (no exaggerations in the style of kondy for several hundred zlotys) and "popular" on SMT and just TL074 in such a version. It will be interesting to compare such constructions both in terms of measurement and hearing.
Why such opamps? Let's say I'll trust Douglas Self, who quite clearly justifies the lack of sense in using "better" chips than the NE5532 in general and the TL07x in applications requiring input FETs.

austriackimalarz wrote:

We settled on the TL071HIDBVR, which is the well-known TL071 in the SOT-23 housing

I'm not sure it's a good choice. The documentation contains only a functional diagram, but if the modern versions do not differ much from the earlier TL06x TL07x and TL08x, they have a primitive and inescapable "short-circuit protection" in the form of a final stage with 100-ohm emitter resistors and another 200-300 Ohm resistor connected in series with the WO output to load. It does not matter much in typical applications of these WO, but here we have an unusual application, because even for the parallel connection of several dozen or even several hundred such circuits with a load in the form of a loudspeaker, this will have an impact limiting the output power and the Damping Factor.

Jaro wrote:

austriackimalarz wrote:

We settled on the TL071HIDBVR, which is the well-known TL071 in the SOT-23 housing

I'm not sure it's a good choice. The documentation contains only a functional diagram, but if the modern versions do not differ much from the earlier TL06x TL07x and TL08x, they have a primitive and inescapable "short-circuit protection" in the form of a final stage with 100-ohm emitter resistors and another 200-300 Ohm resistor connected in series with the WO output to load. It does not matter much in typical applications of these WO, but here we have an unusual application, because even for the parallel connection of several dozen or even several hundred such circuits with a load in the form of a loudspeaker, this will have an impact limiting the output power and the Damping Factor.

Thanks for the warning. In fact, in this situation you would have to go into thousands of WO. But a fellow Austrian painter has just such an intention. I'm even more curious about the end result

marweg1967 wrote:

Will your colleague share details about this security? Input diodes? Because this Behringer makes a really bad noise and I'm getting ready to etch the board to use the workaround described by my friend (I already have the parts). So I would add this protection right away.

Please check the topic again link , I posted new information.

tytka wrote:

marweg1967 wrote:

Will your colleague share details about this security? Input diodes? Because this Behringer really hums badly and I'm getting ready to etch the board to use the workaround described by a colleague (I already have parts). So I would add this protection right away.

Please check the topic again link , I posted new information.

I looked I imagined something more or less like that, except that I was thinking about an anti-parallel connection of sets of ordinary diodes (3-4 in series), or JFETs working as diodes. I wonder if these Zener tubes don't make any noticeable noise, although probably not since you left it like that?

I would like to thank my colleague for keeping his word and best regards

Hi.
So I had a vacation. I will say this - expect an article "The end of a thousand amplifiers" within a week or two It turned out beautiful.

In that case, I'm waiting for my colleague's article. I'd love to see a different approach to this ending than mine.

w3501yyyy wrote:

A block of 250 TL072 units connected in parallel, a total of 500 amplifiers, consumes 1A of current when powered by +-18V.

Just enough? And maybe not more?
And that's only at rest. When you drive this, the demand will be even greater.
I power my "baby" (probably only 140 WO in total) from a 100W transformer.

I signaled that it was heating up

AVE...

With 250 TL072, the maximum quiescent current can be up to 1.25A. The minimum is 625mA. 1A is a good result. Fortunately, with typical speaker impedances and this supply voltage, the maximum power consumption will not increase significantly above this value. For a speaker with an impedance of 4Ω, the maximum power will be approximately 30W. This means additional current consumption of ~840mA per channel. A transformer for such a stereo amplifier should have a current efficiency of ~4.1A and a power of ~150W, with up to 90W dissipated in the ICs in the idle state. It'll be warm...

AVE...

Well, in such a scenario you will need a transformer of at least 400W, but the output power of the amplifier will be 110W per channel into 4Ω. It will definitely be an interesting construction...

AVE...

According to Note, you will lose about 3V of voltage relative to the power supply, i.e. in the bridge, 36V will become 30V. I calculated it for you exactly, because just two hours ago I was editing a text about bridge amplifiers, and I calculated it based on it. So you will need 400W, 200W per channel, assuming that the WO quiescent current will be maximum for each piece. Always assume the worst possible scenario...