Mixer-Amp - TDA2050 - Critique?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Would any kind guru care to take a look at this schematic and, as you will, tear it to pieces? Explanations of how I don't know what I'm doing will be much appreciated.

As I understand, the input impedance of an inverting amplifier is the value of the input resistor (in this case R1a, R2a,...). It makes no difference what's added before the input resistor, the input impedance is always the value of the input resistor, period. Or...does anybody know different?
 

Attachments

  • schematic_01_c.jpg
    schematic_01_c.jpg
    81.1 KB · Views: 209
Last edited:
Thanks for your reply, gootee. Yes I do have those appnotes, much annotated.*

My intention from the first was to have a summing amp (called a virtual-earth mixer in the schematic) feeding master volume control Potvol. I intended that U1 and U2 (and others, indicated by MORE) would be summed by U3, then sent to Potvol.

But from what you say I've missed the mark? I'm glad to find that out.

But look as I may, and referring to the appnotes you mention, I can't find what I'm doing wrong. I've redrawn my schematic in this vain hunt (which I'm uploading), but still no luck. May I have a hint, please?

* For those who don't know, "Foxit Reader" is smaller, faster, and less intrusive than the huge Adobe Reader. Also it allows highlighting text, and saveing the highlights.
 

Attachments

  • schematic_01_d_redrawn_c.jpg
    schematic_01_d_redrawn_c.jpg
    87.7 KB · Views: 183
Last edited:
Would any kind guru care to take a look at this schematic and, as you will, tear it to pieces? Explanations of how I don't know what I'm doing will be much appreciated.

As I understand, the input impedance of an inverting amplifier is the value of the input resistor (in this case R1a, R2a,...). It makes no difference what's added before the input resistor, the input impedance is always the value of the input resistor, period. Or...does anybody know different?

Save yourself some amps:
 

Attachments

  • picbuck-mixer.JPG
    picbuck-mixer.JPG
    63.4 KB · Views: 164
Some 200 people were interested in this thread, so it seems well to close things out on a high note, meaning situation resolved.

Thanks gootee, for your input. And thanks DUG for pointing out the cheap, simple, and equally effective way, although with no amplification.

Meanwhile, I ran across essentially the same situation resolved by real, grownup engineers with degrees, or so I confidently assume. This is from a company whose focus is the pro (not consumer) market. To avoid any teensy little irritations like copyright violation I redrew the schematic, and let the company name remain hidden in the mist, although I doubt very much that they'd mind. Heck, they'd probably come around and help.

Notice that the factory schematic (labeled as such) is direct- (D.C.) coupled. In an add-on version (also labeled) I added a blocking capacitor and its associated resistor. LTspice was OK with this change, but I don't know how it might work out with real parts.

Hope this might be of interest.
 

Attachments

  • factory_01_a.jpg
    factory_01_a.jpg
    65.6 KB · Views: 113
  • added_01_b.jpg
    added_01_b.jpg
    80.2 KB · Views: 116
dc coupling is .. well it can be bad, and it can be good.
it realy depends. if the source has otput capacitors (I.E. no DC at output) then capacitor-less coupling is not bad at all.
if the source has a verry minimal DC , even then it may still not be bad, it all depends on the headroom and the DC level. One might "just " block the DC at the last output of the whole device. Like... having a small 2 channel mixer and an EQ , and a chipamp as a single unit, You may - if it was built properly at least- just use de-coupling at the input of the channels. But not between the mixer-> EQ, EQ->Poweramp.

Other than that input decoupling capacitors are not a bad idea.
Most probably since it may or may not be needed -at least i think- it was not presented on the factory schematic.
But, i must agree, they might have included it markt with an axterix and make a notice, that it is optional, based on actualy circuitry enviroment.
 
Everything you said remaining true, it's a sad fact that the days of published schematics are largely behind us. Usually we shadetree techs (like a shadetree mechanic, but the parts are not as heavy) know nothing about the output circuitry of whatever might be our audio source.

Making things worse, our source tends to be consumer-grade equipment, and we don't know who had the last vote before production started: was it the engineers, or was it the bean counters?

We hear of headphone amps that won't drive headphones, and CD players that won't play CDs. These days it's astonishingly cheap and easy for a factory to turn out first-rate audio, and yet some factories don't do it.

I vote to err on the side of caution. Expect that $29 player to launch a nuclear attack against the amp you spent a week building, and take measures.
 
<< It's confusing, which one is the definitive schematic you want comments on? >>

Very good question. This thread has kind of veered back and forth.

The posts are numbered in the upper-right corner of each post.

My request for comments involved only the schematic in Post #1. All others are discussion.

For clarity:

Post #1 was my original request, and the schematic with that post is the original.

Post #3 is the same schematic redrawn. I posted this in reply to another post.

Post #4 is the [valid] suggestion by DUG.

Post #6 is a final solution (in my opinion), involving a commercial circuit I ran across.
 
Last edited:
OK.
a) I find so called "simulated log" a cheesy solution; getting the proper Log pots is worth it.
Worst case, I prefer the straight Lin pots, go figure.
b) some parts are unnecessary.
Remove Rlog1 , R1a , R1b , R1c, U1 .
Connect Pot 1a wiper > C1a > R1c .
Do the same on Ch2 .
c) if possible get Potvol (25K) in Log taper.
In that case remove Rvol.
Good luck.
 
Thanks for your interest, but you're suggesting the same circuit described by DUG in post #4. A perfectly legitimate mixer circuit, but not the one presented here.

The circuit (not my own) in my own post #6 is a factory circuit guaranteed (I'm confident) to work.

I personally think the log-fake pot idea is kinda neat. But I wouldn't say somebody is wrong if they prefer store-bought to roll your own.
 
Thanks for your interest, but you're suggesting the same circuit described by DUG in post #4. A perfectly legitimate mixer circuit, but not the one presented here.
Of course it's not the one in post #1.
I think the point is to suggest corrections , am I missing something?
And if 2 answers match, maybe it means something ;)

By the way, you complain of
And thanks DUG for pointing out the cheap, simple, and equally effective way, although with no amplification.
Well, yours doesn't provide amplification either :D , that's why we both suggested you remove those unneeded unity gain blocks.

The circuit (not my own) in my own post #6 is a factory circuit guaranteed (I'm confident) to work.
It *does* work, who said it doesn't? ... but still has useless unity gain blocks.
To be more precise, *they* provide the marginal advantage of presenting 10K input impedance while using 5K pots ...but yours does not.

I personally think the log-fake pot idea is kinda neat. But I wouldn't say somebody is wrong if they prefer store-bought to roll your own.
Roll your own would be to open the pot and paint a lower resistance track on the lower half; the crude idea of putting a lower resistance value from cursos to ground is *not* the same.
It has the terrible disadvantage of making the pot value "seen" by the previous 1/10th of what is expected:eek:, throwing calculations, frequency cutoff, etc. out of whack.
That "idea" does not work on:
> "Ampeg" (James) type tone controls
> "Fender" type tone controls
> "Vox" type tone controls
> Gain/Distortion controls as used on 95% of Distortion pedals
> Gain controls in live/recording Mixers
> parametric equalizers
> Audio Oscillator Frequency setting
.
.
> 1000000 other circuits where a *real* Log control is needed.

Yes, same pot may cost 1$ if linear and 3$ if log. So what?

By the way, I buy in bulk, straight from the Importer, and pay 85 cents any type.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.