T-network: the better feedback solution?

Status
Not open for further replies.
Jump to Latest
#441
Yes, the hiss was only audible close to the speakers in my case. But I think that we should aim for total silence"
Click to expand...
Hmmm, I have been struggling with a faint buzz! I thought there was something wrong with my amp 😉 Glad to hear I am not the only one🙂 Well a lot was learned in the troubleshooting. It all wound up to something like this: The noise on the outputs were caused be the rectifier bridge. I replaced it with a P2P rectifier. After that, the amp was completely silent! Except for music, to the relief of some people😉 I tried pretty much to change everything else in advance of that. Then I added a pair of 3.3uF mkp's to the inputs, and then there was a faint hiss! I am sure the big MKP's acts as some kind of antennae, picking up noise! What are you people using for input caps? I need them, since all kind of sources wil be connected to the amp, PC, discman and so on.

Steen.
 
#442
Apologies if this subject has been raised earlier in the thread ( I have only read the beginning and the end), but are you aware that use of the T-network reduces the bandwidth of the amp?

My simulations (with an OPA604) show a reduction in bandwidth from >200KHz with "normal" feedback to about 60KHz using the T-Network with Franz's values (47K, 18K, 18K, 150R). I suspect these figures will be lower with a power op-amp like the 3886.

Is this intended?
 
#443
steenoe said:

Hmmm, I have been struggling with a faint buzz! I thought there was something wrong with my amp 😉 Glad to hear I am not the only one🙂 Well a lot was learned in the troubleshooting. It all wound up to something like this: The noise on the outputs were caused be the rectifier bridge. I replaced it with a P2P rectifier. After that, the amp was completely silent! Except for music, to the relief of some people😉 I tried pretty much to change everything else in advance of that. Then I added a pair of 3.3uF mkp's to the inputs, and then there was a faint hiss! I am sure the big MKP's acts as some kind of antennae, picking up noise! What are you people using for input caps? I need them, since all kind of sources wil be connected to the amp, PC, discman and so on.

Steen.
Click to expand...



Steen, when you added the input cap, did you change the grounding of the input/cable/plug?
What happens when you ground the input pin of the input cap?

You experiences with the rect point to ground loop hum/hiss. The fact that it now comes back may have the same cause.

Jan Didden
 
#446
Dave S said:


Quite possibly, however....

The point is that nobody seems to want to investigate the AC performance of their LM3886 implementations. All I see is endless ramblings about the (not particularly relevant) dc performance!
Click to expand...

OK, here goes. Adjust your T-Network to a gain of 90 and thus get no feedback above 70KHz. Of course you will have too much gain, but in my case I have a CD/SACD Player that is -11dB (560mV) and the gain is welcome. If you got too much gain use some input padding down. Not ideal but if the trade-off is better sound, then it gain gain gain sound wise.

Since the chip is internally compensated (the dominant AC characteristic) when gain is = 90 ninety then the open loop becomes equal or lower around 70KHz, hence no more feedback. Must have the layout right, grounding etc perfect, as PSRR at high frequencies is now poor, but good layout should prevent buzzes.

Less feedback mean less slew rate induced distortion. This happen at HF where the feedback cannot quite keep up. But no feedback means slewing can't happen above 70KHz. Does it sound better? Try and see for yourself.

Joe R.
 
#451
I'm jumping into what appears to be a very long thread, so forgive me if I cover material that's already been discussed. The tee network can be used in place of a single resistor in any application that has a known source resistance and a virtual ground load. Let R1, R2 be the series elements of the tee, and R3 be the parallel element. The exact equivalent resistance of the tee can be calculated as Re = R1 + R2 + R1R2/R3 (Eq 1). So, for the original Gainclone amplifier, R1 = R2 = 10K; R3 = 100 Ohms. So Re = 10 + 10 + 10X10/.1 = 1020K = 1.02 Meg. You can design a tee easily. Arbitralily select two of the resistors. Calculate the third by solving Eq 1 for the remaining resistor.
.
The tee network has the disadvantage of producing a larger output error due to the Op-Amp input offset voltage than is produced using a single feedback resistor.
.
As far as bandwidth is concerned, it is the same, regardless of whether a single resistor is used or an equivalent tee is used. Connect a capacitor from the output to the inverting input, and the erffect on the bandwidth will be exactly the same whether a single resistor or the equivalent tee is used.
 
Status
Not open for further replies.
Top Bottom