Op-amp with darlington in feedback loop.

Status
Not open for further replies.
Jump to Latest
#22
EUVL said:
If you want something simple, use a LH0033 Sziklai.

Ðèæñêèé áóôåð-3.
http://www.diyaudio.com/forums/head...portable-headphone-amplifier.html#post4760389


Patrick
Click to expand...

It's more complicated than op-amp+one transistor.

I will use BD139 or something better, bias it using LM317 to around 500mA, throw in huge heatsink and this should suffice. I just can't think of something simpler (like I said before - I need to use ADA46x7-1 in this one, I just can't let this chip sits on the shelf).
 
#24
tybysh said:
Let's look at Douglas Self design from "Small Signal Audio Design" (I hope It's ok to put it here):
An externally hosted image should be here but it was not working when we last tested it.
Click to expand...
Ah, that's along the lines of what I was talking about earlier.
(Depending on opamp, a series resistor between opamp output and buffer input may be needed. The one used here has several hundred ohms already built-in, in some other cases an external 47-100 may prove helpful. Reserving a spot for a double-digit pF capacitor from opamp output to inverting output is also advisable.)
tybysh said:
An externally hosted image should be here but it was not working when we last tested it.


Is this really that bad thd? Can You hear 0.01% distortions?
Click to expand...
Not likely. Most of that actually ought to be coming from the TL072 if you compare his opamp measurements. A well-biased buffer tends to perform extremely well.
tybysh said:
I'm still leaning to one NPN inside the feedback loop, biased to class A using a simple resistor, or even simple CCS made from voltage regulator.
Click to expand...
Actually there isn't too much wrong with an EF + CCS buffer. I think it was our Bonsai who suggested (and simulated / measured) such a circuit here. Efficiency is not exactly its strength though - you'll need twice the bias current of the equivalent push-pull Class A stage, so it wouldn't be unusual to see 150 mA of quiescent current here, with heatsinking to match (that's a few watts to dissipate there).

Note that an LM317 CCS will appreciate some series resistance when more than a few volts are dropping over it, otherwise it may oscillate. EFs also tend to be more finicky with capacitive loading than MOSFET followers (ever wonder why a Szekeres is a DIY classic while there are hardly any similar projects involving BJTs?).

As you can probably tell, even a simple circuit tends to require some extra measures to make it work well in real life.
 
#25
sgrossklass said:
(Depending on opamp, a series resistor between opamp output and buffer input may be needed. The one used here has several hundred ohms already built-in, in some other cases an external 47-100 may prove helpful. Reserving a spot for a double-digit pF capacitor from opamp output to inverting output is also advisable.)
Click to expand...

I always use resistors on op-amp input/outputs and transistor base.


sgrossklass said:
Note that an LM317 CCS will appreciate some series resistance when more than a few volts are dropping over it, otherwise it may oscillate. EFs also tend to be more finicky with capacitive loading than MOSFET followers (ever wonder why a Szekeres is a DIY classic while there are hardly any similar projects involving BJTs?).
Click to expand...

Can You say something more about that LM317 CCS configuration? Maybe use something like this:

jung_13c-scaled1000.jpg

(source of image: https://vwws.wordpress.com/2011/06/21/dn2540-lm317-cascode-ccs/)
?
 
Status
Not open for further replies.
Top Bottom