repost -- son of screamer schematic differences

[mordechai]

The schematic for this circuit on the beavisaudio site and in RG's "Technology of the Tubescreamer article" have some significant differences.  Here are links to each schematic:

http://www.geofex.com/article_folders/tstech/tsxtech.htm  (SCROLL DOWN TO THE BOTTOM OF THE ARTICLE)

http://beavisaudio.com/bboard/projects/bbp_BoutiqueTubeScreamer_Rev1_1.pdf


I'm wondering if anyone can please explain the following:

1. On RG's schematic, the power section seems more complex.  There is a series resistor of unspecified value coming off the battery line, and there are two 47uF electrolytics in the power section here.  In the Beavis version, these items are missing. 

2.  In RG's schematic, the voltage dividing resistors are valued at 10K; on the Beavis version, they are 100K.

3.  What purpose is served by the unspecified series resistor in RG's version that is absent from the Beavis version? 

[ubersam]

1. On RG's schematic, the power section seems more complex.  There is a series resistor of unspecified value coming off the battery line, and there are two 47uF electrolytics in the power section here.  In the Beavis version, these items are missing.
...
3.  What purpose is served by the unspecified series resistor in RG's version that is absent from the Beavis version?

The unspecified R serves to give the circuit a little isolation from the power source, and along with the first 47uF cap provides a degree of filtering for the incoming power. I usually use a small resistor, 47-100 ohms, and a bigger cap, 100-220uF

2.  In RG's schematic, the voltage dividing resistors are valued at 10K; on the Beavis version, they are 100K.

This should explain that: http://www.geofex.com/circuits/Biasnet.htm





[/quote]

[mordechai]

Thanks...that URL does explain quite a bit, but I am still not totally clear.  A 10K/10K network vs. a 100K/100K network seems to be sending a very different amount of current into the Op Amp (45uA vs. 450uA if I remember the comment in the article correctly).  How is it that the same SOS circuit can take either option and still function in a consistent manner?  I guess what I'd like to know is which option, the 10K/10K or 100K/100K, provides a better scenario for stability, lower current draw, lower noise, etc.

[ubersam]

The bias network isn't really "sending" any current to the opamp, think of it more as having current available for the opamp's input bias requirements. The R values for the bias network are calculated so that there is enough current available so that when the opamp draws its input bias current the voltage at the junction will not fluctuate significantly (<10%). According to the datasheet, the typical input bias current of a JRC4558 is 25nA (or 0.025uA) which is less than 1% of either 45uA or 450uA. Either network would provide sufficiently stable Vref & bias current for opamp in the TS/SOS circuits.

As for noise, both networks have a big cap to ground at the junction that sinks noise from the power supply. Look up noiseless biasing.

[mordechai]

That really helps.  Thanks so much.