Bootstrapping question... maybe for Gus?

[earthtonesaudio]

I've just started experimenting with bootstrapping, and I'm noticing that the source/emitter bypass cap seems to roll off highs if you attach it directly at the source/emitter, but not if you have some resistance in there.

Anyone have experience with bootstrapping, either in buffer or gain stages, who would like to comment on why this might be?

I also noticed the beginner project is bootstrapped so I figured I'd see if Gus wanted to chime in.

Thanks,
Alex

[Sir H C]

You can't bypass the emitter/source resistor and have bootstrapping work.  That cap makes source/emitter an AC ground, so now you have that resistor from the bootstrap shorted to ground at AC frequencies, lower R as frequency goes up, killing your highs.

[earthtonesaudio]

I guess I mis-typed when I wrote "bypass cap."  I am working with the standard bootstrap circuit, where you take signal from the source/emitter and feed it into the bias divider network, which is connected to the gate/base via a resistor. 

What I want to know is, when I have some resistance between the emitter/source and the "bootstrap cap," it seems to work fine, but if I eliminate that resistance and take the cap directly off the emitter/source, then there is a high frequency reduction.  I'm using a 1M resistor from the voltage divider to the gate/base, so adding 100 ohms or so in there shouldn't change the roll off frequency that much, unless there's something else going on that I'm not aware of.

[Sir H C]

Is the cap directly connected from the base to the emitter?  A drawing of the circuit would help.

[frank_p]

This is the way I see it but I might be wrong (I am still breaking my head on that circuit).

If you put the resistor it will attenuate the amplitude comming from the emitter going to the base bias.
So, if you take from the emitter, you take the "X" of the same phase output (but filtered) and re-inject in the base, what you get ?
If you put a resistance, the effect of filtering  the signal will be less, and will have less effect on filtering high frequencies.

So what happens is that the impedance at the input gets higher with bootstrapping, but if you put a resistance in series with the bootstrapping cap  the effect of bootstrapping becomes efficient in essence.  That is the loading of the input impedance is now less large for AC signal with the added resistance than it would be with only the cap.

The more the resistance the less current you bring to the base. The more the current gets to zero in the bias voltage of the base, the more the AC resistance is high ( R = V(bias) / I (feedback) ).  The increase of R in AC resistance augmentation is then multiplied many times.

So the more R is high the more the gain stage acts like a buffer.  In fact it's like if the voltage divider was becoming a buffer in itself: you got to look at it like if this V divider would become a gate at which the AC resistance would be high when looking towards it, and would be small when looking from the other side.

Again I might be completely wrong, as this is a guess.

[earthtonesaudio]

Yes, some drawings would probably help matters a lot...
Best example of what I'm doing:
http://web.telia.com/~u43200663/blocks/bootstrapping.htm
The first picture there, they have a 25uF cap going from the source to the voltage divider bias point.  Also attached to the bias point is the 2M2 gate resistor.  My circuit was essentially identical to this, except for the 10k source resistor, I used a 10k pot, and attached the 25uF cap to the wiper.  As the wiper approaches the top of the track (closest to the source) the highs get rolled off. 

Now, from a non-bootstrapped point of view, this makes perfect sense, because since I'm taking signal off the source, putting a cap to "ground" (via the voltage divider resistors' parallel value) obviously creates a high frequency roll-off.  But the literature on bootstrapping says nothing about how the very action of adding the bootstrap cap creates a low-pass filter. 

Here's what I think the problem is in my case: my voltage divider is not the same as the one in the picture.  I used two 10k resistors.  That would present an AC resistance of 5k, which is lower than the source resistor, and therefore the signal would rather take that path to ground.  That seems like a reasonable explanation to me... anyone agree/disagree?

Thanks for your input so far.  I think I'll get the hang of the whole bootstrapping concept eventually.

[Sir H C]

Yes that is the problem.  You want the voltage divider to look reasonably like an open relative to the source impedance.  10 to 1 is viewed as a good ratio for resistances.

[earthtonesaudio]

Cool, maybe I'll have another go at it then.