What is bootstrapped?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
AX tech editor
Joined 2002
Paid Member
Bootstrap

The name comes from the expression "pulling oneself up on one's bootstraps".
Generally it means that a circuit node pulls itself up by itself (huh? did I say that?).

An example: suppose you have a preamp with an input resistor of 1k to ground. That preamp has then an input resistance of 1K, that's what the source sees as load. Now suppose you disconnect the ground pin of that 1k resistor and connect it to a signal point that has the same level as the input level. The current through that 1k goes to zero (both sides have the same voltage), and the source now sees an effective input resistance of infinity. You have now bootstrapped the input resistor: it has pulled itself up by itself (the input signal).
There are many other cases, and I'm sure members on this forum will be too happy to point it out, but the concept is always the same.

There is also an equivalent case in software: when your computer "boots", it is loading what used to be called a "bootstrap loader", which is a small piece of software that can load itself without having a program to load it in the first place, hence the name bootstrap loader. After that, the bootstrap loader loads the operating system.

Jan Didden

"There's nothing so practical as a really good theory" - Ludwig von Boltzmann
 
AX tech editor
Joined 2002
Paid Member
Bootstrap

Yes I think you're right, I think it was the (in)famous Baron von Münchausen. I think he even did that while sitting on his horse, pulling up horse and all.
The same guy that jumped on a cannon ball fired from his own army to go over to the other side, do some reconnaissance, then came back by hitching a ride on a enemy cannon ball.

Jan Didden
 
Optical said:
I hear its a patented design of some sort and results in low noise?

is it a circuit of some sort?

It has nothing to to with low noise and I don't know anything about any patent.

A good example is here

Check the 100 µF cap. It serves like a battery if we talk AC-signals. This cap "lifts" up the potential in phase with the audio output.
 
Little Niles, big Frasier.

Imagine you are using a crane to raise a load by 10 metres. All you really care about is how far the load is above the ground. As you start to raise it, your friend who has a MUCH bigger crane than you, decides to lift up *your* crane with *his* by the same amount that the load went upward, i.e. 10 metres. When everything has stopped, his crane has done almost all the moving and yours has moved hardly none. What use is that? Your crane has to do almost zero lifting but it is in control of the whole process.The bootstrapping part is where the little crane lifts the load and therefore causes *itself* to be lifted by *ideally* the same distance. I say "causes" because it doesn't actually lift itself, something else does.

In an amplifier cct, the bottom end of a load resistor is allowed to go upward and this point couples to the output stage. The output rail then lifts up the supply rail to the top of this resistor by the same amount that the bottom of the resistor went up. The voltage *across* the resistor has not changed therefore the current through it hasn't changed, so the transistor hanging of the bottom of this lovely load resistor thinks it is connected to an *infinite* resistance (the current doesn't change even though the voltage at the bottom end of the resistor does). Therefore it is in transistor heaven and makes LOTS of gain and LOW distortion. :cool:

GP.
 
Thinking about it further, it's more like having a spring dangling from your ceiling and you have it pulled in a stationary position halfway down to the floor with a constant 10kg tension. As you pull it further down toward the floor, your friendly next-door-neighbour watches what you are doing and as you move first she co-operates by lowering your ceiling by *an equal amount*. Same goes for when you let the spring go upward.

The result is that you feel no change in spring tension because the length of the spring (theoretically) doesn't change. The good part is that if you now say to yourself "I shall *try* to apply an extra .0001kg to the spring even if my hand has to go beneath floor level" and the ceiling comes down on your head. Then you change your mind and *try* to reduce the spring tension by .0001kg and your ceiling goes all the way to the moon. You now have lots of gain. Next you invite your neighbour over and have tea and biscuits with her while congratulating each other on how clever you both are.
 
What's a Bootstrap?

I think of a bootstrap this way. It is a means of applying identical voltage changes to the top and the bottom of a resistor to increase the dynamic impedance of that resistor. Effectively, it turns the resistor into a constant current source (or sink).

Let's say we have a 10K bias resistor from the base to ground on a power amplifier. Ignoring the input impedance of the base, which is very high anyway, the source sees a 10K load with respect to ground. This is a stiff load. :xeye:

Let's disconnect the ground connection on our 10K resistor and hook it to a voltage divider coming off the output of the amp. (A separate resistive divider with the cold end grounded is often used in this situation.) Now, as the input goes up and down at the base, the other end of the resistor also goes up and down equally with it, since the bottom of the resistor is connected to a voltage divider off the output which mimics what is happening at the input.

The interesting thing is that the top and the bottom of the bias resistor are moving up and down in step. An approximate analog in the mechanical world is the Weston Differential Pulley - the endless chain pulley used to lift engines from autos. Thus there is no variation in current through the resistor, which actually defines a current source. Ordinarily, if the current does not vary with changing voltage input, we have a near infinite resistance, since by simple Ohms Law, R = deltaE/deltaI. This is boostrapping, and it's very effective.

The usual bootstrap application is voltage amplifiers in older SS PP designs, such as the Citation 12. Two resistors feed the collector of the NPN voltage amplifier, with a large electrolytic connected between the output of the amp and the midpoint of these two resistors. Thus, as the output of the amp hikes up and down, so too does the potential at the midpoint of the feed resistors. Thus the voltage across the lower of these resistors scarcely changes, leading to a constant current, and bingo, the virtual impedance of this lower resistor becomes very high.

And so the collector of the voltage amplifier sees almost an infinite value of resistance as load, and this in turn leads to the very high open loop gain so necessary for proper operation of a feedback amplifier.

What is not clear here is that the circuit which drives the bootstrap must have substantial power, so must be of a lower intrinsic impedance than the circuit being driven. Furthermore, the HF behaviour of the bootstrap electrolytic complicates things; as the frequency rises, so too does the ESR of the cap, reducing the bootstrap action.

Cheers,

Hugh

www.askaonline.com
 
So for example, on the esp website under project 30, the mixing desk, he mentions that the balanced input circuit would be better if it were 'bootstrapped'.. how would this circuit be modified?

would i be right in thinking from previous posts that at the input, where the 100k resistor goes to ground, could be connected to an attenuated output of the circuit?

btw thanks for all the advice and knowledge guys :)

An externally hosted image should be here but it was not working when we last tested it.
 
AX tech editor
Joined 2002
Paid Member
Optical said:
So for example, on the esp website under project 30, the mixing desk, he mentions that the balanced input circuit would be better if it were 'bootstrapped'.. how would this circuit be modified?

would i be right in thinking from previous posts that at the input, where the 100k resistor goes to ground, could be connected to an attenuated output of the circuit?

btw thanks for all the advice and knowledge guys :)

An externally hosted image should be here but it was not working when we last tested it.

Well, it would, except in this case there is a hitch (assuming the + and - input are equal in amplitude but opposite in phase). Because of the balanced inputs, the signal at the top of the 100k is already (drum roll) ... zero! So there is no point in making zero what already is zero. There is only DC through the 100k, to provide a return path for the input bias current.

Now, if you use this circuit single ended, with one of the inpuits grounded, then yes it would work as you say.

One thing to watch out for: if you put the bottom on a signal that is larger than the input, the circuit will start to generate its own input without external help. Such circuits are commonly known as "oscillators". (In fact, bootstrapping is a form of positive feedback. The gain around this loop must always be less than 1). The max bootstrap voltage would be exactly the input voltage, which would make the resistor appear infinite. In practice it will be very difficult to exactly define that point. But if you would use a bootstrap level of 0.9 input voltage, you are on the save side and the resistor will appear 10 times as large.

And, oh, make sure the bootstrap signal is in phase with the input signal. Otherwise - you guessed it - the resistor appears much lower than it really is. Can you tell me why?

Jan Didden
 
AX tech editor
Joined 2002
Paid Member
Bootstrap

Well, I assume that with "improvement" you mean the input impedance? The differential input impedance consists of R2 & R3, increasing them would help, but I haven't really analysed this circuit. Probably one could separate R2 & R3 and connect their cold ends to a suitably attenuated version at the output of the first set of opamps.

But, note that this discussion started about "what is bootstrapping". I don't want to imply that it is always a good thing or even recommended. But in some cases it can be a quick & dirty way to increase input impedance.

The case with the bootstrap capacitor mentioned above by Hugh Dean to linearize the collector load in a Vas stage was widely used, until someone introduced the current source loading of a Vas collector and that then became the trend.

Jan Didden
 
Nothing exceeds like excess.

janneman said:
The case with the bootstrap capacitor mentioned above by Hugh Dean to linearize the collector load in a Vas stage was widely used, until someone introduced the current source loading of a Vas collector and that then became the trend.

Jan Didden
Of course, you can go quite silly like me in my class A amp somewhere else on Diyaudio, and use a (voltage) bootstrapped current source. All part of the DIY ethos actually. ;)
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.