bootstapping comparison

[tubenit]

I was reading this thread and I am curious about DummyLoad's (Pete's) bootstrapping schematic?   The schematic I am posting shows a difference between Merlin's and Pete's idea.

http://el34world.com/Forum/index.php?topic=17781.0

I tried Merlin's once and it was really too bright for my taste; however, in the D'Mars circuit with the OD having a tone stack ....... it might work just fine.  Merlin's bootstrapping definitely added quite a bit of gain in what I was hearing.

The thing about Pete's idea is that the amount of bootstrapping gain might be able to be controlled by different sized resistors (470k or smaller) or by using a 500ka pot.

Anyone ever made a comparison of these two different approaches and can offer a idea of the difference tonally?

I love the D'Mars OD, but if I could boost it up maybe 25-50% more ............ I might be interested in that.

Thoughts on the different approaches?

with respect, Tubenit

[jazbo8]

I'm not sure you would call Pete's circuit a boostrap, may be I got my definition wrong?

[sluckey]

The schematic I am posting shows a difference between Merlin's and Pete's idea...

...The thing about Pete's idea is that the amount of bootstrapping gain might be able to be controlled by different sized resistors (470k or smaller) or by using a 500ka pot.

I don't recall ever seeing that Merlin circuit. But Pete's bootstrapping circuit deals with how the cathode follower is biased. Since it is a cathode follower, the gain is less than one. You can't get any more gain by doing anything to a cathode follower.

Cathode followers or cathodyne phase splitters (inverters) are very similar in operation. They can be dc coupled to the previous stage (like the Marshall CF tonestack driver). This type requires no additional resistors for biasing the CF. Or they can be cap coupled from the previous stage (Pete's example, or 5E3 cathodyne PI). When you block the dc plate voltage from the previous stage with a coupling cap, you will have to set up the bias for the CF separately. You can do this with a bootstrap resistor (Pete's example) or you can use a fixed bias voltage divider between B+ and ground to put a fixed positive voltage on the grid of the CF (the Paul C. mod is an example of this).

Bottom line though is that you cannot increase the gain of any style cathode follower. Gain will always be less than unity. Any additional gain boost will have to come from the gain stage that precedes the CF (or in a following gain stage).

[tubenit]

Sluckey,

Thanks for the reply. I am not sure if you are saying Merlin's bootstrap technique for more gain does or doesn't increase gain?  I tried it on a previous amp and there seem to be a substantial increase in gain, but it was very bright in tone. 

http://www.valvewizard.co.uk/dccf.html

I am understanding that Pete's design would not increase gain?

With respect, Tubenit

[SILVERGUN]

From memory, without pulling the book out,,,I think Merlin was suggesting that the bootstrapping increases the gain of the preceeding triode....

[SILVERGUN]

There's some good stuff from HBP in this thread:
http://el34world.com/Forum/index.php?topic=16458.msg161812;topicseen#msg161812

Here's the link to the Merlin page:
http://www.valvewizard.co.uk/dccf.html

[sluckey]

Thanks for the reply. I am not sure if you are saying Merlin's bootstrap technique for more gain does or doesn't increase gain?  I tried it on a previous amp and there seem to be a substantial increase in gain, but it was very bright in tone. 

http://www.valvewizard.co.uk/dccf.html

I am understanding that Pete's design would not increase gain?

I never made any comments about Merlin's circuit other than to say I've never seen that circuit. After reading the link I understand his circuit and see that the gain of the preceding stage will be boosted/increased. That is a form of 'bootstrapping' that I was not familiar with. But this is not the same idea as bootstrapped biasing as seen in Pete's circuit. All my comments about a cathode follower are still true. The gain of any cathode follower (not talking about any previous stages) will always be less than unity.

Merlin and Pete were both showing totally different examples of bootstrapping.

[2deaf]

Anyone ever made a comparison of these two different approaches and can offer a idea of the difference tonally?

The classic Marshall-style CF (Merlin's w/o frequency cap.'s) is extremely Lo-Fi and adds a significant sound to the circuit.  This is desirable in a lot of guitar amps.  Pete's CF is much more Hi-Fi and shouldn't add any sound to the circuit.

[SILVERGUN]

I love the D'Mars OD, but if I could boost it up maybe 25-50% more ............ I might be interested in that.

Thoughts on the different approaches?

I remember trying to increase the OD in my version of the TBM, and one of the things I did was increase the value of the OD drive pot to 500K....pretty obvious tweak and works as expected, adding more of a 'bassy' drive.
I guess another simple way to get there would be to a (switchable?) resistor (or pot, for adjustability) in series with the grounded tab of that pot.

If you did that PLUS bootstrapped the CF in that OD section,,,,that would definitely accomplish wayyy more drive.

[tubenit]

I never made any comments about Merlin's circuit other than to say I've never seen that circuit. After reading the link I understand his circuit and see that the gain of the preceding stage   will be boosted/increased. That is a form of 'bootstrapping' that I was not familiar with. But this is not the same idea as bootstrapped biasing as seen in Pete's circuit. All my comments about a cathode follower are still true. The gain of any cathode follower (not talking about any previous stages) will always be less than unity.

Thanks!  I always value your perspective and knowledge. I appreciate the clarity.

My respect, Tubenit

[jazbo8]

Merlin and Pete were both showing totally different examples of bootstrapping.

What you said made perfect sense wrt the CF, Pete's circuit deals with biasing but does not really alter the input impedance of the CF per se. Since Tubenit's original question was on gain, to some of us at least, it seemed to fall outside the commonly known definition of boostrapping, which can be found here.

[sluckey]

Pete's circuit deals with biasing but does not really alter the input impedance of the CF per se.

Yes it does increase the input impedance. You have essentially the same voltage at the top side of the resistor as you have at the bottom side (bottom side is more like .9). This results in much less signal current flowing thru the grid resistor and that makes the grid resistor look much bigger than it really is. Merlin used that same idea to make the plate resistor appear much larger than it really is and that's how he boosted the gain of the previous stage.

You can see evidence of this increased impedance by measuring the dc voltage on the cathode and then the grid of Pete's AC coupled CF. The grid voltage will be unrealistically low compared to the cathode voltage. In fact it will measure so much lower that you would swear the tube would have to be in cutoff! But it's not. The lower grid voltage reading is a result of the parallel input resistance of your meter loading the boosted impedance of that bootstrapped grid. Even a 10M DMM will give a bogus reading.

Look at the voltage chart numbers for V2B on the schematic of my 5E3 which has an AC coupled, bootstrap biased cathodyne PI (same as Pete's CF except it has a plate load resistor too).  The grid measured 15v and the cathode measured 41v. The tube could not possibly be conducting with those numbers. But notice on the schematic the 40v (in red) that I measured at the bottom of the grid resistor. That's the real voltage that appears at the grid too. We just can't measure it correctly with a 10meg DMM.

Here's my 5E3 schematic with voltages.... http://home.comcast.net/~seluckey/amps/5e3/5e3.pdf

The typical LTP PI grids have boosted grid impedance also. Ever wonder how Fender could use a 500pF coupling cap in a AA864 bassman? You cannot measure the grid voltage in the LTP circuit either.

[jazbo8]

Yes it does increase the input impedance.

I guess I was more focused on the intent of the circuit which was to bias the CF so it operates within spec, what you wrote made sense, but I was not able to verify the results when I sim it.  As you can see, with the boostrap, the input impedance becomes lower instead of higher (R4, the boostrap resistor, it is open and R3 shorted for plain old CF operation):

[sluckey]

I'm not always good at expressing my thoughts with a keyboard. I'm a little better with pen and a napkin. Morgan Jones explains it much better...

[HotBluePlates]

From memory, without pulling the book out,,,I think Merlin was suggesting that the bootstrapping increases the gain of the preceeding triode....

What Silvergun just said is exactly correct.

Pete's circuit deals with biasing but does not really alter the input impedance of the CF per se.

Yes it does increase the input impedance. You have essentially the same voltage at the top side of the resistor as you have at the bottom side (bottom side is more like .9). This results in much less signal current flowing thru the grid resistor and that makes the grid resistor look much bigger than it really is. Merlin used that same idea to make the plate resistor appear much larger than it really is and that's how he boosted the gain of the previous stage.

And all of what Sluckey just said.

... I was not able to verify the results when I sim it   As you can see, with the boostrap, the input impedance becomes lower instead of higher ...

The sim is lying to you. A "normal, non-bootstrapped" stage has what, a 1MΩ resistor from grid to ground? Then how does the graph show input impedance up to 100MΩ? Ahhh... It's only accounting for the input capacitance of triode, not for the grid resistor that's normally there.

And if I read the scale correctly, the 470kΩ appears like a ~10MΩ resistor... That's bootstrapping, since it should look like no more than the 570kΩ from grid to ground (at d.c., or 520kΩ at a.c.).

[jazbo8]

I'm not always good at expressing my thoughts with a keyboard. I'm a little better with pen and a napkin. Morgan Jones explains it much better...

You already gave a very good explanation, I was just trying to figure out why the sim results did not match up with the expected results... Now it's sorted, see below.

The sim is lying to you. A "normal, non-bootstrapped" stage has what, a 1MΩ resistor from grid to ground?

Nope, it's my own fault... it's a case of "garbage in garbage out", I was not comparing apples to apples, I mixed up the AC and DC coupled CF cases, duh! I am removing the previous post/chart to avoid causing confusion to the others.