Another Discrete Amp - GainBloak

[tk@halmi]

Hi,

There is a JFET input version of the GainBloak now that should suit users with lower output DC offset needs. The 2SK170 is an excellent part for the job, but other N-Channel FETs should work with adequate ratings and suitable pinout.

Tamas

PS: Boards for this opamp are going to be up for sale in a week. Waiting for a big bag of Mill-Max pins. Each board will come with its own set of pins so you won't have hunt for them.

 

[featherpillow]

Hey, that's neat! I just downloaded the PDF for this yesterday, and was going over it along with your B7 opamp, the National Semiconductor opamp notes, and the Douglas Self "Distortion in Power Amplifiers" recommended in the Meta.

I downloaded the whole Self page, converted it to a Word document so the page breaks were evened out, increased the font size, and then printed all 45 pages of it.

These docs have been my morning reading on the train ride to work, and have been quite instructional.

I'm very surprised that Mr. Self doesn't seem to like FETs very well. I can understand not using them on the output, but he seems a bit biassed against them altogether. No pun intended.

 

[bcarso]

[quote author="featherpillow"]Hey, that's neat! I just downloaded the PDF for this yesterday, and was going over it along with your B7 opamp, the National Semiconductor opamp notes, and the Douglas Self "Distortion in Power Amplifiers" recommended in the Meta.

I downloaded the whole Self page, converted it to a Word document so the page breaks were evened out, increased the font size, and then printed all 45 pages of it.

These docs have been my moring reading on the train ride to work, and have been quite instructional.

I'm very surprised that Mr. Self doesn't seem to like FETs very well. I can understand not using them on the output, but he seems a bit biassed against them altogether. No pun intended.[/quote]

My impression is that Doug is usually in self-enhancement mode :razz:

 

[Consul]

[quote author="tk@halmi"]PS: Boards for this opamp are going to be up for sale in a week. Waiting for a big bag of Mill-Max pins. Each board will come with its own set of pins so you won't have hunt for them.[/quote]

You rock! I'll definitely be in for a few...

 

[Carl_Huff]

Tamas,

Please point me to your variation with the JFET front end and I'll put one together this weekend to play with.

 

[adrianh]

Can we change Doug to depletion mode.
self-enhancement mode is so "Self" centered.

 

[bcarso]

[quote author="adrianh"]Can we change Doug to depletion mode.
self-enhancement mode is so "Self" centered.[/quote]

It's up to him, but at this point in his life it's probably hard to get a tiger to change his band structure :grin:

 

[tk@halmi]

D. Self also thinks that CFB amplifers suck. He is just missing out on all the fun. :green:
His work concentrates on power amplifers so I can understand why JFETs have very limited use there, perhaps on the input.

 

[tk@halmi]

Here is the drawing for the FetBloak. It is really just a matter of replacing Q1 with the FET and adjusting R4.

http://mysite.verizon.net/res75okq/sitebuildercontent/sitebuilderfiles/fetbloak.pdf

Notice that the value of R4 has changed to make clipping more symmetrical. If you want it to clip the same top and bottom you have to play with R4. The value will be somewhere between 120 and 150 ohms. I really don't worry about perfect clipping because I run them on +-22V and the following stage always overloads before the opamp ever clips.

I am working on a fully complementary version of the opamp. It actually looks like and works like the Borbely phono and line amplifiers. Most of his goodies are CFB amps as well.

 

[featherpillow]

I can understand why JFETs have very limited use there

I think he thinks they aren't linear enough; like you have to go to hell and back to linearize them so it's not worth it. I'm not really in a position to argue with him since I don't know as much, but I still disagree.

Look at the work of Borbely--he's ALL FET pretty much--thinks they're better than tubes even. I would consider the two of them to be among those at the forefront of the field, so I'd have to chalk it up as a difference of opinion.

 

[featherpillow]

I have a question--when subbing the SK170 for the original BJT input, are we going to see much of an effect on the slew of that stage?

The SK170 has a good sized amount of input capacitance, around 20pf IIRC...or do you think the lower transconductance of the FET compensates?

 

[tk@halmi]

[quote author="featherpillow"]I have a question--when subbing the SK170 for the original BJT input, are we going to see much of an effect on the slew of that stage?[/quote]

Well, it may be 100uV/S instead of 200uV/s.


[quote author="featherpillow"]The SK170 has a good sized amount of input capacitance, around 20pf IIRC...or do you think the lower transconductance of the FET compensates?[/quote]

That is an excellent question! I will measure the roll off. Of course the higher the impedance of the previous stage the more roll off one gets.

 

[featherpillow]

Looks like Ciss is 30pf, and Crss is 6pf.

 

[tk@halmi]

Using the 2SK170BL as the input device the roll off does not seem to be an issue at all. I put a 20K ohm resistor in series with the function generator. Then I measured the amplitude of the output signal and it is exactly the same at 1KHz, 10KHz, and 100KHz. Repeated measurements at 10dB and 20dB gains. I think theory says that the input is really like a buffer in CFB amps.

I tried to compare the slew rates to the BC550 input version, but I do not see anything different. The BK scope I got isn't really helping...

Cheers,
Tamas

 

[bcarso]

The input capacitance is mostly bootstrapped out with the series feedback. You have then the drain-gate C and its Miller multiplier, but that is also loop gain/phase dependent---at mid frequencies it doesn't take much voltage swing at the drain to achieve equilibrium, due to all the gain that follows.

EDIT: Even open loop there's not much Miller, with R14 500 ohms and the drain load about 1k.

 

[Svart]

I've been looking for the Blackmarket thread for the Gainbloak and FETbloak but cannot find them.. does anyone have the link?

also I am interested in using these in a channel strip, I don't recall any discussion on use as a buffer or any low gain setups.. any info on that?

 

[Flatpicker]

Fabio used to have boards - check with him.

 

[tk@halmi]

Mail me to gfargone at yahoo dot com if you want boards for these.
You could try them as buffers with a 10k resistor connected from output to negative input.
I use them with MJE700/800 darlingtons in place of BD139/140 for a VERY low distortion-low impedance output opamp.

 

[Svart]

awesome. email sent.

I'm gonna give these a royal flogging.. :green:

 

[adrianh]

Svart
You will like them!
Build and you will be happy!