Different differential

[PRR]

> I am getting confused.

Me too.

This is conceptual; it is too hot to wrestle the PC to get a detailed schematic. Input drive and biasing same as my earlier plan. This has in-phase nonlinearity and no large feedback loops. Gain is easily 72dB, adjusted downward by R2, and R6 to get to very low gain. It is a pig for power: you might scale-down the Q5 current and add a buffer.

I have no idea what these transistors are: they popped-up in 5Spice and I took them. {later} ah, they would be good, except BC8x8 is 30V, try BC8x7 for a 50V part. I only found surface-mount; Q5 dissipates over 2 Watts so you need sumthing bigger there.

 

[tk@halmi]

Thank You for the clarification. This is what I came up with last night. It is a version that uses the additional gain stage and buffer.

Diff-Diff with second stage and p-p buffer

The common emitter amp gives an additional 6dB of boost with no bypass cap and utilizes a lot of local feedback. The output buffer is push-pull.

I think I am going to build this version.

Cheers,
Tamas

 

[tk@halmi]

Ohtay, so I built the first part of the circuit, just the four transistors, no additional stage or buffer. It even amplifies. :grin:
It really needs those capacitors between the emitter and collector of the input transistors or you get a heck of a bandwidth (oscillations too).

I need to work on a better matched balanced signal to feed it. Right now I run into the primary of a 600:600 transformer from the signal generator and feed its secondary into the Diff-Diff preamp.

Some issues that need to be dealt with:
- It does some really ugly things when the input voltage goes near a half of a volt.

More later.

Cheers,
Tamas

 

[tk@halmi]

I am not sure I understand the distortion characteristics of this thing. When overdriving the sine wave grows a "finger". Where have I seen that? It looks what the old diode base "noise supressors" produced.
Any thoughts?

 

[tk@halmi]

This circuit is shaping up nicely. Some odd artifacts were due to the switching power supply I was using. but now I have a 22 ohm resistor between two large capacitors on the 48V rail, and things look a lot smoother.

The basic circuit consumes about 16ma, requires about 20 seconds to warm up and reach equilibrium.

The step response is nice and the slew rate is not too fast. Some peakiness from the transformer I use for balancing the signal is actually smoothened out.

Right now using 300pF between the emitter and collector of the input transistors seems to allow for stable operation.

I have yet to test it at large gains like 40dB and over, and at very low gains like 10dB.
All my test were done with a fixed resistor that gave about 20dB.

I think this circuit deserves more experimentation and some thoughtful signal conditioning on the input. I will try to put it in a metal enclosure and give it a sound check, but now...

I have to go fill up with beer and potato salad until I can't move. Later in the evening I will blow my fingers off with fireworks. Happy Fourth of July!!!

Tamas

 

[Rossi]

Bump. (just so you know you're not talkling to yourself :grin: )

 

[tk@halmi]

It is on hold for now. I got frustrated with oscillations and strange distortion problems at varying gains. My measurement methods and test bench hygene are always suspects. I am thinking to add a follower to isolate it from downstream crud on the breadboard.
In the mean time, I have found out that a nullator and a norator is a nullor. How about that?

 

[tk@halmi]

I am done with this. On the simulator it seems to make decent signal. However, on the breadboard it creates some ugly stuff. Perhaps someone else can take a stab at it and prove me wrong.

 

[tk@halmi]

Since I have no time for anything new these days I dug out the breadboard for a version of this circuit and fired it up. (good thing I am a total packrat) It really deserves more study. I will play with it in the evenings. I really would like to get it to the point when measurements and test recording can be attempted. Right now it seems to have a severe bandwidth problem, but it could be a faulty measurment approach guided by beer fumes.

Cheers,
Tamas

 

[chrissugar]

Hey Tamas,

Looking forward for your results. And how it sounds.

I'm curious how it would look this differential-differential if it would be powered differentially (+/-24V) and how it would be the output buffer to avoid capacitor coupling.

chrissugar

 

[tk@halmi]

[quote author="chrissugar"]I'm curious how it would look this differential-differential if it would be powered differentially (+/-24V) and how it would be the output buffer to avoid capacitor coupling.
chrissugar[/quote]

Wasn't the attraction of the circuit that it runs on the same voltage as the phantom power? So one could build a simple single rail power supply for the preamp. You will never get low offset output on this puppy anyhow.
Operating on +48V is the perfect circuit for capacitor coupling as the cap will likely have enough DC bias on it to be well formed and sound at its best.

 

[tk@halmi]

The input part is shaping up. I have found a wiring probem on the breadboard that threw me off. I have pics of the step response at different frequencies and gains, and the circuit itself.
My Verizon website is full and 90% of the files are linked to this forum so they will disappear if I clean up. I requested an account http://groupdiy.twin-x.com/index.php. I will post when the account is approved. In the mean time I am working on a simple single ended output.

 

[tk@halmi]

Looking at the gain, using the lowly 2N4401 and 2N4403, transistors the maximum gain I get without 'straining' the input stage is 40dB. After that the step response starts to get too much garble. So I have decided to use a simplified version of a bloak for the output. This will allow the builder to customize the gain structure easily as needed. It should drive 600 ohm inputs without blinking once.
Again, the idea PRR proposed is to keep the circuit single ended as long in the recording chain as possible to preserve 2nd and other even harmonics. For the initial development I don't want to introduce slew rate smear either.
The input diff amp gain can be set with a rotary switch and the volume can be controlled with a regular audio pot between gain stages. This alleviates the need for pain in the butt to revlog parts. The circuit consists of nine pedestrian, non-esoteric, BJT transistors.

Drawings and pics to follow this long wekend. Four days off!!!

 

[Rossi]

Go, Tamas, Go! :thumb: :thumb: :thumb: :thumb:

PRR's different differential plus your own bloak seems like an excellent project. Makes me wanna try this myself sometime soon!

 

[PRR]

> 40dB. After that the step response starts to get too much garble.

Can you hear this "garble"?

Did you play with C7 C8, compensation caps? There is no perfect compensation for this amp, but those "470pFd" values will depend on devices and layout.

Speaking of layout: ANY inductance in the R2 C1 path is going to be trouble, unless it is holistically designed into the compensation. I think this was a concern: how to keep the path from Q1 emitter, through physically large cap, through pot/switch, to Q2 emitter, very-very low inductance. Layout may be tough (may be impractical).

 

[bovox]

Hey, anybody done this? PRR,s version or TK@Halmi´s mod.
Cheers Bo

 

[adamasd]

I played about with PRRs version long ago.

It worked pretty much as advertised. I either had to do no or very little tweaking, can not remember how much of that tweaking was just my curiosity vs how much I had to do.

I liked its sound and ment to continue on with it, but I borrowed its power transformer for something else and it never made it back. Forgot all about it, should dig that out again.

 

[clintrubber]

[quote author="PRR"]this is a differential amplifier:

I found this in a 1967 book.[/quote]

FWIW, it might have inspired the circuit of this 1970/71-patent for a negative capacitor:

http://www.pat2pdf.org/patents/pat3578911.pdf

Ran into that one by accident and it reminded me of the pic above.

Regards,

Peter

 

[tk@halmi]

I am going to revive this project. I think I have found a good match for the output stage.

Just as soon as I finish the tubification of an ST51 mic... :grin:

 

[clintrubber]

:thumb: