Signal Compression.......

[Jack_Hester]

Take a look at the schematic of this vintage amp: 

http://www.el34world.com/charts/Schematics/files/gibson/Gibson_GA-100_BASS_AMP.pdf

1 - Can someone explain how this Compressor circuit works? 

2 - Why do you suppose that the PI uses pentodes?

3 - And, what is the advantage to having a voltage regulator tube vs. not having one (as is typical in most amps). 

Thanks for any info.  Have a good one.

Jack

[sluckey]

1 - Can someone explain how this Compressor circuit works? 

Hurts my head to think about it. 

2 - Why do you suppose that the PI uses pentodes?

The PI is V1A, a triode.

3 - And, what is the advantage to having a voltage regulator tube vs. not having one (as is typical in most amps).

Insures the screen voltage (and downstream voltages) will be safe.

[Jack_Hester]

The PI is V1A, a triode.

V1a? Or V1b? 

So, V2 and V3 are just more gain, between the PI and the power amp.  Any reason to do it there, rather than upstream of the PI? 

A friend of mine told me that a compressor in a Bass amp makes not-so-precise playing sound better.  Is that true?  Does it serve another purpose?

Thanks for the quick response.

Jack

[sluckey]

V1a? Or V1b?

Oops! V1B is the PI. Gibson was a light weight in the amp department. They tried a lot of non conventional stuff in their effort to compete. And they lost.

[eleventeen]

I'll certainly accept PRR's take on circuit description, but that resembles the classic Gates style compressor.

[Jack_Hester]

that resembles the classic Gates style compressor.

Can you point me to a link of this style?  I want to read more about such.  Thanks. 

Jack

[eleventeen]

At the risk of grossly overstating my grokking of circuitry....

http://headfonz.rutgers.edu/comp5/Comp5.html

(actually may have been posted by PRR circa 2003!)

This (the above, not the below, though the below is a similar schema) appears to be a differential amplifier (as more frequently used in instrumentation and test gear) with the input signal split, then applied equally (if out of phase) to two symmetrical triode grids. The grid feeds come from opposite ends of an input transformer, thus they are out of phase. The outputs are fed to opposite sides (also out of phase, then) of a transformer, so they get back together in phase, while the cathodes are basically tied together. In "the above" there is a balancing pot that presumably affects the level at which the thing starts to act by balancing out the gains of the mirrored triode-stage. In the Gibson, there is no pot, there is only an on-off switch that turns the compressor (and...it may only be a limiter, not an AGC-type compressor...I don't know) at some Gibson-selected level.

The point being, crudely stated, to generate an "error" signal and to feed the error signal into a gain stage such that any over-level signals act to reduce later-stage gain.

[Jack_Hester]

I fixed lunch for the guys in the shop, on Thursday (meat lovers baked beans).  A friend came up from Raleigh to eat with us, and we were discussing this a bit.  He has a recording studio and is a musician.  Both being life-long passions.  

I didn't have the Gibson circuit handy, but tried to describe it from memory.  I also tried to give as much of your reply (Eleventeen) as I could remember.  He repeated much of what you said, and his take on the Gibson circuit.  So, I went home and finished re-drawing the compressor circuit, with components that are off-the-shelf, as opposed to finding old tubes.  That's when it became obvious to me what you and he were saying.  

The compressor takes a signal from each side of the PI, each signal feeding the grid of a separate triode.  The resulting amplified (AC) signals are fed into opposite sides of a bridge rectifier.  The resulting DC is used as a continuously changing Bias for the pentodes downstream of the PI.  And, I'm thinking that the pentodes (V2/V3 of the Gibson circuit) may have been placed there as part of the compressor circuit, and not as some additional gain stages.  

Please comment of my simplification of the circuit, and correct my wording, if it can be stated better.  Now, to go back and look at the compressor switch circuit a bit.  Thanks for the comments and much help.

Jack

[eleventeen]

I'll have to look at greater length as to your redrawing the Gibson ckt.

But this is the basic notion of "regulation"...whether it is a power supply, an AGC-type compressor (maintains constant level, eg; brings up weak signals and attentuates strong signals), a limiter (only works on the upside to prevent too-strong signals), the governor on a steam engine, or some chemical reactions.

We sample the output
We generate an "error" signal (could be over/under voltage, excess/under speed, or too much heat/temperature)
We feed the error back to a gain or gate or process control stage to produce the effect we're interested in.

For electronics, since we can build whatever we can think of and implement (within the limits of whatever parts we can get) we probably wish to control how fast the ckt responds to level changes and corrects them. For example, in the case of an AGC compressor, we probably don't want the thing to snap to maximum gain as soon as it senses a quiet (no signal) condition so that the first microseconds of the next note blast through, then the ckt senses a too-high signal and crushes the overall gain...because you'd hear those sudden, exaggerated (over)corrections in the output. Thus ADSR "attack-delay-sustain-release" controls for the most sophisticated "envelope" circuits or more commonly for a compressor, attack and release.

There are a few different ways to create the error signal. The differential amplifier is a pretty sophisticated way, IMO. One could use a bridge circuit as used in early test/measuring devices...those can be VERY sensitive..or, a diode bridge. In the case of a power supply, we are only talking about volts, and DC volts, so a direct sampling of the output does the deed.

[Jack_Hester]

Our discussion at work was also around the use of tube diodes versus solid state.  Being that tubes would be a bit delayed in correcting, and solid state would be quick.  I would have to hear them both to know for sure the effect. 

I'll document as I go (the research, that is), so that I can have it at hand, if it appears practical enough to build as an experiment.  Thanks for your input, and please verify the little bit that I've redrawn.  Have a good one. 

Jack

[Jack_Hester]

Amazon sent this as a book of which I might have an interest:

http://www.amazon.com/gp/product/B00E7HN89M/ref=oh_aui_d_detailpage_o00_?ie=UTF8&psc=1

$2.99 for Kindle book.  I was interested.  Short book that I can read several times in a short period (days, weeks).  And, hopefully get a better understanding of this topic.  Not a difficult project, but busy.  It would have to be something for the Winter, when I'm house-bound.  Had some of those days last Winter, which is highly out of the norm for this part of N.Carolina.

Anyway, it anyone has a Kindle book, or the reader for PC (free), 3 bucks is a deal. 

Jack

[terminalgs]

1 - Can someone explain how this Compressor circuit works?

gain of a triode is altered by applying a control voltage (CV) to the grid.  CV can be generated by taking the input VAC signal and rectifying it into DC.  a small reservoir cap (that .05)  is used to "store" that CV.  series resistors,  and resistors to grid in the CV path can be used to alter 'release' and 'attack', etc...  usually, a "make up" gain stage follows the CV governed gain stage.


The input signal (an AC sine wave) when rectified to produce a VDC relative to ground, only produces CV 50% of time (if the sine  wave is 50% above 0v and 50% below 0v).  If you take the sine wave and invert it, and produce CV from both the normal sine wave and the inverted sine wave, you get a consistent CV that includes both the upper and lower halves of the input signal.

CV voltage can make the device a limiter or a compressor, or  even a gate.

3 - And, what is the advantage to having a voltage regulator tube vs. not having one (as is typical in most amps).

Lots of old tube compressors used a voltage regulator like an OB2  or an OC2.  I always figured it was to maintain an exact plate voltage, so the exact rise and fall of the grid bias consistently effected compressor/limiter effect. (which the engineer controls by adjusting presisely marked control knobs and monitors by visually watching a VU meter the front).


diode tubes are cheap..  6AL5's or 12AL5's are as cheap as they get. 6FM8's are cheap too.. There are lots of triode + diode or triode +  twin diode tubes out there.   You are using the triode as "side chain" , so you don't care about the fidelity of the  triode.  You just want to isolate, and maybe amplify the input signal, so you can produce some CV.     

As for the GA100.  I don't think this was a popular bass amp, and the compressor feature, which is cool for us to look at, maybe wasn't as useful to bass players as they had hoped it  would be.   Sort of like the string snubbers that used to come on basses (which players usually took off and threw away).

tube compressor circuits are fun to play with, but for guitar or bass SS stomp boxes are 10X cheaper, 10X smaller, and 100X more versatile and audibly pleasing.   Making a tube compressor for the studio is a fun project, but you'll quickly find the input/output/interstage transformers make such a project incredibly expensive...

[John]

Very informative post, thank you!

[Jack_Hester]

Well, I had a nice reply and submitted it.  It's gone, and going back gave me a blank screen. 

So, in a nutshell, the book is a great read, and I'm on my second time around (only 52 pages). 

Jack

[kagliostro]

I don't know if this thing can be of interest


I found this schematic attached to a Bass Amp project of a guy from East Europe


The only info I salvaged with the schematic is that inverting the diode (D) it will be an expansor instead of a compressor


Ciao


Franco

[Jack_Hester]

Franco -
 
I'm glad that you posted that drawing.  The author of the book stated that pentodes were used a lot in AGC, but he chose all triodes (in particular, 12AX7's) because of the desirable 2nd harmonic created by them, that gave the fuller sound to the end result.  As in, the much desired Hi-Fi quality of the old mono systems that he wanted to replicate, from the 50's and 60's.  But, he used it for stereo FM broadcast and built dual AGC. 
 
Jack

[kagliostro]

Glad it is of your interest


Franco