Gibson GA-83S Stereo Vibrato

[Willabe]

Can someone explain how this vibrato circuit works? I only get little pieces of it to make sense in my mind.

I see V4a/b set up as the phase shifter but I don't see a variable resistance like a varisistor in a Maggi or an SR coil in a Hammond/Sluckey Warbler. Then it goes to a driver/gain stage? Has a LFO tied in also.  

They made less than 200 of these in total so few have ever heard this amps vibrato. It would be a big pain to build this only to find out it sounds only OK and not great.

(DL, you wouldn't happen to be interested in breadboarding the vibrato from this amp would you?    

I'd bet you already have the tubes for it.)


                   Brad        



http://www.el34world.com/charts/Schematics/files/gibson/GA-83S_STEREO-VIB_(PWR_AMP).pdf

http://www.el34world.com/charts/Schematics/files/gibson/GA-83S_STEREO-VIB_(PREAMP).pdf

[Willabe]

Is the 6CG7 being used as the variable resistance some how?


               Brad     

[sluckey]

V12 is the Oscillator/phase splitter. V5 and V6 are the modulators (one for each channel). With that in mind, look at page 9 of my scrapbook for a theory of operation of a circuit that is functionally identical.

http://home.comcast.net/~seluckey/amps/misc/Amp_Scrapbook.pdf

[Willabe]

Thanks Sluckey, I see it better now. I'll have to put both schemos side by side as they are a little different. The Fender circuit is complicated but it's still not as complicated as the Gibson even in mono. More than 1 way to skin a cat, maybe Gibsons engineers out foxed them selves on this circuit? Overly complicated without much benefit in the effects circuits tone?  

But the million $ question is; Is this true pitch shift vibrato?

For a vibrato circuit to be true pitch shift it has to have a variable resistance part like a varisistor or SC coil?

I've read both sides of the argument on if the Fender vibrato is or isn't true pitch shift or just sounds like it is. I played through a brown face concert years ago but I don't know if it is/was true pitch shift or if it matters? It did have a good sound to it.

One thing I did notice was that once I played through the vibrato in that amp after I turned it off I still could hear the tone/pitch/voicing of it just without the modulation. I don't know if the amp needed a little work or if that circuit just acts like a filter of some kind on the overall voicing of the amp? It would make sense since your still going through part of the vibrato circuit all the time.


              Brad        

[Willabe]

Ok, I copied the 2 vibrato portions of the 2 amp schemos and will post at the bottom for anyone who's interested.

Looking at them side by side with the GA-83S in mono it's really just a few more parts. Not as bad as I 1'st thought. With both using 5 triodes to get done.

The order is slightly different between the 2.

Fender 6G _ _ type has > LFO/CF driver > PI > frequency splitter > modulator with guitar signal fed into the frequency splitter.

Gibson GA-83S has > cathode driven frequency splitter > modulator with the LFO/PI driver fed to the modulators grids.

So I guess that even though the order of the different stages is different the 2 vibratos accomplish the same thing effect wise?


              Brad      

    

[sluckey]

Take a look at this. Compare page 2 to page 3. I bet Vox copied the Gibson circuit.

[Willabe]

Thanks Sluckey for pointing that out and the 2 drawings. The highlighted/labeled Gibby drawing is very helpful to me. I was wondering which was the hi and which was the low on the freq. splitter.

They are the same. Vox is said to have copied the top boost circuit from Gibson so why not copy another circuit as well?

I have a few of questions;

1. Vox is using a (5 stage) hi-pass filter (?) on the vibratos output? Is that to smooth the vibratos waveform? If so then each stage smooths it a little more? Fender didn't feel/see the need to use it?

2. How does the Vox mixer work? Looks like a split load on the plates, at least where the output is taken from? And why the cap (C42) on 1 of the plates?

3. Why are the plates in the Gibby tied back to the hi-lo-pass frequency splitter through a series C and R? I see they are part of the freq. splitter but why/how does that work?

4. Are the Fender and Gibby hi-lo-pass filters anywhere near each other in Hz? I don't know the math to figure it out.  

      
             Brad    


Edit; I've been thinking about Q. 3.

Since the plate and cathode are 180 deg. out of phase it would cancel (full and going up) where they set 3db knee as allowed by the limiting R in series with the cap? And would roll off less as the frequency goes lower, again set by the rolloff knee created by the cap? But why did they do that?

To help set the amplitude at a certain frequency or is this just part of how they tuned the frequency splitter?  

[Willabe]

 

[sluckey]

Sorry. I had meant to get back to you on this, but just got busy closing out my career and forgot.

1. Vox is using a (5 stage) hi-pass filter (?) on the vibratos output? Is that to smooth the vibratos waveform? If so then each stage smooths it a little more? Fender didn't feel/see the need to use it?

Yes, yes, yes, yes, and I guess so. If the trem signal cancels perfectly in the mixer, why would you even need a filter? Maybe Leo felt it was good enough. Doug's revibe vibrato sounds pretty good and it doesn't use that 5 pole filter.

2. How does the Vox mixer work? Looks like a split load on the plates, at least where the output is taken from? And why the cap (C42) on 1 of the plates?

Maybe Vox just needed to change the circuit just enough to say it was original.

3. Why are the plates in the Gibby tied back to the hi-lo-pass frequency splitter through a series C and R? I see they are part of the freq. splitter but why/how does that work?

That's just the way they chose to do it. The guitar signal on the plate is the INPUT to the filters. The series C and R are part of the filters. The size of the series caps (we also call them coupling caps) is a big clue as to which is the high filter and which is the low filter.

4. Are the Fender and Gibby hi-lo-pass filters anywhere near each other in Hz? I don't know the math to figure it out.

Probably, but I really can't say. That's way beyond the level of understanding I need about this circuit.

I just know that I like the way that trem sounds. Listen to some of Doug's stuff where he is using his revibe. I know enough about the circuit to understand it on a block level and that would enable me to build and/or troubleshoot successfully.  Maybe some of the deep theory guys will take it on down to a component level.

[Willabe]

Sorry. I had meant to get back to you on this, but just got busy closing out my career and forgot.

That's what I thought.        So I waited a few days to give the topic a bump.

I just know that I like the way that trem sounds. Listen to some of Doug's stuff where he is using his revibe.

Yeah I've listened to most of Doug's sound clips and yes that Re-Vibe sounds real good to me.

The guitar signal on the plate is the INPUT to the filters. The series C and R are part of the filters. The size of the series caps (we also call them coupling caps) is a big clue as to which is the high filter and which is the low filter.

Ok so I have the cart before the horse? The signal comes from the plate and is tied back to the cathode and ground through the filters? When I saw you listed the stage as a "buffer" I thought the signal came from the cathode for low Z drive. But there's still got to be cancellation between the plate and cathode that is part of the filters tuning?

Maybe some of the deep theory guys will take it on down to a component level.

Yes I'd like to understand more about how this circuit works.

Thanks Sluckey.


                  Brad      

[PRR]

Fender had patent on a similar plan. Gibson would be careful not to copy too close.

V4 is wired as split-load cathodynes. The plate signal is equal to cathode signal but out of phase.

Taking signals from both sides allows greater phase shift. I think.

There may be cancellation.

I'm not going to analyze a 4-R 4-C network which is known to work.

You might learn more by poking a working example.

The 5-stage hi-pass is needed in a single-end modulator, but "not" with this push-pull affair. I suspect the VOX was some other plan hastily modified to this plan, and not fully re-simplified.

[Willabe]

Thank you PRR for shedding some more light on this.

You might learn more by poking a working example.

I'm going to have to build one and see how it sounds.


            Brad     

[Willabe]

Hey, look what I found. An early stand alone vibrato from Gibson. Some of the same circuitry in it.

             Brad     

[Willabe]

V4 is wired as split-load cathodynes. The plate signal is equal to cathode signal but out of phase.

Taking signals from both sides allows greater phase shift. I think.

There may be cancellation.

So is this circuit true pitch shift?

Can you run 2 or 3 stages in series? I think yes but you would have to include the split load triode, the modulator and at least the 2 mixer/combiner R's then go to the next split load grid. So 1 1/2 tubes for each phase shift stage.


               
               Brad       

[HotBluePlates]

I dunno for sure. Probably.

Best way to find out is to breadboard, the view output with an oscilloscope.

Despite what Zack proclaimed at Vibroworld, you could have vibrato and trem going at the same time, which amounts to signal level not being constant at the vibrato output.

Which reminds me of why Hewlett-Packard became a big company. They built an oscillator that had a constant-amplitude output, because competitor products tended to have changes of amplitude when they were adjusted to output a different frequency.

Anyway, the thought occurs to me that Zack might have seen the less-than-perfect designs have a volume variation while the frequency was varying. This would be the wave getting taller/shorter, when he would have looked for wider/narrower (indicating frequency change). It might have looked confusing depending on scope settings.

[Willabe]

Despite what Zack proclaimed at Vibroworld, you could have vibrato and trem going at the same time, which amounts to signal level not being constant at the vibrato output.

That might have a good sound of it's own?

Which reminds me of why Hewlett-Packard became a big company. They built an oscillator that had a constant-amplitude output...

Anyway, the thought occurs to me that Zack might have seen the less-than-perfect designs have a volume variation while the frequency was varying.

1. So how complicated is Hewlett-Packard's design?

2. What would have to be changed/added to make a more perfect LFO out of the Gibby/Vox design?

3. Could it be made adjustable to dial in between less than perfect (as is) to closer to perfect LFO?



                Brad          

[Willabe]

Can you run 2 or 3 stages in series? I think yes but you would have to include the split load triode, the modulator and at least the 2 mixer/combiner R's then go to the next split load grid. So 1 1/2 tubes for each phase shift stage.

And have I got this right?


             Brad     
    

[HotBluePlates]

1. So how complicated is Hewlett-Packard's design?

2. What would have to be changed/added to make a more perfect LFO out of the Gibby/Vox design?

It's not about whether the LFO is perfect in the guitar amp circuits, but how the phase shift stages are implemented (if my hunch is right).

How complicated is HP's circuit? Simple and not, at the same time. They did a standard R-C oscillator, then used NFB around the oscillator/amplifier circuit. The NFB and original LFO/amplifier form two arms of a bridge, with a small lightbulb being incorporated in part of the NFB bridge arm. the lightbulb provided local NFB that kept the amplitude nearly rock steady.

Anyway, what reminded me of this was the fact that the Magnatone patent arranges the circuit as a bridge, then has variable impedance elements. Signal input/output are two opposite corners of the bridge, LFO inputs are the other two opposite corners.

All this is probably a side-show. Build the thing, listen, probe with a scope if you're interested.

[Willabe]

Build the thing, listen, probe with a scope if you're interested.

Ok, thanks.

I think it's time I get some kind of bread board together. I have some ideas and some parts on hand that should be able to get it done.


             Brad     

[PRR]

> Hewlett-Packard ...built an oscillator that had a constant-amplitude output, because compatitor products tended to have changes of amplitude when they were adjusted to output a different frequency.

Trivia.........

I'm not sure this is all true.

I'm not sure what "compatitor products" existed.

A lot of work got done with fixed-frequency oscillators.

The main variable frequency oscillator was the BFO. Fix one oscillator at 200KC. Vary another 200.1KC-210KC. Beat them together, you get many products, one of them a 100CPS-10KC tone.

Since it works on small-differences, every part must be very frequency stable. However the amplitude stability was excellent because the frequency shift was small.

A good BFO was a HUGE machine and expensive. Not real comparable to the compact and not-so-costly 200AB. There might be one BFO for a whole studio. 200AB was used in much larger numbers.

> you could have vibrato and trem going at the same time

Sure can. Probably happens in some vibratos.

> How complicated is HP's circuit?

It looks like a wiring error. There's a pilot-lamp where you expect a cathode resistor.

But that's not the whole story. If you improve the oscillator's amplifier, it gets -less- stable. It worked for H and P because their amp was sloppy. It was almost 50 years later that someone worked-out the "why" of this.

[PRR]

...why Hewlett-Packard became a big company. They built an oscillator that had a constant-amplitude output, because competitor products tended to have changes of amplitude when they were adjusted to output a different frequency. ...

Here is how it was done in 1933 (well before H-P), from RCA Broadcast News April 1933.

I'm not sure a BFO's output changes with frequency much. It is a simple straight-ahead circuit with no large feedback loop (the Wien tends to bounce when changed quickly). BFOs are dastardly fussy to build and to keep calibration at low frequencies, tending to lock.

[PRR]

And here is what you do with a variable oscillator: calibrate microphones (etc).

Note this is NOT a "strip chart", and the pen is not automatic. Standard paper tacked to the stage. One motor drives a carriage carrying the pen and also cable-drives the BFO big knob. The mike output works a meter needle. The operator eyeballs the needle and moves a spot of light on a handle. The operator IS the pen-servo! Using arm-motion to move a pen too heavy/draggy for the meter to move.

[PRR]

A BFO from 1936:

[uki]

@PRR  Wow !! Sir I'm impressed how you tracked this all this years!!   

[octal]

To circle back to where we started, I actually own one of these amps. (Well, the two chassis- I bought it without the cabinet.)  It sounds nice- a bit Leslie-ish, but it doesn't have quite the extreme pitchbending effect of one of the higher end Magnatones. Years before I bought the GA83s, I made a solid state version of the vib/trem using op-amps and some Philips electronic volume chips. (their gain is controlled by a DC voltage, which I modulate)  I scaled the phase shift networks to adjust for low driving impedance of the op amps.