AC30 Vib/Trem

[G._Hoffman]

Can anyone point me to a good technical explanation of the vib/trem circuit for the AC30?  Preferably with pictures/oscilloscope plots, and maybe even a bit of math?  I'm thinking about building an AC30, and I want to understand the Vib/Trem circuit.  I know I could just blindly copy, but that is not nearly as much fun, and makes it ever so much harder to figure out what will inevitably go wrong when/if I build it!  Besides, I'd kind of like to maybe replace the oscillator with a solid state circuit to save some current for an extra tube so I can put in an EF86 normal channel.

I've done a quick Google search, but I've had a hard time finding anything in depth.  I'm thinking of just building one up on a bread board using FETs and seeing if I can figure it out that way, but that seems a good way to get very frustrated.

Oh, and a book would be fine with me, as long as it isn't TOO expensive.  I don't mind spending a bit of money, if the product is worth it.


Gabriel

[kagliostro]

I haven't the kind of info you are looking for

But may be you can find interesting info here

http://www.guitargear.net.au/discussion/index.php?topic=30731.msg326407#msg326407

if you read all the post on the link, you'll find some that are very interesting

K

[HotBluePlates]

Can anyone point me to a good technical explanation of the vib/trem circuit for the AC30?

I don't have pictures & plots for you. Yet.

I assume you have a copy of the schematic for a Vox amp with vibrato, but for those that don't here is the Vox AC-15. I will be referring to this for the explanation.

The signal from the Vibrato channel jacks goes to a section of V5 (12AX7) and gets amplified; this is dry guitar signal and is initial boosting of the guitar signal.

The next section of V5 is a self-biased split-load inverter with 22kΩ plate and cathode loads (R39 & R37). The output of this stage is 2 equal and opposite signals with phase 180 degrees apart.

• At this point you should know that if you apply 2 signals of identical phase to 2 different circuits, one made of pure resistance and the other made of resistance and capacitance, the two signals will no longer have identical phase. That is, reactance in the circuit appears to advance or delay the voltage output compared to the pure resistive circuit. Therefore, you can have 2 identical signals which differ only in their phase relative to one another.

The plate output of V5's split-load is applied to the series RC network made up of C27, C28, R49 and R50 (if you look carefully, this is 2 separate series RC paths). The cathode output is applied to the parallel RC network made up of C24, C26, R46 & R48 (if you look carefully, this is 2 separate parallel RC paths).

• These 4 total paths form a bridge circuit. The 4 arms of the bridge are (starting at the plate) C27/R49 (in series), then C24/R46 (in parallel), then C26/R48 (in parallel), then C28/R50 (in series).
• The 2 parallel RC circuits of C23/R45 and C25/R47 are ground references for the two outputs of the bridge.
• The point of the bridge (and its reactive components and their arrangement) is to get 2 output signals whose relative phase is 90-degrees apart (rather than the 180-degrees apart at the 2 input points of the bridge)

Now forget the 90-degree signals (also termed "two signals in 'quadrature'" as you will see in very technical references). Look instead at V7 (12AX7). The left-most triode is your standard phase-shift oscillator, as you'd find in any tremolo circuit.

The output of the phase-shift oscillator is applied to the other section of V7, which is another self-biased split-load inverter. Its outputs are 2 copies of the trem oscillator signal, but 180-degrees apart.

[HotBluePlates]

V6 (12AU7) is a differential amplifier stage, whose sections have individual plate loads (R34, R35 at 10kΩ each) but also shared plate load resistors R32 (47kΩ) and R20 (12kΩ). This mixes the 2 outputs and biases each triode properly, though C42 bypasses some of the plate load for the a.c. output. The output signal is passed to the multi-stage steep low-cut filter (C13, R12, C14, R13, C15, R14, C16, R15) in the manner of a tapped plate load resistor.

• Now back to the inputs of V6: the 2 outputs of the bridge are applied via caps C21 and C22 to opposite inputs of the differential amp.
• The two oscillator output coming from the split-load inverter of V7 pass through blocking caps, then meet 470kΩ resistors R40 & R43 (which isolate this output from the bridge) and 1MΩ grid leak resistors R41 & R44.

Do you know how the Doppler Effect works? A siren traveling away from you emits sound waves, which reach your ear mixed as two waves whose phase is diverging (at the rate the siren is moving away from you); this sounds like a dropping pitch. When the siren moves towards you, the waves have converging relative phases, and the pitch seems to rise.

• The opposite phase oscillator signals mix with the 2 90-degree-apart bridge signals to produce an output at the shared plate load of V6 which is a combination of signals whose relative phase alternately diverge and converge. That is, the relative difference shifts from 90-degrees to more- and less-apart.
• Because the mix contains changing-phase signals the result is frequency modulation (pitch shifting/bending).
• Fundamentally, this works exactly like the Magnatone circuit, though it doesn't use varistors within the bridge circuit for mixing in the oscillator signals for pitch bending (yes, this counters the claims of some, but your ears will prove it's true). Some Magnatone circuits will have a lusher sound because there are more than one stage of phase-shift/modulation.

The push-pull application of the oscillator signal to the differential stage helps to cancel out the oscillator signal from the output (reducing pumping), but (probably) supply voltage limitations prohibited Vox from using a large tail resistor to insure common-mode cancellation. That's the real reason for the mass of caps & resistor ahead of the Vibrato channel volume which feeds one side of the phase inverter... These are high-pass filters to strip out the common-mode oscillator signal, and multiple stages made for a very steep (fast) roll-off.

This explanation is captured (with some better schematic drawings) in The Universal Vibrato and The Wurlitzer Vibrato. A careful reading of Don Bonham's patents for Magnatone-style vibrato shows that circuit works under largely the same principles.

[HotBluePlates]

Oh yeah... There is a Vibrato/Tremolo switch.

This grounds one side of the bridge output/oscillator output by way of an attached blocking cap. Because our ear does not detect absolute phase of a single signal but needs a reference to compare against, killing one of the signals which get mixed by V6 stops the vibrato effect.

However, the volume of the remaining signal is modulated by the remaining oscillator signal, and gives a trem effect.

[kagliostro]

This is a stand alone unit from Vox circuit by Merlin Blencowe



K

[G._Hoffman]

I haven't the kind of info you are looking for

But may be you can find interesting info here

http://www.guitargear.net.au/discussion/index.php?topic=30731.msg326407#msg326407

if you read all the post on the link, you'll find some that are very interesting

K

Yeah, I found that after I posted, but he kind of skipped over some of the details.

...exactly what I was looking for, elided for length...

THANKS!!!!  The bridge is still a bit of a black box in my mind (I get what it is does to the signals, I just don't get how it does it!), as the only electronics bridge I've dealt with was an H bridge I used to switch a bunch of two wire bi-color LEDs.

One small bit of pedantry (because I'm a bit of a pedant), but the Doppler effect doesn't just appear to raise/lower the frequency, but from the perspective of the stationary observer, it DOES raise/lower the frequency.  Relativity, don't you know.  Sorry.  I'm just like that, I can't help it.


Gabriel

[HotBluePlates]

The bridge is still a bit of a black box in my mind (I get what it is does to the signals, I just don't get how it does it!) ...

There is a simplistic first-principles method to understand what the bridge does, and a more complex and accurate way to understand what the bridge does.
 
Look at the image of a bridge circuit below. This is exactly the same as the Vox circuit I described, as far as the first split-load inverter and the RC components after it.
 
The tube itself is replaced by the a.c. voltage generator in the diagram (the squiggle inside the circle). Call the upper output of the generator "the plate output," and "the lower the cathode output." Z1-Z4 each take the place of a Resistor-Cap pair (a couple of those are series-resistor-cap, a couple are parallel-resistor-cap). Where "Null" is shown are the 2 outputs which run to the differential amp stage I described before.
 
Not shown are 2 parallel resistor-cap pairs which run from the 2 outputs of the bridge to ground. They are both parallel-resistor-cap pairs. Forget they exist for now.
 
Bridge Operation:
Look at the bridge as having 2 parallel paths from plate output to cathode output. Each of those paths is made up of an RC network instead of Z as shown in the simplified diagram.
 
In guitar amps, we typically think "caps roll off lows" because that is often the sonic impression that results. In reality, the cap has a changing reactance (depending on applied frequency) so that it forms frequency-dependent voltage divider with the resistor; in other words, the attenuation of the circuit changes with frequency.
 
Look at the upper line of the Filter Plot below, which is Amplitude vs. Frequency. The RC filter is a simple low-pass filter (a series resistor and a cap-to-ground), so attenuation is zero at a low frequency then gradually more as frequency rises. That's easy enough to grasp and fits with the basic understading of how caps work in an amp.
 
But reactive components store and release energy into a circuit, so the phase of applied voltage and applied current is not the same, as it would be in a resistive circuit/divider. So the phase of the voltage at the output of the filter also varies with frequency (if you need to know the exact phase at one frequency, you will need to use vector algebra to determine it). So the lower graph on the Filter Plot is Phase vs. Frequency. At zero attenuation, phase shift is zero. Phase shift approaches 90 degrees as the circuit approaches infinite attenuation.
 
The bridge essentially has two parallel paths from the plate output (bridge input 1) to the cathode output (bridge input 2). These signals are applied to one parallel path (say, Z1 & Z2 below); at the midpoint, the RC circuits for that side of the bridge shift the signal phase a certain amount. The same signals are applied to the other parallel path (Z3 & Z4), and at their midpoint the RC circuits on that side shift the signal phase a certain amount.
 
What is key is the RC components of each path through the bridge have a phase shift which is 90 degrees apart from the other bridge path at any given frequency. In other words, take the lower graph of phase on the Filter Plot, mentally duplicate it, and slide one of the graphs to the side so that at a given frequency one output is 90-degrees away from the other.
 
In practice, this is accomplished by changing the cap values between the same-positioned cap in the two bridge paths. The resistances stay the same. For example, the plate output of the 1st split-load on the Vox schematic shows two paths with a cap in series with a resistor. The resistors are both 68k, but the cap values are 500pF and 2000pF (0.002uF).
 
The phase of the voltage across the resistor of the series circuit is given by the equation at the bottom. When I ran the numbers for 68kΩ and 500pF or 2000pF at a frequency of 1kHz, I got that the phase for the 500pF case was ~30.33 degrees, while the 2000pF case was ~77.94 degrees. Those are ~47 degrees apart.
 
But that doesn't account for the input from the cathode signal and the shifts found due to their cap values. You will find the result is at the brdges two midpoints, the two outputs stay 90-degrees different regardless of the frequency applied and the total amount of phase shift in the circuit.
 
So then the modulator's job is pretty easy: make these signals alternate between less-than-90-degrees-apart and more-than-90-degrees-apart.
 
 

If you're willing to skip to solid-state, then you can find circuits called "Quadrature Oscillators" which have 2 outputs 90-degrees apart. This usually simplifies parts count and build complexity. See pages 18-19 of this TI Application Report.

[G._Hoffman]

Thanks again, HBP.  I'm starting to get it, but I don't think I'll ever really understand it until I actually set it up and look at on the scope. 

Actually, how different is the Magnavox circuit - I've got a friend's M6 here that has some odd fizzing noise that I haven't tried to work on yet (I think it just needs a new tube or two, but I've been so damn busy!), but this could motivate me to do something with it in the next few weeks. 


Gabriel

[HotBluePlates]

If you're willing to skip to solid-state, then you can find circuits called "Quadrature Oscillators" which have 2 outputs 90-degrees apart.

I'll need to retract this statement; after re-reading, you need the quadrature guitar signals, not quadrature oscillator signals. But the report may come in handy some other time.

... I've got a friend's M6 here that has some odd fizzing noise that I haven't tried to work on yet (I think it just needs a new tube or two...) ...

Right, it's probably an issue of a tube or dirty contact (or hissy resistor).

Actually, how different is the Magnavox circuit - I've got a friend's M6 here ...

Turns out, different enough to be different.

Some very good references can be found at:
F.M Vibrato
Don Bonham
R.G. Keen Phase/Flange Technology Article

And if you search for the patent numbers listed in the Don Bonham page, you'll find free copies of all 4 patents related to the Magnatone Vibrato circuit.

The R.G. Keen article does the best job of breaking down the explanation to first principles then building up to the Magnatone circuit. Basically, the Magnatone circuit works on the basis of varying phase-shift just like the Vox. But the Magnatone uses a variable resistance as the mechanism to cause a phase shift (as it works against a fixed-cap). Additionally, the Magnatone incorporates a bridge circuit, but uses the bridge as a way to both inject the oscillator signal and to ensure that it will be eliminated from the output after its job is done.

After you read the R.G. Keen article and grasp the basic function of the circuit, look at the Magnatone M8 schematic (my understanding is the M6 is similar and has only one phase shift stage).

Channel 1's 3rd triode (let's call this V2A for reference) is a split-load which divides the dry signal into a path for the phase cap and the "phase resistor" (which will be where the control voltage is applied to the varistors). The 820pF cap connected to the plate of that split-load is your phase cap.

The 2nd 12AU7 triode (call it V2B, right after that split-load inverter) is itself another split-load inverter, but is used to derive +LFO and -LFO signals.

The cathode output of the V2A leads to the "phase resistor" portion of the simple phase shift circuit R.G. Keen described. The bridge circuit for the application of the LFO signal to vary the "phase resistor" is made up of the two 0.047uF caps and the two varistors. The junction of the caps is the "left corner" of the bridge, and the junction of the varistors is the "right corner" of the bridge. The LFO phase splitter has outputs which directly connect to the "top-" and "bottom-corners" of the bridge.

This reverses the position of single blocking cap vs. bridge blocking caps shown in Keen's simplified diagram, but that doesn't change the circuit's function.

I asked the question several years ago when Magnatone vibrato came up if the varistors could be replaced with encapsulated LDRs, now that LDRs are cheaper and easier to come by. I never did sit down and try to build the circuit. Of course, now it just seems easier to implement the Univibe circuit using cheap & easy Vactrol LDRs (using either a tube or solid-state inverter); Much of the Magnatone circuit seems to be about adapting to the unique problems of using a varistor instead of an LDR, and how to remove the large LFO control voltage from the output.

[sluckey]

Here's a little more fodder for the fire...

     http://home.comcast.net/~seluckey/amps/magnatone/NON_LINEAR_IMPEDANCE_ELEMENTS.pdf

[HotBluePlates]

Here's a little more fodder for the fire...

Thanks for posting that!

Look at Figure 1 on the first page. The 2 a.c. voltage generators marked 15 & 16 are, in essence, the plate and cathode output of the split-load inverter which has the dry guitar signal. Point 13 is simply ground, or a 0vac reference.

The upper right arm of the bridge is what Keen called the "phase cap" and is the 820pF cap I mentioned earlier in reference to the Magnatone M8 schematic.


The lower right am of the bridge is of course where the varistors are, taking the place of Keen's "phase resistor". The a.c. voltage generator shown there (between 28 & 29) is the LFO signal, which we know to be derived from its own split-load inverter.

You could re-draw the lower right arm of the bridge to form a more-obvious diamond-shaped bridge circuit; essentially, the arm is another bridge circuit within the bigger bridge, to apply a big control voltage to the varistors but have that control voltage cancelled at the output of this "mini-bridge".

I think you'll agree that if you replace the left half of the overall bridge with a phase inverter, and the lower-right arm of the bridge with an equivalent variable resistor, that the whole thing reduces to the phase shift network Keen introduces first in his article. This fact is why it finally dawned on me that you don't really need varistors at all (or to copy all aspects of the Magnatone circuit) to get the Maggie's vibrato: it's cheaper, easier and lower-parts-count to implement the circuit with commonly available devices.

[sluckey]

I have a Magnatone M10A and I recently built the Vox circuit. The Vox circuit sounds good but the Maggie sounds much, much better. In fairness, I should say the M10A uses a dual modulator (two sets of varistor bridges)

[HotBluePlates]

And the M15 uses 2 stages of phase shifting, with a 2nd parallel set of 2 phase-shift stages. I bet that thing sounds thick...

[Willabe]

Hey, there's a bunch of schemos on this website that Doug doesn't have including the M15.   

http://magnatoneamps.com/


             Brad     

[sluckey]

Hey, there's a bunch of schemos on this website that Doug doesn't have including the M15.

Yes. And the owner of that site is a member here.

[Willabe]

Yes. And the owner of that site is a member here.

Soooo.... Can we add these schemos that Doug/we don't have??????


                  Brad       

[terminalgs]

Hey, there's a bunch of schemos on this website that Doug doesn't have including the M15.

Yes. And the owner of that site is a member here.

as owner of that site,  i have no problem with those schematics being spread around! most of them are copies of the originals.

There were at least six revisions of the M15 (before the M15A),  mostly changes to reverb sections, but also changes to signal path stuff.  I've got 2 M15 schematics up there, and an M15A (also a snippet of a running change).

[Willabe]

Well, thank you terminalgs!

And may I say your web site is great! Very informative.

I need to spend some time going through it all.


                Brad     

[EL34]

Yeah, cool, thanks for the schematics

I'll add them to the library

[MWagemaker]

hi, I am new here, this is because i newly adopted the hobby to do the repairs on my amps myself.
Old post, but from what i see on the internet maybe i can find some data on measurement on this vox vib trem circuit.
Anyone measured actual the phase shift at Point A and B, when refering to the point where the presumed phase shift is 90 degrees in the Wurlitzer mod 44, and fed to the mod tube? at different frequencies?
Point is, I want to check the vibr.trem system in my vox ac30-6 nTB, without desoldering all components to check if they are in range.
Anyone a idee?

greetings

[HotBluePlates]

... I want to check the vibr.trem system in my vox ac30-6 nTB, without desoldering all components to check if they are in range.
Anyone a idee?

Play it.  If you hear the pitch bending, the Vibrato circuit is working.

If you don't hear the pitch bending, do you at least hear the volume changing when switched to Trem mode?  If not, open the amp and adjust the depth reset trimpot inside to increase the depth.

If that doesn't work, make sure the tubes are strong.  There are two 12AX7s and one 12AU7 in that Vibrato circuit...  If other-testing confirms the tubes are strong, then check DC volts at each pin.



The point there is that you don't need "critical balance" to get the Vib/Trem circuit to work.  It will work (though maybe weakly) even if things are way-wrong.  It almost requires disconnecting parts to get it to fail.

[MWagemaker]

thanks HotBluePlates, checked the phase shift and amplitudes over the frequency range anyway. looks like it is working, but it would be  nice to compare data to some other same circuits.