Virtual Earth mixer for Fender Twin, Super, Bandmaster

[Chevy]

Hi guys,

Anyone have any success using a virtual earth mixer stage in place of the standard reverb mixer stage on a standard Fender reverb topology? In other words, use the V.E. mixer stage to mix the Normal, Vibrato, and reverb return signals. Haven't tried this yet myself..... recall someone saying the V.E. stage sounds kind of sterile/undynamic or something like that.
Best,

[Chuck H]

Anyone have any success using a virtual earth mixer stage in place of the standard reverb mixer stage on a standard Fender reverb topology?... recall someone saying the V.E. stage sounds kind of sterile/undynamic or something like that.

Works just great. I've done it with good results. Much better separation between the dry and reverb signals. More clarity. If by "sterile" that "someone" means that your tone won't turn into reverb garble soup then I guess so. But I really enjoy the added clarity. Also, maybe unrelated, but my reverb is quieter (noise wise) than other "Fender" amps. The reverb cannot get as heavy in the mix. But who in the world ever uses a standard BF type reverb on 10.

I did mine kind of like a 6G16 Vibroverb using a split plate load but wired the recovery stage with a 22uf cathode bypass cap, a 100k "feedback" resistor and a pair of 100k resistors for mixing instead of the 470k R's in the schem.

Chuck

[Steve Conner]

I've tried it too, it works fine.

If someone says it's "sterile", well, you are putting negative feedback around a tube, and that is known to reduce distortion. So it won't produce as much tubey warmth as if you ran the tube open-loop and threw away the unwanted gain. 50 times less warmth, in the case of Chuck's unity gain example.

This also implies that you could use a MOSFET if you were running out of triodes :-)

BTW, I hate to be a nerd, but what in the world does "separation between dry and wet signals" mean in the context of a circuit whose sole purpose is to combine them into one signal?

[Bruce / Mission Amps]

Works just great. I've done it with good results. Much better separation between the dry and reverb signals. More clarity. If by "sterile" that "someone" means that your tone won't turn into reverb garble soup then I guess so. But I really enjoy the added clarity. Also, maybe unrelated, but my reverb is quieter (noise wise) than other "Fender" amps. The reverb cannot get as heavy in the mix. But who in the world ever uses a standard BF type reverb on 10.

I did mine kind of like a 6G16 Vibroverb using a split plate load but wired the recovery stage with a 22uf cathode bypass cap, a 100k "feedback" resistor and a pair of 100k resistors for mixing instead of the 470k R's in the schem.

Chuck

I have a 6G16 in for a tune up and one of his biggest complaints is that the Vibroverb's "reverb" is weak and doesn't have anywhere near the depth of his other Fender amps.
I can't really find anything wrong with the amp, but he's right... the reverb on 10 sounds like reverb on 2 1/2 to 3 1/2 on any newer blackface or silverface Fender reverb amp.
******************

With respect to the two 470K resistor divider and using 100K resistors instead, how do you keep the 100K reverb level pot from effecting the overall output level of the preamp stage at the split load?
The reverb 470K resistor appears to be used more as a isolation resistance to the grid of the recovering triode.
In other words with the reverb level set to zero.... does the overall gain drop quite a bit?

[Chevy]

With respect to the two 470K resistor divider and using 100K resistors instead, how do you keep the 100K reverb level pot from effecting the overall output level of the preamp stage at the split load?

Not an expert on this, but if I understood correctly, the virtual earth mixer is supposed to keep the inputs totally isolated from each other by virtue of the "virtual earth" point at the junction of the mixing resistors.

[Chuck H]

Yes, The impedance at the input (the junction of the three 100k resistors) is very low. You could actually ground the opposite end of the reverb mix resistor and I don't think you would notice a change in the dry signal.

And yes, the reverb level is too low. I am using a 500k pot on mine and it's still low. I plan to try a smaller reverb mix resistor in the circuit to bring it up. I won't know the drawbacks of this idea until I try it. But I haven't fired the amp up up in a long time (it's actually in storage right now).

And yes, any mixer stage is intended to combine the two signals. But since the VE has almost no interaction between the signals prior to mixing AND is wrapped in NFB so many peripheral distortions that result from blending signals is reduced, the two signals after mixing sound more like they did before mixing. Not as much garble and loss of nuances in the character of either signal. More clarity.

Chuck

[Chevy]

And yes, the reverb level is too low. I am using a 500k pot on mine and it's still low. I plan to try a smaller reverb mix resistor in the circuit to bring it up. I won't know the drawbacks of this idea until I try it. But I haven't fired the amp up up in a long time (it's actually in storage right now).

Chuck

Chuck, from what I undestand, you don't have to make the v.e. mixer stage unity gain. You can use a 220K feedback resistor, for example, with 100k mixer resistors, and give it a gain of around -2. This would beef up the reverb level. Then to make preamp levels correspond with the reverb level, you can reduce the level of the preamp stages via a) the selection of different split-plate load R's or any other way you choose or b) perhaps use 220K mixer R's for the two preamp channels, making them unity gain at the v.e.stage. I don't know if mixing values in the v.e. mixer resistors will mess up anything, but I doubt it.

[Chuck H]

Yes you could bring up the reverb by decreasing local NFB. You noted that the VE doesn't NEED to be a unity circuit. For my amp it just worked out that unity gave me what I wanted. But with that in mind it also stands that the VE shouldn't need to be at unity on all inputs to do a good job mixing. Maybe better, but perhaps not needed for this application. By reducing the input R for the reverb you effectively change the ratio for the reverb without changing the dry signal. I suppose you could do the same thing by changing both the feedback R and the dry signal input R to the same higher value. Like 220k. Then you would still have unity gain on the dry signal and the reverb should be louder in the mix. It seems like the same difference to me. But my by ear testing helped me to choose the 100k feedback value. Maybe something to do with the fact that the plate R is also 100k. Could be some kind of electron dance party going on that I'm not aware of, but can hear just the same. And by lowering the reverb input R instead of both the feedback and dry input R's, I only need to change one resistor.

Chuck

[Chevy]

Gotcha....
I misunderestimated the situation What you say makes good sense.

I guess my only concern would be using a mixing R of less than 100K...... don't know how happy that would make the reverb recovery stage (or other preamp stage for that matter) in terms of it's loading.

Perhaps I'll get around to trying this some time soon... sounds promising. Wonder why it wasnt' a Fender standard thing, as they did use it on one of the Bassman amps to mix 2 preamp channels.

[Chuck H]

Wonder why it wasnt' a Fender standard thing, as they did use it on one of the Bassman amps to mix 2 preamp channels.

Very true. I've never had or built one on those but from what I've read here they sound great. And the only circuit changes I read about in those amps is the removal of the cap across the plate load in the early preamp to make the bass channel more guitar friendly. After that they're said to sound better on the bass channel than the normal channel...Hmmm.

Chuck

[diagrammatiks]

was just going to post a question on this and this thread was right on top...hope no one minds an additional question here...

in merlin's preamp book he says that the virtual earth mixer doesn't need a ground reference at the grid. However, When I don't have a ground reference at the grid and only the 1meg resistor to the output side of the coupling capacitor I'm getting a ton of buzz/noise that is unaffected by how much signal the mixer is receiving.

When I have a resistor to ground right at the grid of the mixer the buzz goes away.

Any clue what I'm doing wrong? Is the stage still a virtual earth if I have a grid to ground reference?

[Chuck H]

I'd need to see a schematic of what your doing to be sure... But I will say that VE mixer or otherwise any tube grid needs to have an actual (not virtual) 0VDC reference point to bias correctly. Since capacitors block DC I don't see how providing a 0VDC reference is possible without a grid to ground resistor. You could always put a 1M resistor in the circuit to see if this is the source of your noise.

[Wombaticus]

was just going to post a question on this and this thread was right on top...hope no one minds an additional question here...

in merlin's preamp book he says that the virtual earth mixer doesn't need a ground reference at the grid. However, When I don't have a ground reference at the grid and only the 1meg resistor to the output side of the coupling capacitor I'm getting a ton of buzz/noise that is unaffected by how much signal the mixer is receiving.

When I have a resistor to ground right at the grid of the mixer the buzz goes away.

Any clue what I'm doing wrong? Is the stage still a virtual earth if I have a grid to ground reference?

Depending on how you set up the VE mixer, it *is* possible sometimes to "borrow" the grid reference from the stage that follows, perhaps this is what you are talking about....this really only applies when you have the mixer setup so that the feedback resistor wraps around the coupling capacitor between the mixer and the next stage. You can't rely on this when the feedback loop has a resistor AND its own dedicated coupling cap, which is a common arrangement.

There are a number of variants to the circuit. The arrangement I favour for VE mixers is to have a feedback resistor coming directly from the plate to the "mixing point" where the inputs meet... but this mixing point is separated from the grid by a cap, typically 100nF or so. In this case, the grid may have its own reference resistor without worrying about messing things up. This is the last arrangement mentioned by Aiken in his great tutorial: Designing Single-Stage Inverting Feedback Amplifier