AC10 Intensity in AC4? <<Solved>>

[dwinstonwood]

Hi all,
Recent talk of tremolo/vibrato volume issues led me to revisit my AC4 where I used a Vibro Champ tremolo circuit. The effect has good range, but even after tweaking various values I still get a volume drop with the Intensity turned up.

So, that got me thinking... would the AC10 Intensity work with my AC4/VC tremolo?

Here is my proposed schematic. I added that 82K load resistor on the cathode follower's cathode. Is that necessary, or a sensible value?
The values around the EF86 are directly from the AC10:
https://el34world.com/charts/Schematics/files/Vox/Vox_ac10_2.pdf

I like the fact - mentioned by HBP - that when the effect is turned off, the Intensity pot becomes a variable resistor for the EF86's screen.

Thanks!

[HotBluePlates]

... I still get a volume drop with the Intensity turned up. ...

I'm not convinced this is a problem, because:

   -  Trem turns the tube OFF half the time; apparent volume naturally drops.

   -  This fact caused Fulltone to add a boost after the trem circuit in their Supa-Trem, with a Volume control so users could match the loudness of no-Trem.  Compare the versions photographed below.

 


Apart from that, the AC10 doesn't have a "cathode follower" and it's better to look at its cathode-connection a different way:

   -  The oscillator tube passes plate current at a rate/quantity set by the oscillation.

   -  The tube physically passes current only from Cathode to Plate.

   -  Connecting the Oscillator's cathode to the EF86 cathode allows the Oscillator to pull current through the EF86's cathode resistor, changing the voltage across that resistor, which varies the bias of the EF86, which imparts tremolo.



    Your circuit is halfway between the Vibro Champ & AC10.  There is potentially too-much signal and/or an overheated buffer tube if there is not a big resistor between buffer cathode & EF86 cathode (because it is also the "plate load" and "bias resistor" of that buffer tube).  But it's a "try & see" situation unless you just connect the Oscillator's cathode to the EF86 cathode, as the AC10 did.

    The Intensity control in the AC10 effectively changes the bias on the EF86, deciding whether to not the Oscillator's cathode-current is a big effect or not.

[PRR]

...I'm not convinced ...

Agree. Trem "should" go up and down. But a Champ-scale amplifier is already maxed-out on gain. It has no more "up" to give. We can only pump "down".

There are other ways to do it. More complicated. Lose the simplicity of a basic amplifier.

Push-pull power stage can help. SE bias trem thumps. That's why we sometimes have a C-R-C-R-C-R filter after the modulated stage.

[dwinstonwood]

Thanks HotBluePlates and PRR for your explanations, as usual. 

Looks like the way I have it is good enough. But, tremolo circuits are a thing I need to get a better handle on. I'll try to study these two circuits and get a better understanding of what's going on with each of them.

[dwinstonwood]

The two are almost identical. If I went with the original AC4 circuit (just a matter of changing a few resistors), maybe I could use a 47K cathode resistor on the CF, and a 50K Intensity pot to equal the AC4's 100K resistor at full turn? I'll also add that 10M resistor that comes off of the last B+ node. So, I'd be all the way back to the original amp.  It's easy enough to experiment with.

[66Strat]

An alternative could be to replace R9 with a 250K to 500K pot and then connect R9 between the between the 12AX7 cathode and the pot wiper.

EDIT:
Another option could be to implement the AC10 trem circuit. NOS ECF82/6U8a tubes are readily available. The only downside that I can see would be an additional 0.15 amp current draw (12AX7 with vs ECF82/6U8a). The upside would be an additional triode to play with.

[dwinstonwood]

Thanks 66Strat. I'll have some time this week to try these things out. It's all fun.

[dwinstonwood]

Well, I got things sorted out. I went back to the original AC4 V-1-1 EF86 resistor values, and the tremolo is now working better than it ever has.

The biggest change seemed to come from going with the 5.6M screen resistor. This made a huge difference in the depth and strength of the tremolo. I don't know why? There's still a volume drop, but not like it was. It's easily fixed with a small turn of the guitar volume. It's not bothersome.

Pretty much everything I've read on the web recommended either a 1M or 2.2M screen resistor for the EF86, so the 5.6M was the last change I made. It's sort of a weird value, and I wonder if Denney worked up to 5.6m after trying other values first. Who knows.

The tremolo circuit is a mix of the AC4 and Vibro Champ. I went with the AC4 Speed control, but didn't add the 10M kick starter.

Thanks everyone for all the input and help!

[kagliostro]

Ciao Dwinstonwood

Please can you post a schematic with the last mod

Many Thanks

Franco

[dwinstonwood]

Thanks kagliostro!

Here it is:

If I was building this over I'd use a Hammond 194A choke instead of that first 1K dropping resistor. If you lean towards the amp there is a little hum.

[HotBluePlates]

... I went back to the original AC4 V-1-1 EF86 resistor values, and the tremolo is now working better than it ever has. ...

Congrats!!  Glad the amp is working well now!

... I went back to the original AC4 V-1-1 EF86 resistor values, and the tremolo is now working better than it ever has.

The biggest change seemed to come from going with the 5.6M screen resistor. This made a huge difference in the depth and strength of the tremolo. I don't know why? ...

Pretty much everything I've read on the web recommended either a 1M or 2.2M screen resistor for the EF86, so the 5.6M was the last change I made. ...

Looks like "copy the original schematic" would have been wise!   

Why:

We bias Triodes by setting the grid-to-cathode voltage.  But Pentodes have 2 useful grids, and so we add screen-to-cathode voltage as a "second bias" for the tube.

If we tie the screen to the plate to operate the Pentode as a Triode, it will exhibit an "Amplification Factor from Grid 1 to Grid 2."  And where Triode Amplification is an expression of how much more effective a voltage-change on Grid 1 is at controlling plate current than a voltage-change on the Plate, this "Mu g1-g2" is an expression of how much more effective a voltage-change on Grid 1 is at controlling plate current than a voltage-change on Grid 2 (screen).

But changing voltage on the screen does change plate current.  And screen-current is usually a fixed ratio of plate current, so using 5.6MΩ instead of 1MΩ lowered the screen voltage, which then lowered the plate current.

Now that plate current is less, a smaller voltage from G1-to-cathode (a "smaller bias") will give our desired idle current.  This also means the stage accepts less drive signal, or said a different way a smaller drive signal gives all the output the stage can make.  This observation leads to noticing that pentodes have a higher voltage-gain when their screen voltage is smaller (though there are competing tradeoffs).

     So What?

What this means for you is the larger screen resistor lowered the screen volts, which moved the bias of the tube, increased the tube's apparent Gain, which allowed a smaller tremolo signal to have greater impact.

... Pretty much everything I've read on the web recommended either a 1M or 2.2M screen resistor for the EF86 ...

When you don't know, copy a known-good plan.

Others didn't know so they copied the AC15's EF86.

You didn't know, so you copied what others had copied.  Except you were building an AC4 that wiggles the EF86, so you should have copied the AC4's EF86.

[dwinstonwood]

Thanks HotBluePlates! Those are great explanations as to why/how the pentode's screen voltage plays such a big part. I'll still need to read through your post a few times and really think about it. I want to get a better visual grasp on how the lower screen voltage shifts the bias point and increases gain. I'm sure there's a way to see that on an EF86 datasheet graph. I'll try to work that out.

And your right, I should have stuck with the original schematic. Why I think I need to "improve" an amp before it's even built, I don't know.

Thanks again! And, it is a great little amp.

[HotBluePlates]

... I want to get a better visual grasp on how the lower screen voltage shifts the bias point and increases gain. I'm sure there's a way to see that on an EF86 datasheet graph. I'll try to work that out. ...

Start by reading Kuehnel's Pentode information.

Part 2 (right at the top) is where Kuehnel explains why the lower screen voltage results in higher gain.  And he does it with a set of transfer curves from the EF86 data sheet, which satisfies your "from the data" desire.


The tradeoff to see (and this is common with most tubes) is that low screen voltage gets the plate current swing with a smaller grid-volts input, but it takes a higher screen voltage to get a larger plate current output (if that is desired).  This sorta implies that a large plate voltage output (albeit at lower gain) probably requires a higher screen voltage.

To make sense of that last bit, assume the plate load is 100kΩ in either case, then use Ohm's Law to determine the output voltage for a 1mA peak plate current swing, and a 4mA peak plate current swing.


You can use the information above to think about output tube pentodes/beam tubes, too.  So just remember that our EF86 is typically used in the Input Stage, and the guitar signal is rarely more than 100mV unless boosted by pedals.  Then consider how "getting all the output with a small input" is likely more desirable...

[kagliostro]

Many Thanks for the schematic Dwinstonwood
Franco

[dwinstonwood]

Thanks for all the info HotBluePlates!

I'm reading the Kuehnel (a few times over), and Blencowe's chapter on pentodes for the nth time. It all makes a lot more sense having actually battled with a real EF86.

Here are my final drawings. I sort of wish I had a Vibro Champ to to pit head to head with this amp. I think this amp would hold its own. One last change I made was to raise the heater ground up to the EL84 cathode voltage.