Merlin's 12AU7 effects loop

[kuDo]

Hello Everyone,

My Name is Douwe and I Joined this forum a couple of weeks ago.
I am into Amp building and so far I completed 1 tube pre-amp which is working beautiful.

This time my plan is to build a complete Marshall 2204 inspired amp. A entire amp is the next step for me. It is al about learning for me and the joy of building.

Oke, the amp that I want to build must have a effects loop. This is the final part off the puzzle for me, but I have trouble understanding it entirely.

Attached you find a scematic of a effects loop that Merlin has designed.

Merlin says "The ECC82/12AU7 is well suited to this job because it is cheap and plentiful, its low internal anode resistance allows it to deliver relatively large signal swings into heavy loads, and its moderate transconductance will guarentee fairly low output impedance which ensures wide bandwidth. The second triode in the ECC82 could of course be used to amplify the returning effects signal."

This sounds good to me, so I would like to understand it a little better.

A big part is explained here:
http://valvewizard.co.uk/accf.html


Question

This sounds good to me, but what I don't understand is this.

1. He says that the load resistor Rk is preferably smaller than the impedance of the stage that it will be driving. Why is that? Isn't the output impedance of this cathode follower already somthing like 420R? (420R is including the 18k Rk)

2. In the attached scematic you can see that after the first stage there is a potentiometer to adjust the signal to whatever pedal is used.
If the pedal resembles a load of 10k, how does this al work? The cathode follower is stated to be 420R, A low output impedance, then a pot is 22k follows and parallel to that is the load of 10k.

Is there someone on this forum that would be so kind to help me out? I have been searching for a while the last couple of weeks.
But I just don't understand how it works.

Greetings,
Douwe

[Ten Over]

The output impedance forms a low-pass filter with the capacitance of the cable plugged into the FX Send. A low output impedance is desirable because it allows you to use long cables without treble loss. The output impedance doesn't come into play when determining what load a cathode follower can drive.

Drawing load lines is a good way to see what happens when you use different Rk's and biasing with various loads. It usually comes down to what amount of current you are willing to use and what kind of signal swing you need. You don't see 12AU7's in guitar amp FX loops very often because they tend to use more current than the entire rest of the preamp by the time you get a decent signal swing out of them.

This isn't the first time I have seen that loop that supposedly came from Valve Wizard, but I have never been able to find it.

[kuDo]

Never thought of the cables running to the effects pedal. That makes sense.

But the low impedance, high impedance thing is also important when stages connected with each other.
What I red about it is that I should see the connected next stage or load as being in series with the output of the cathode follower. This never lands with me, because to my eye they seem paralel. So the output impedance is parelel to the input impedance of the load. What hard about that is that they share the same voltage. So the voltage transfer doesn’t seem to be an issue. But I know that must be wrong. Do you know how I should see this?

I am not sure this schematic is from Merlin, but it is always mentioned when I see it.
Maybe I should go with the version with the 12ax7’s also precent on this forum.
The current draw of these 12au7 might be a thing. I can see those tubes create a hefty voltage drop over the dropping resitor between de stages that are been supplied of their power needs.

You said:

The output impedance doesn't come into play when determining what load a cathode follower can drive.

What do you meen by that?

Thanks for your help so far!

[martin manning]

Check this thread out: https://ampgarage.com/forum/viewtopic.php?f=6&t=25351

[Ten Over]

But the low impedance, high impedance thing is also important when stages connected with each other.
What I red about it is that I should see the connected next stage or load as being in series with the output of the cathode follower. This never lands with me, because to my eye they seem paralel. So the output impedance is parelel to the input impedance of the load. What hard about that is that they share the same voltage. So the voltage transfer doesn’t seem to be an issue. But I know that must be wrong. Do you know how I should see this?

For AC purposes, the output impedance of the voltage source is in series with the parallel combination of the DC cathode path, the send pot, and the input impedance of the FX device. But the voltage source output impedance is so much smaller than the other impedances that it can safely be ignored for load-driving design purposes. This is the kind of impedance mismatch you want for voltage transfer, which is what we are interested in.

When we determine what load a cathode follower can drive, we use the DC resistance between the cathode and ground for the DC load line like Valve Wizard did. A lot of the time the bias resistor is ignored when Rk is significantly larger. Then we decide on a bias point and run the AC load line through it. The AC load line is basically everything in parallel. Now we know how much signal swing there will be for that particular load. The cathode follower output impedance was never used in this process just like the plate output impedance is never used when drawing those load lines.

The design parameters that you want to meet will determine the values you choose. The only time the output impedance of the cathode follower comes into play is when determining the worst case output impedance that is driving the patch cable.

[TUBEDUDE]

Greetings KuDo,
If you use an ECC832 or a 12DW7 for the loop, you get one triode from a 12AU7, and one triode from a 12AX7. So you get a good current driver for the loop send and a higher mu recovery stage for better gain in the recovery stage.

[kuDo]

by martin manning » Mon Nov 02, 2020 12:37 am

Check this thread out: https://ampgarage.com/forum/viewtopic.php?f=6&t=25351

Thanks Martin, I am inclined to go with that schematic, but I would like to understand this topics schematic a bit better.
I have got a lot to learn that's why Is started this topic here.

When we determine what load a cathode follower can drive, we use the DC resistance between the cathode and ground for the DC load line like Valve Wizard did. A lot of the time the bias resistor is ignored when Rk is significantly larger. Then we decide on a bias point and run the AC load line through it. The AC load line is basically everything in parallel. Now we know how much signal swing there will be for that particular load. The cathode follower output impedance was never used in this process just like the plate output impedance is never used when drawing those load lines.

The design parameters that you want to meet will determine the values you choose. The only time the output impedance of the cathode follower comes into play is when determining the worst case output impedance that is driving the patch cable.

Thanks for the information Ten Over!

I just made myself a couple of load lines for the cathode follower part of the schematic. When I look at those load lines I think al three of them would be useable.
The al seem to provide enough voltage swing for my purpose. I am however curious which design decisions you would make and why for this setting.

by TUBEDUDE » Mon Nov 02, 2020 10:13 pm

Greetings KuDo,
If you use an ECC832 or a 12DW7 for the loop, you get one triode from a 12AU7, and one triode from a 12AX7. So you get a good current driver for the loop send and a higher mu recovery stage for better gain in the recovery stage.

Thanks for the tip. I will thake a look at it. After I decrypted the version with the 12au7's.

[kuDo]

Isn’t it weird that you use the parallel resistance for drawing the AC load line. The output impedance value is much lower. If I was to draw a loadline with that, the voltage swing would be a lot smaller. What is the Impedance value good for anyway?

[kuDo]

If I was to go for this schematic I would probably go for these numbers for biasing.

bias current = 9mA
output swing max = 43V.
Output impedance = 625 Ohm

There seems to be lots of voltage swing capability. The output impedance is low as it is expected to be.
Al is within the maximum power rating of the tube.
The downsize is that compared to a 12ax7 there is more current needed.

If I was to use this plan, I think I need to attenuate the input signal as I think it will be much to hot for a signal going to the effects pedal.
In front of this loop will be 3 stages of a JCM 800 preamp and via the tone control circuit.

For attenuation I can put a voltage divider in front of the input. But Is that a good idea? This stage is having a really high input impedance. Won't a divider just ruine that?

[Ten Over]

Isn’t it weird that you use the parallel resistance for drawing the AC load line. The output impedance value is much lower. If I was to draw a loadline with that, the voltage swing would be a lot smaller. What is the Impedance value good for anyway?

Notice how the cathode follower load lines are exactly the same as common cathode gain stage load lines? The output impedances are radically different clearly demonstrating that they don't come into play when determining the behavior of the triode.

What the output impedance is good for has already been explained.

[kuDo]

Notice how the cathode follower load lines are exactly the same as common cathode gain stage load lines? The output impedances are radically different clearly demonstrating that they don't come into play when determining the behavior of the triode.

Yes, I see they look a lot like the normal gain stages. Although the cathode follower has 100% feedback for the cathode so this doesn’t show what actually is going on, if I am not mistaken. Amplification just below 1.

What did you mean by clearly demonstrating that they don’t come into play? What do you see?
Sorry if I am a little slow.

[Ten Over]

What did you mean by clearly demonstrating that they don’t come into play? What do you see?

I don't think restating it is going to do any good. How about this: The internal resistance and, therefore the output impedances, are already included in the plate characteristics curves. You wouldn't add them in again when drawing load lines.

[Ten Over]

Let's talk about 2204 FX loops.

The signal from the tone stack needs to be attenuated. There are three main choices: Before the cathode follower (CF), after the CF, or both.

I have documented 110Vp coming out of a 2204 tone stack. If this signal is not attenuated, the CF will need to swing at least 110Vp. This is a problem, so the signal is going to have to be attenuated at least somewhat before the CF.

The tone stack in a 2204 is loaded by a 1M pot, so whatever attenuating circuit you put in front of the CF must present a 1M load to the tone stack in order to prevent changes to the tone stack characteristics.

The input impedance of the AC coupled CF is very large, so the parallel contribution to the 1M tone stack load is minimal. The input capacitance of the CF is very low, so high frequency attenuating is not an issue.

The attenuating circuit can be the send level control if you desire.

[kuDo]

Thank you, dat is very helpful.

The signal from the tone stack is around 110v? Wow that higher than I thought.
That is indeed way to hot to pass through to the effects. Is this setup how you personally have build and used it?
The signal needs to be attenuated a lot. Doesn’t that mean that the biggest part of the pot is unusable?
The signal needs to go down to input signal lvel right? What is that, about 150mV?

So the impedance will be around 1M with the pot in place, that’s oke, because there was already a pot in place. The master volume. Seems logical.

But if I attenuate, the recovery stage needs to be able to get the signal back to 110 volts again.

[kuDo]

I don't think restating it is going to do any good. How about this: The internal resistance and, therefore the output impedances, are already included in the plate characteristics curves. You wouldn't add them in again when drawing load lines.

I see you have replied twice. Thanks for explaining. Sounds so simple really.

Am I right when I say that you can’t see the effect of the feedback within the CF? As in, when you put signal in, the graph says it will amplify, but I know the signal won’t get stronger than the amplitude that gets in due to the feedback.