Adding preamp stage.

[slash1986]

Hi all, I built this: http://paulamps.com/Schematic6gw8.gif amplifier, which sounds pretty good. What I wanted, however, was a little more distortion, and then I decided to put a 12AX7 in the diagram. the new circuit is attached
.
The resistance on the anode of the 12AX7 is 100k. have any suggestions?
Now the amp is very noisy and does not sound good. In some moments especially with the volume to the maximum you can hear a hint of auto-oscillation.

I then replaced the cathode resistor of 100k 12AX7 with a 56k but things have not improved. You will know show me the right values ​​for the resistance of the cathode and anode. I also removed the capacitor 1uf baypass from the cathode of the triode pcl86.

[zendragon63]

Hi slash, if you really built it based on the schemo you attached, you will have an abundance of noise--due to where you're pulling your preamp supply voltage from. The first stage is a particularly sensitive section because its signal gets amplified over and over through progressive stages. You should consider moving the plate supply for your 12AX7 addition further on down the line--maybe right off of the 272V point and then tweak your current 18K dropping resistor to suit your fancy for quiescence. Just an observation based on prior experience. The real pros here will jump in and make additional recommendation.

Kinda has that Matchless look  . As far as your Rp on the paralleled preamp triode sections, the folks at Matchless have already done a lot of R&D so you might start with their values http://www.el34world.com/charts/Schematics2.php and tweak from there. Regards

dennis

[SILVERGUN]

Hey slash,
I took a couple of minutes and edited your schematic to reflect what I would try if I was adding distortion to your preamp....
Just an idea and I'm no pro,,,but this will add distortion  , and it address the point that zendragon brought up
This will add a lot more gain than the paralled 12ax7 in your original schematic.......
It might get you closer to sounding like Slash 

[terminalgs]

yup,  what zendragon63 said.

I'm not sure what the original schematic designer was envisioning (Paul?).  maybe he pulled the circuit from a radio audio power section.   Or was it simply a challenge of a 3 tube push-pull amp?   Either way, for guitar the circuit needs more preamp gain (like you found).


Make your first gain stage (the 12ax7) similar to the 6GW8's gain stage , 220K plate R, and 2700 Cathode R.   are you using a 10K plate resistor as an attempt to lower gain?  If so there are better ways to do that.  what you've got would be the start of a decent reverb tank driver, not the 1st gain stage for guitar.  a 10K R needlessly increases current through the 12ax7 and will shorten its life,..  maybe thats a typo and you are using a 100K R...

if you want to start simple, eliminate the 18k/47uf (how'd you pick 47uf?)  and tie your 1st gain stage plate resistor to the 272V point (HT for the V1 6gw8),  and one use one stage of the 12ax7...


A more traditional way to pull this circuit off is remove that 500pf and 1M Log pot between V1/triode and V2/triode and change that 1K R under it to (at minimum) 1M.

replaced the cathode resistor of 100k 12AX7 with a 56k but things have not improved

I don't see a 100k cathode R ,  unless you are talking about the phase inverter.

[HotBluePlates]

I'm not sure what the original schematic designer was envisioning (Paul?). 

Let's guess Paul designed the amp to work with a guitar as the input.

Slash added a paralleled 12AX7. The amp already boosted the guitar's signal enough to presumably fully drive the output section. The tack-on gain stage was probably too much gain, made the signal too-big.

I think Silvergun has the answer. Whether or not you use the 12AX7 to be 2 gain stages cascaded, or stick with the original paralleled 12AX7, you almost certainly need the voltage dividers between the added gain stage and original circuit to knock down the signal size.

[scrimpus]

I have a very similar amp, this design is originally taken from a Hammond AO-44 reverb amp from an M-100 organ. I have added half a 12ax7 to mine and even that required a voltage divider and/or a tone control to reduce to signal to a manageable level. I love the sound of the ECL86s, very bright and jangly vintage sound for sure.

[slash1986]

First of all, thank you all for the answers and the help you are giving me.

If I understand correctly, the gain with the 12AX7, is too large, and when it enters the first triode of pcl86, creates noise. The way to reduce this signal is a voltage divider connected between the 12AX7 and the first pcl86.

Not being an expert on the issue  , I ask you if you can tell me how it's done a voltage divider and how to calculate the values ​​of the resistors.

[kagliostro]

You can do the thing using math

or simply try a pot (try with some different value pot) connected before the gain control between V1 and V2

this pot, when you find the right value, can be substituted with a pair of resistors of proper value or by a trimmer to be set and discard

however if you want to go directly with the partition resistor this can help you

http://amps.zugster.net/tools/voltage-divider

if you consider as example 20v input, you can go on trying different resistors values till at the output you have 10v, this way you know the partition resistor give you an output in the order of 50%, you try it and decide if the reduction is enough or redundant and then you can adjust further the output changing the resistors value

don't use too low resistors values, I think the whole partition resistor (R1 + R2) must be at last a 100k, not less, but may be someone can give a council about a better minimum value for this purpose

---

I agree with the council to modify the intake from which you get the voltage, the more B+ is filtered before you add the new node for the new tube, the better

you can also add a 200n cap between the filament wires directly on the pins of V1 this will help to prevent interference coming from the HV side of the PS

if happens you have a choke use it in the circuit, this will help you on lowering the noise due to the HV PS

Enjoy

K

[sluckey]

Could you show us a high resolution picture of the wiring for the 12AX7?

[SILVERGUN]

Not being an expert on the issue  , I ask you if you can tell me how it's done a voltage divider and how to calculate the values ​​of the resistors.

Not simple stuff  ....here's some great free info to get you going:
http://www.valvewizard.co.uk/Common_Gain_Stage.pdf

Otherwise, it is pretty common to just use proven values that tend to show up over and over again in common schematics.....most of us have stared at schematics for countless hours trying to "absorb" this info  ,,,,,it's a great place to start

Try to clean up your schematic, and make it accurate, so these guys can see what's really happening there 
You've got the attention of some of the best minds here (me excluded) 

[DummyLoad]

it's based largely on the hammond ao-44 reverb amp. originally had a germanium transistor driving the works. i thought that schema looked familiar...the pot though, something about the pot...
 
so no, i don't think it was paul. i'm betting it was laurens hammond's army of engineers that did most of the grunt work.

--pete

[HotBluePlates]

If I understand correctly, the gain with the 12AX7, is too large, and when it enters the first triode of pcl86, creates noise.

Let's define our terms.

"Gain" is just "amount of voltage amplification". It tells you that when a small a.c. voltage is applied, how big the output voltage will be.

The problem is 10lbs of sh!t in a 5lb bag.

A typical 12AX7 stage will have a gain around 45-60. Your parallel 12AX7 stage might be more like 70 (I haven't calculated it based on your schematic, simply a knowledge that parallel 12AX7's often sound ~30% louder than single 12AX7's).

The input ECL86 was likely designed to accept the 100mV or thereabouts that a typical guitar pickup with a typical strum will produce. If you notice the voltage across the cathode resistor is 1.5v. This means that the maximum signal input it can accept without saturating, spitting and sputtering is 1.5v.

Now let's assume 100mV input and a gain of your added 12AX7 of 70. 100mV * 70 = 7v(!)

The original ECL86 input stage is getting too, too overdriven. A voltage divider is normally 2 series resistor, with an output signal taken from their midpoint (like the 470k & 500k after the Gain control in Silvergun's schematic). A volume control is also a voltage divider, but allows you to vary the ratio of the size of the two resistors for different output levels.

So try this: You have a 1M volume control indicated between the 12AX7 and the ECL86. Turn this down to zero and (very slowly) increase it while playing. You will quickly hit some maximum point above which you just can't get any good sound. Turn the amp off and measure the resistance from the wiper to either outer lug of the pot. Write these down somewhere.

The amount of resistance from your C1 to the wiper of the 1M pot probably represents a resisatnce that is always need in the circuit to knock down the "size" of 12AX7 output. You won't be reducing the "gain" of the 12AX7, because that stays the same as it always was. You'll simply reduce how big the signal voltage is.

Let's say from C1 to the wiper you measure 600kΩ. That tells you the usable range of your 1M pot was 0-400k. Take out the 1M pot, and install a 500k pot. Add a 470k-680k resistor between C1 and the "top" of the 500k pot. Now you have a total resistance about equal to your original 1M pot, and the series resistor means you can't turn it up above some middle point on the original 1M pot.

We could calculate and guess that you need to knock 7v down to 2v or less and figure likely resistor and pot values, but it's faster and more accurate to try it and drop in appropriate parts like I described.

[SILVERGUN]

Great simple explanation HBP 

Slash, you could read 10 books and still not walk away with a simple understanding of that topic, and how it applies to your situation,,,,just because of the wording
The fact that these guys (who really know what they're talking about) take the time to slow this stuff down for us is priceless

If you can read his post 2 or 10 times and completely understand it,,,,it's worth every second