30 watt from scratch (guidance please)

[hesamadman]

This is in relation to the 30 watt I had built over the summer. It's been a great amp but I have had excess noise. I have done multiple things to eliminate it. Moved the OT around. Tried to fix some grounding but I think my layout was poor. Etc.


So I am going to re do it, and document the entire thing. After reading tons of material over the passed couple of weeks, I have decided to try the galactic star ground approach and place the filter caps on the board close to their associated components. I have a layout that I re-worked. I have not laid one out like this before so I would love some input on this.


I also want to mention that I used a 17" x 10" x 3" chassis. This is a combo amp so the tubes were mounted on the bottom which left about a 6 inch gap from tube sockets to the board, so my wires are WAY too long. This right here could very well be the culprit for my noise. It could also be how my wires are ran from tubes to board. During the new application, I plan to twist all of my associated wires together like in Merlins book (lead dress section). He has each section of a triode twisted together (pins 1, 2, 3 together and pins 6, 7, 8) which I have not tried.


I also ordered new capacitors, my original design was using dual caps.


My layout is not yet 100% as I still need to connect components to the appropriate tube sockets of the EL84.


I did want to see if I could get input on the placement of my 8uf filter cap for the PI and where it is grounded. I wasnt sure the best place to put it. I would think its fine where it is but this grounding style is new to me.

Some values may be different as when I changed some in the amp, I changed it on the layout and not in the schematic

[Willabe]

Are you going to re-use the same 17" x 10" x 3" chassis?

[hesamadman]

Shoot no. 17 x 6.5 x 2.5

[Willabe]

    Ok, just checking.   

[hesamadman]

Doing some reading. Stumbled upon grid leak resistors. It is optimal to have it placed before the grid stopper rather than after. On v2a of my schematic I have a 470k grid stopper bypassed by a 500pf. Following this, I have a 470k grid leak. Would it be a good idea to change those around so that the grid leak comes before the stopper. Also, considering putting the grid stopper resistor and cap directly to the tube socket. Doing this I feel I should use a shielded cable after the cc though. Any thoughts?

[Willabe]

Doing some reading. Stumbled upon grid leak resistors. It is optimal to have it placed before the grid stopper rather than after............  Would it be a good idea to change those around so that the grid leak comes before the stopper.

Technically a grid stopper works best mounted directly to the tube socket's grid pin with only a very short lead from the resistors body to the grids pin. Even though Fender and others did not do it this way for the 1st stage grids, they did usually mount the power tube grid and screen grid R's directly to the tube socket.

The deep theory guys can tell you why but it's got to do with capacitance/resistance/grid pin reacting with each other to kill off the gremlins for stability. (HBP and probably PRR and others have explained this before, I should have it memorized.)

It's more important with higher gain amps as they can have stability issues because of all the gain, but still very good build practice to mount them directly to the grid pin.

On v2a of my schematic I have a 470k grid stopper bypassed by a 500pf. Following this, I have a 470k grid leak.

What you are describing here is not a grid stop but a high pass filter to limit/stop too much signal from slamming the next grid. So you would leave it as is. Most classic amps only use grid stoppers on the 1st input gain stage grid, the power tube grids and on the output power tubes screen grids. (What I said above still holds true.)

You will see grid stops on hi gain amps in 2nd, 3rd, ect, stages along with, and/or series hi pass filters.

The hi pass series filter limits the low end of the signal but lets the hi end pass according to the caps value. The smaller the caps value the less hi end can pass around the series R.


                        Brad   

[hesamadman]

I'm definitely going to lay the tubes out in a way that I can keep leads as short as possible. Should I nix the shielded cable on the grids of stages after the CC or should I stick with it?

[Willabe]

I'm definitely going to lay the tubes out in a way that I can keep leads as short as possible. Should I nix the shielded cable on the grids of stages after the CC or should I stick with it?

Maybe, maybe not.   

When you say 'as short as possible', 1 mans ceiling is another mans floor. As possible might mean, because of the layout, ~6"/7"/plus inches and it can depend on what those grid wires are close too, ie, hi voltage/hi current wires and transformers. So it depends on the amps circuit, your layout, the chassis size and type (mild steel is a better shield than alum.), ect.

Some (most) guys go with shielded cable for any/all long grid runs, some start with just the input jacks, TS grid wires and the power tube grid wires (another would be the reverb recovery stage grid wires), then go from there adding shielded cable if  they encounter problems.

Another thing is the tone stack. After the tone stack caps, which are used as the coupling caps in most amps for the TS recovery gain stage, all those wires going to the tone controls and then to the recovery stage grid are grid wires. So to shorten up these wires some guys mount the TS caps (and TS slope R) directly on the tone controls.

They still use shielded cable going back from the TS to the recovery stage. AND you don't need to use shielded cable for the short little wire runs between the TS caps.

The questions your asking are great but I think it would help you greatly to look at Doug's and Sluckey's builds to see what where they are using shielded cable. 'A picture says a 1000 words." I'm trying to say to add them to your studies. I've posted the links for Doug's and Sluckey's in the grounding thread and other threads.


                             Brad             

[hesamadman]

Well I'm definitely going to go from a 6 inch gap between tube sockets and board to a maximum of 1.5". The only long grid run I should have would be input to v1a grid and preamp volume to v1b grid. Which is similar to a 2204 layout. I'll shield those.

I found a mistake on my layout. Filter cap to PI is incorrect. B+ should go through the 22k 3w an THEN hit filter cap and plates of PI.

I'm going through this with a fine tooth comb this week in hopes to assemble the board this weekend.

I have a SS reverb going in. Power supply will mount on small board just under amp power supply. Reverb components are on a custom PCB. Very small. Will mount it just below the board under the two CC between last stage and PI.

[John]

I have a SS reverb going in. Power supply will mount on small board just under amp power supply. Reverb components are on a custom PCB. Very small. Will mount it just below the board under the two CC between last stage and PI.

That power supply. If you're running off your 5V winding (nothing else connected) that's fine. If you're using your 6.3 heater supply, I found that I got a significant hum doing that, no matter where or how I routed the wires. I used a small 6.3 v transformer for it and the hum disappeared. The little tranny I actually mounted inside the chassis, and still no problems after that.


If you already had that figured out, then you know more than I did.    If not, maybe that'll save you some troubleshooting.

[hesamadman]

I'm actually using a separate 36v transformer to get my +15/-15 I need for the TL072??? I think that's the model number of the IC anyway. My amp was noisy but not due to resistor circuit. I removed it completely and no change.

[John]

Oh, okay! You're using a different 'verb than I did then. Carry on. 

[hesamadman]

Another thing is the tone stack. After the tone stack caps, which are used as the coupling caps in most amps for the TS recovery gain stage, all those wires going to the tone controls and then to the recovery stage grid are grid wires. So to shorten up these wires some guys mount the TS caps (and TS slope R) directly on the tone controls.

Did a comparison of Sluckeys ac15 to Dougs ac30. http://el34world.com/Hoffman/files/Hoffman_AC30.pdf Looked like Sluckey used shielded cable for inputs only. On Dougs, he has his input lines drawn in purple and he has them labeled as shielded cable. He also has his wires from TS caps to controls drawn in purple and his wire from tone stack controls to PI in purple. They are not labeled specifically but I wonder if the purple is meant to imply shielded?


Edit: Actually the two other purple wires are both just going to PI. One from normal channel volume control. The other from Top boost tone stack to PI.

[hesamadman]

I changed the layout slightly. I moved the PI cap next to the PI. Grounded it with the cathode of the PI. Also repaired a couple mistakes. The way it sits now, I have the ground of the PI and the B+ of the PI nd of crossing at an x pattern. I will put the B+ jumper under the board and the ground on top. I wouldnt think that would cause me any issues but you tell me.


Also, with my PPIMV, should I ground it at the power amp side of the grounding scheme since its between the PI and output?

[sluckey]

I have the ground of the PI and the B+ of the PI nd of crossing at an x pattern. I will put the B+ jumper under the board and the ground on top. I wouldnt think that would cause me any issues but you tell me.

No problem.

with my PPIMV, should I ground it at the power amp side of the grounding scheme since its between the PI and output?

I would ground it at the same place you ground the cathodes of the output tubes.

EDIT... I would not install grid stoppers between pins 1 and 2. Several brands of tubes have pin 1 internally connected to pin 2.

[hesamadman]

EDIT... I would not install grid stoppers between pins 1 and 2. Several brands of tubes have pin 1 internally connected to pin 2.

hmm.... im going to check continuity on the tubes I have right now and see what they tell me. If that is the case. then the grid stopper is useless in this application. The signal is bypassing the 1.5k resistor by taking the shorter path from pin 1, if connected to pin 2.

[John]

EDIT... I would not install grid stoppers between pins 1 and 2. Several brands of tubes have pin 1 internally connected to pin 2.

hmm.... im going to check continuity on the tubes I have right now and see what they tell me. If that is the case. then the grid stopper is useless in this application. The signal is bypassing the 1.5k resistor by taking the shorter path from pin 1, if connected to pin 2.

But even if these tubes do not have that internal connection, the next set you (or the next guy) pop in might. And we don't want unhappy musicians, now do we? 

[hesamadman]

But even if these tubes do not have that internal connection, the next set you (or the next guy) pop in might. And we don't want unhappy musicians, now do we? 

Thats a good point. I was curious though if that might have been a possibility as to where some of my noise was coming from. Though its obvious its from grounding and lead dress (lengths of wire also). But no continuity between pin 1 and 2 of my JJ Tesla El84's.


Other than attaching grid stoppers to my board, any suggestions how to place them securely if not on pin 1?

[John]

If you have room (and I'm not familiar at all with the inside of the amp or how it's laid out) you could install a small terminal strip. Or even just solder directly to pin 2 and the other end to the signal wire and heat shrink it.

[sluckey]

Other than attaching grid stoppers to my board, any suggestions how to place them securely if not on pin 1?

I like to use standoff turrets next to the sockets. Some people like to connect the resistor directly to the socket with a 'flying' lead connected to the other end of the resistor (I hate this idea!). Some people like to put the grid stopper between pins 2 and 8. I've never ran across an EL84 that had anything internally connected to pin 8 but that doesn't mean that some one may do so! I suggest leaving unused pins alone on all EL84s just to be safe.

[hesamadman]

Other than attaching grid stoppers to my board, any suggestions how to place them securely if not on pin 1?

I like to use standoff turrets next to the sockets. Some people like to connect the resistor directly to the socket with a 'flying' lead connected to the other end of the resistor (I hate this idea!). Some people like to put the grid stopper between pins 2 and 8. I've never ran across an EL84 that had anything internally connected to pin 8 but that doesn't mean that some one may do so! I suggest leaving unused pins alone on all EL84s just to be safe.

How can I obtain stand off turrets?

[sluckey]

You'll have to buy them somewhere. Take a look. They're not hard to find.

[John]

https://www.surplussales.com/Hardware/Standoffs/dap.html


https://www.surplussales.com/Hardware/Standoffs/threaded-base.html

[hesamadman]

https://www.surplussales.com/Hardware/Standoffs/dap.html


https://www.surplussales.com/Hardware/Standoffs/threaded-base.html

Thanks a ton John!


After "completing" my layout drawing, I realized something and I would like to talk about it. Perhaps this has given me problems and I didn't even know it.


PPIMV Shielded Wire
I have dual conductor shielded cable for my PPIMV. However, I MAY have the wrong wires paired in the same sheathing. Correct me if I am wrong, but I SHOULD have the wires on the middle terminals of dual pots ran to 1.5k grid stopper of EL84 inside the same sheathing and shielded together. I should also have wires on right terminals of dual pots going to preceding CC in same shield. Is this right? Im not sure how my current 30 watt is set up but i am going to check.


Updated layout also:

[hesamadman]

This is not how I have my PPIMV shielding.


Currently the middle terminal and right terminal are together and shielded on one half of the dual pot. The same is done with the other half of the pot.


So basically I have the in and out of the one half twisted together and shielded, and in and the out of the other pot twisted together and shielded. Im not sure. Could this be a problem.


I feel like the 2 inputs from the PI should be together and shielded, and the 2 outs to the grid stopper pin 2 EL84 should be twisted and shielded.


I also see 2meg resistors across the two terminals. Whats the purpose of this? Should I add these?

[John]

looking down at your amp,  the left lug on the pots is where the signal wire comes in from the PI. The right lugs get grounded, preferably to the same point your cathode resistor gets grounded. The middle lugs feed your grid pins on the power tubes.


Ideally, the signal wire feeding your dual pot, and the signal wire leaving your dual pot and going to pins 2 on the power tubes is shielded. The braid on that shielded wire gets grounded at one end only. If you ground both ends of the shield, good potential for ground loop/hum.


Those 2.2M resistors going from the wiper (middle lug) to the right lug (ground lug) are so that no matter what, the input grids of your power tubes have a path to ground in case that wiper in the dual pot should ever lose contact inside. Rarely happens as far as I know, but for .20/resistor....


They are 2.2M so that your dual pot stays at around 250K (tad less because of ohm's law).


Reading your post, I think you have the pots wired wrong. The way the schematic drawing is for those dual pots, it's easy to do. But you want to wire each "half" of the dual pot just like you would any other individual pot. They're exactly the same as any other pot, you just have 1 knob to turn instead of 2.

[John]

Here's a schematic that shows better what I mean, if you have SCH to open it?


*edit* picture attached.  Note, I omitted the 2.2M r's for clarity (just like Merlin  )     (his books are invaluable, btw)

[Willabe]

I feel like the 2 inputs from the PI should be together and shielded, and the 2 outs to the grid stopper pin 2 EL84 should be twisted and shielded.

Technically yes, because then the 2 out of phase signal lines from the PI will cancel any noise.

But I'm not sure that the way you have them wired up now would cause any problems? They still are shielded.   

[hesamadman]

Eh. I'm just trying to cover every angle before I start the new one.

[hesamadman]

But man does this sound fantastic. I threw a 12ay7 in v1 today. Aroused my tone buds.

[Willabe]

But man does this sound fantastic. I threw a 12ay7 in v1 today. Aroused my tone buds.

               

[John]

Hey, I just realized that I did not understand your question, and answered something you didn't ask. Sorry!

[Willabe]

I mis-read questions here all the time and you did explain to him what the 2 2.2M R's are for.   

[DummyLoad]

Other than attaching grid stoppers to my board, any suggestions how to place them securely if not on pin 1?

I like to use standoff turrets next to the sockets. Some people like to connect the resistor directly to the socket with a 'flying' lead connected to the other end of the resistor (I hate this idea!). Some people like to put the grid stopper between pins 2 and 8. I've never ran across an EL84 that had anything internally connected to pin 8 but that doesn't mean that some one may do so! I suggest leaving unused pins alone on all EL84s just to be safe.

How can I obtain stand off turrets?

current production stuff...

http://www.keyelco.com/category.cfm/Terminals/Threaded-Turret-Terminals/p/515/id/778/c_id/941

http://www.keyelco.com/product.cfm/product_id/3279/checkStock/1

http://www.mouser.com/Electromechanical/Hardware/_/N-5g2c?Keyword=11352&FS=True

http://www.mouser.com/ProductDetail/Keystone-Electronics/11351/?qs=%2fha2pyFadujSb4e1CbFZMNs%2fOnramsOVwdUbUoofUKA%3d

--pete