Blues Deluxe to ODS #183 Thread

[pompeiisneaks]

That sounds logical. However, how would you test if 1 interacts with 3 or 2 interacts with 4?

First the short answer: By opening up all the controls to 10, with nothing plugged into the guitar input, and observing the PI input with a scope for signs of oscillation. When nothing is plugged in, the first stages grid is held at ground, so it is not possible for that stage to oscillate. If you see oscillation under these conditions, the oscillation is occurring between the second and fourth stages.

I actually think the question has already been answered, but I am having a difficult time believing it. [Apparently, I am from the "show me" state, and never knew it]. My gut feeling on this is the oscillation is occurring between the second and fourth stages. This is based on my own amp-building experience. If this were the case, then grounding the grid of the first stage, by not plugging a cable into the guitar input, would result in the amp still breaking into oscillation - and this is not what I have experienced in my own builds. Phil says there is no oscillation with nothing plugged in. I had asked him to reverify this (because I have never seen it happen this way) by opening all the controls up to 10 and looking for oscillation at the input to the PI with nothing plugged into the input. I was just thinking perhaps there was an oscillation with nothing plugged in, and Phil hadn't noticed it yet. But again, that's just based on my own experience. Also from experience, these oscillations occur way above the range of human hearing. But on the other hand, other strange noises (like sudden excessive hiss or 120Hz hum) as the volume control is increased to the point of oscillation usually show up that will imply an oscillation.

I think I can answer that with empirical evidence.

I can literally put a dmm probe or chopstick between the standoff where the grid input from the guitar and the grid resistor is. this is V1B triode in your board, and it's right next to the OD tube in this case. when I do the sound of the osillation changes literally when I touch the standoff, or when I block it with almost anything. It almost disappears when I put the DMM probe in front of it. It' seems very clearly coupling between V1B and V2A, (on your board, basically the N/C point is where it is supposed to be, but obviously that says N/C because you connect from input directly to the grid of V1B. (pin 7). Pins 6 and 8 are also connected to what you have marked as V1.6 and V1.8 appropriately. Those are what is coupling with V2.1/V2.2/V2.3 right adjacent to it.

To me in a nutshell this says that my problem is potentially one or both of the above mentioned issues, either proximity is just too close from grid input and the OD tube and pdf64's suggestion to move it would resolve it, or the swapping of the leads from the V1B to V1A and vice versa might as well, as the recovery stage of V1 is less prone to noise than the first stage. I'll test just swapping the leads first and see what I get. If that doesn't remove it, I'll also add a physical grounded barrier of steel as well, to just add that extra protection.

~Phil

[JazzGuitarGimp]

This is really awesome! Thanks to everyone who pitched in with suggestions; especially pdf! You guys are the best.

[pompeiisneaks]

Yes this fixed it.

I just resoldered the 123 and 678 pins on V1 vice versa, and the oscillation is gone! It may be wise in the docs to make that clear that either or both should happen:

1. make sure the V1 tube has good separation from V2
2. the gainier the amp is going to be, the farther away you want the first grid connection, so leave that over on the V1a pins 1,2,3 instead of V1b pins 6,7,8.

Thanks!

I'll have a video posted on youtube in the next day or two showing this all off.

Yes thanks to everyone that helped, especially pdf64! my hero

By the way I never discounted what you were saying, I just didn't fully understand the impact/import and was trying everything I could to NOT have to drill a new hole in the chassis. I did that, but did what you needed anyway.

~Phil

[pompeiisneaks]

A new video showing off the fix:




~Phil

[norburybrook]

yay I love a good detective story and it took a Brit for the rescue well done @pdf64

Phil you can sleep now with a clear mind , I Know what it's like having a problem with a build.

@Lou, I hear you about schematics/layout , I was one of those people who started building from layouts, it was my way in as I'm a visual learner and had scant electrical knowledge. I can now make sense of a schematic but still have trouble picturing actual physical layout of components and as we can see from this thread, actual physical layout is very important.

enjoy your amp Phil, threads like these are great for the community.


M

[pdf64]

Yippee, that’s great news!
There’s always some degree of compromise involved with layout, and so local internal screening plates are a very handy means of reducing unwanted coupling / improving an amplifier’s margin of stability. I use them as a default if something is obviously ‘non ideal’, eg OT mounted at input end of amp chassis. Maybe the circuit would be unconditionally stable without it, but can’t rule out the possibility that if I chanced my luck, then eg a particular set of tubes might tip it over the edge.

[martin manning]

Glad you got it sorted, Phil.

I don't think it's the proximity of the tube sockets. That looks pretty typical, and I've seen them spaced much more closely. I believe the problem is just lead dress. Virtually every amp layout uses the outside pins of the first tube for the input stage, and the following stages are arranged so that adjacent leads are out of phase. Dumble layouts are no exception. I guess this problem occurred because you followed the printing on the board when hooking up the leads, with the sockets rotated 180 degrees from the assumed orientation. That would put the Clean 1 leads close to the OD2 leads.

I would also like to see the filament leads run on the other side of the sockets to get them away from the leads going to the board.

[pompeiisneaks]

Glad you got it sorted, Phil.

I don't think it's the proximity of the tube sockets. That looks pretty typical, and I've seen them spaced much more closely. I believe the problem is just lead dress. Virtually every amp layout uses the outside pins of the first tube for the input stage, and the following stages are arranged so that adjacent leads are out of phase. Dumble layouts are no exception. I guess this problem occurred because you followed the printing on the board when hooking up the leads, with the sockets rotated 180 degrees from the assumed orientation. That would put the Clean 1 leads close to the OD2 leads.

I would also like to see the filament leads run on the other side of the sockets to get them away from the leads going to the board.

Does that mean I put the sockets in 'upside down' from convention then? Or are you more meaning that the a/b pinout sides are less important in general than putting the first stage on the side farthest away from all other stages as a point? I know i've seen a lot of amps that used pins 6,7,8/V1b as the first stage in the drawings and I wasn't exactly sure why... maybe because of how they also installed the sockets?

As for the sockets, you mean having them go up towards the open side of the chassis instead of down to the bottom edge? The way I saw it, is that the leads basically come in from the top and don't get close to them no matter where they are. I also did my last build with straight wires, no twisting and directly away from the tubes at an almost right angle from the incoming leads and had 0 noise issues. I'm starting to believe that for guitar amps, anyway, that heater hum is rarely a problem unless you're horrible with lead dress, but maybe I've been lucky.

~Phil

[martin manning]

Look at the 183 layout. The preamp stages are arranged in order from right to left. Since each stage is inverting, adjacent leads going to the sockets alternate in phase. By flipping the socket and following the pin numbers as printed on the board you brought the first stage leads closer to the OD, and placing the turret out there made it worse. I can’t tell which stage on the second tube was next to the input stage before you swapped the leads. Is that OD1 or OD2?

The 678 triode may be less susceptible to heater hum, so it is often used for the first stage.

[pompeiisneaks]

The 678 triode may be less susceptible to heater hum, so it is often used for the first stage.

I fully understood the first part for sure, it's more that second part you stated about the less susceptible to hum. I didn't realize that, but now that I think about it, it makes sense, 1,2,3 have only one side that has a heater connection, leaving it a bit 'unbalanced', but the 6,7,8 side has each phase on either end, which should likely do some minor cancelling. I think anyway .

Yeah I definitely didn't think about the positioning of the first input and that OD tube for sure, and I won't make that mistake again!

~Phil

[xtian]

I oriented my tube sockets to match Lou's PCB layout, to keep wiring neat. Again, no issues with oscillation over three builds.

IMG_3342.jpg

[pompeiisneaks]

Yeah it looks like to me, though, that you probably have at least another 1/2" or more of distance between V1 and V2 than I had. That distance is pretty important, just swapping sides of the tube solved mine, so that's less than 1/2" I think.

~Phil

[rootz]

Not just the extra distance as far as I can see. Shorter wires (really neat lead dress), sockets placement matches with their components on the pcb. I made a Bogner extacy blue channel in a Marshall Jmp chassis some time ago. I had the second gain stage on v1a too without any problems. Lead dress was tight, sockets were placed differently.

My latest Dumble like build has the preamp sockets close together, but no oscillations that I know of.

IMHO this oscillations look like a combination of at least three factors and luckily you solved it with the simplest solution. Congrats on a successful build!

[martin manning]

Blues Deluxe spacing looks pretty reasonable IMO.

[pompeiisneaks]

I guess looking at that, I do see now especially 2 things:

1. mine, due to having the standoff, is maybe 1.5 circles away from the other tube instead of almost 2 full circles like xtians
2. mine, the OD wires go directly beside the tube as well, instead of to the rigth of it like xtians. So that's more telling evidence that the longer runs from the board to the tube socket and the tube socket being too close to the wires on the board from the OD section are a recipe for disaster. I'll have to watch that a lot closer in the future. I still would think, as a general rule, if you're going to setup something like this, doesn't it make more sense to always have the grid of the first stage on the farthest side from the next gain stage though? If you're looking for that 6,7,8 pin setup for lower hum, then rotating the tube 180deg as well would get you the best of both worlds no?

Things I'll never make a mistake on in the future for sure

~Phil