18Watt.com

Dangggg!!! I'm having trouble fixing a buzz/hum in my last build. I tried about everything, but I can't get it out.
It has a LarMar PPIMV and the buzz/hum increases with the MV. Volume or gain pots have no influence, nor do tone pots.
I checked every wire, every solder joint, filter cap, changed grounding wire connections, tube rolling, diodes, omitted the relays system while disconnecting its power supply, played with heater wiring,.. you name it!!
I have a scope but IDK where to begin...
The buzz stops when I ground the PI grids. Without preamp tubes V1 to 4 (preamp + reverb) it still buzzes

Listen:
https://www.dropbox.com/s/rrhz72jb0s85p ... d.m4a?dl=0

Have you tried with the LarMar PPIMV removed from the circuit?

JMPGuitars wrote: ↑

Tue 03/22/22 2:34 pm

Have you tried with the LarMar PPIMV removed from the circuit?

Yes. Replaced it with 2x220k to ground.. no change at all.
Just disconnected all wires from the relays circuit and disconnected its PS again.. no change whatsoever...

Transformer bolts tight? Have you tried a different outlet?

JMPGuitars wrote: ↑

Tue 03/22/22 4:55 pm

Transformer bolts tight? Have you tried a different outlet?

Yes they're tight. I even loosened the transformer and put 1mm rubber washers between it and the chassis to prevent vibrations. I did that because when the power transfomer rested on my workbench I had a rather loud vibration. Maybe a clue this hammond is flakey?
Same problem on every outlet..

But, I might be onto something. Yesterday I disconnected the 50V winding to the tremolo bias thing and the hum reduced dramatically. Instead I got a lot of heavy hiss. This hiss is probably due to the 50V wire hanging in the air (I disconnected it at the diode connection)
Tonight I will go deeper down that route. I'm going to disconnect the tremolo circuit everywhere it's connected and see where that will lead to.
On the layout there is a blue jumper wire from the diode to point "G". When I first started the amp up I got about -70V there so I lowered it to 50V by replacing that jumper by a 100k 2W metal oxide resistor. Maybe I have to reverse that, because at the time of initial start up I had extremely much AC from the outlet (245V .. that's normally 235V)

NOTE: the first time I built this amp the PT didn't have a dedicated 50V tap on the HT, so the neg bias was tapped at the rectifier. I did put the diode and 1M dropping resistor right next to the rectifier socket. So on that amp the 50V wire crossing all my power stage wires was DC loaded. On the current build it's AC right onto the board where it connects to the diode. Maybe I better copy that principle on the current build. The 50V wire runs under the board , so it's hard to see if it has a chance to couple to a DC loaded wire somewhere (or affect OT primaries)

Sorry if I'm guessing too much, but I'm on the end of my rope here...

Here's a link to all pictures:

https://www.dropbox.com/sh/36qip83onofg ... jNy5a?dl=0

I'm starting to think the config of neg bias for the trem might be the culprit...
Things I did wrong IMHO:
- grounded the cap C35 for the bias at the preamp ground bus. Should be at the power amp star ground I think.
- running that 50VAC wire across the 6V6, OT and PI wires is bad practice, even if it crosses straight angled
- the grounding of the tremolo intensity pot lug 3 is joined with the grounding of the speed pot lug 3 and ran to the preamp ground bus. While it worked perfectly for the former build I think I'd better follow the circuit of my 6G3, being, speed pot to preamp bus and intensity to the neg side of the bias cap at the conjunction of the cap and diode. The 6G3 system works perfectly.

Ok! It's solved. Moved all the ground connections of the tremolo circuit to 1 small buss, and that buss to power amp star ground . Also moved the diode to the transformer area and put a shielded wire from there to the board.
Now I have to address some pretty loud hiss when volume and MV is turned up. But that responds to the volume and reverb. Cathode bias is set at 82% dissipation, so that's just fine.
One other odd thing it still has is the volume drop when trem is engaged. Have to look into that too...
At least that darn hum is gone

A part of me wonders if you still have some grounding issues. You show that you have the input grounded directly at the input. This is good practice, but only if that becomes the ONLY grounding point in the amp for the whole of the amp. The power amp section can then be grounded elsewhere, so long as it contains ALL the power section within it. You have nodes split up so that they are set next to their respective Tubes. This means that you have B+ grounding in more than 1 place, at the input, and at the center tap. It would now have a reference to ground at the input and the output section's grounding points. This is why cap cans are generally loathed because they share a common ground among what are technically different nodes. This adds noise. If your preamp B+ has a different ground path than your power tubes B+, you are crossing the streams so to speak. The ultimate objective is that there is ONLY 1 grounding location in the amp, or you make sure all your grounding paths are locally grounded to points in the amp that are descending in path length to the safety ground in the amp.

Excessive hiss and noise can be due to lead dress. You have A LOT going on in a small chassis. The wire runs are relatively short, but as a result, you also have lots of runs crossing over each other. Try moving wires around or getting some wires up and over the layout to see if that helps. Also your schematic/layout and your build look completely different. The schematic shows a center tapped heater circuit, but you are using elevated heaters in your build ( I can't tell if your PT is center tapped for the heaters ? ) and you should not use both, only one or the other. If your PT is center tapped for the heaters, you can't also use elevated heaters.

Also, it looks like your G2 ( screen grid ) is coming directly off of the B+ that feeds your Pi plates. This should have its own node, usually taken off the second #B dropping node. The Pi should be on your third #C node, which means you're sharing your G2 with the same thing that is feeding the plates of your Pi. The screen grid should have its own dropping resistor and decoupling cap and or run directly after your choke. In your case, you are using B1, B2, B3 etc. except you skip B2 which is where your choke returns at. This is where the G2 is generally pulled from, you have it coming off of B3. I would still use a screen grid resistor of 1k for each tube, inline from where your choke returns, to the G2 grids.

Objectively, you are trying to keep each section of the amp separated from the other, that is the idea behind the decoupling caps ( the B+ nodes ). It decouples the preceding stage from the last one reducing noise progressively as you go from the power amp to the first stage of the preamp. In your case, you have multiple paths where loops occur between the power section and the input section. Especially around your Pi with the G2 grid tied directly to the Pi's plates! Your grounding scheme for the inputs and outputs would work if ALL your nodes were referenced to only one point. In your case, you have the output and the preamp referenced at two different points. This carries current via B+ to those points adding noise. Another way to better this is to have multiple star points where each node and its respective circuitry is grounded directly to the chassis in a way that they descend in path length to the safety ground. Imagine now that you have your Pi node tied directly to your G2, you have a path to ground in two directions from that point. One through the power amp section, and one through the preamp section, all via your ground plane. Think of it in terms of current paths. If the current generated can go only one way, there is no loop. If there are two ways that the current can get to ground, then there is a loop. This is what the decoupling nodes do, is close loops when they are connected to ground. But if the nodes are all tied to a single grounding path ( wire ), there is only ONE way that the current can get to ground. So if a piece of circuitry touches TWO circuits and each has a different path to ground, there then becomes a loop. The B+ is a circuit, and you technically have two paths to ground for it. One in the power section where you ground all your power amp circuitry, and one from your input that your other B+ nodes are in.

As you can read above the problem was solved. I have built this amp 3 times with no problems. It was some interaction on the grounding part of the tremolo.. but I fixed it.
G2 has 2x 1k by the way..

Thx

As a general rule, buzz is power supply noise and ground loops cause a loud sine wave humm
Finally, I'm sure everyone around here knows that in my honest opinion, PPIMVs don't belong in 18W amps!