Reverb Hum-How to get rid of it

[Willabe]

Yea, I think at least part of my problem is that Node E is carrying 258VDC and it is grounded on the low voltage ground buss on the circuit board, which then connects at only 1 point to the GND BUSS..........   Does this make sense to do?

No.

And it's not the voltage, it's the current.

You have 2 busses, that's not good either. And the ground connections to the busses are supposed to be laid out, gathered together in order, organized with the other grounds from that same circuit. Not randomly connected.

You don't just daisy chain all the B+ filter caps together then connect the buss ground to the filter caps daisy chain.

Look at how Merlin did the ground buss in his pictures that I posted.

And read his grounding link.

[wsscott]

Gotcha.

[wsscott]

Okay, so I've read the article by Merlin, and that is helpful.

I've gone back to my Layout and the Circuit Board, and identified the Filter cap Nodes for the connections on the circuit board's "Buss" grid.  So I will rewire them so that they are connected to their appropriate Filter caps.

I will also remove the daisy chain wiring between the filter caps, and have them all go directly to a single star point, which I will connect to the Buss Bar at its far left end.

Then there are the connections to the Buss Bar from Tab 1 on some of the pots, the Input and Output jacks, the Vibrato footswitch, and the Reverb Input and Output jacks.  Should all of these connections be removed from the Buss Bar, and wired to another star point, maybe using a terminal strip, and then wire that directly to the Buss Bar?  Or should their Grounds also go back to the Filter Nodes that they relate to?

[tubeswell]

Grounds go back to the Filter Nodes that they relate to

^This^


If you’re doing a signal star ground it’s got to have the loop blocker

[wsscott]

I guess I was planning on doing what is called the multiple star grounding in Section 15.8 of the article.  I guess under that method, I could leave the 5 filter caps still daisy chained together, but run a single wire to each filter cap node from the group of ground wires coming off the signal circuit associated with each node.  Those 5 nodes are daisy chained and connect to the Ground BUSS at only one point.  And I wouldn't need to do the loop blocker.  Does this sound correct?

If so, how about the other grounds that I mentioned that are currently grounded to the BUSS, ie. as mentioned below, the pots, jacks, foot switches?  Do they need to also connect to the stars for each of the signal nodes described above, or can they stay where they are?

[Willabe]

I will also remove the daisy chain wiring between the filter caps, and have them all go directly to a single star point, which I will connect to the Buss Bar at its far left end.

I wouldn't do that. I would do it like Merlin and K. O'Connor and do a wired galactic ground, ie, multiple ground stars daisy chained together and grounded to the chassis at only 1 point.

Bring all the grounds together for all the tube sections being fed by a filter cap node. Then run a wire from that filter caps ground lead to those gathered grounds. That's a ground star. Then you run a wire (acting as a buss) from that ground star to the next ground star. And somewhere along that chain you run a wire for the chain to the chassis for the chassis ground.

You could have put each filter cap right next to it's star ground on the buss, look at Merlin's multi star buss drawing. It's really very easy to do.

And the chained/wired buss on the board that goes to the other buss, those board grounds need to go to the tube ground star they belong too.

And all the isolated jacks, including the ft sw jacks need to have their ground wire go to the ground star they belong too.

You never answered me if the verb ft sw jack is isolated or grounded to the chassis where it's mounted?     

[Willabe]

And I wouldn't need to do the loop blocker.  Does this sound correct?

You still need that loop blocker. Because it's 1 unit plugged into a 2nd unit. They both have chassis grounds with a 3 prong power cord. And the guitar cord plugging the preamp/vib/verb to the amp it's ground makes the loop complete.
Now if you get your grounding right, then use a 2 prong adapter, you might not need the loop blocker. 

If so, how about the other grounds that I mentioned that are currently grounded to the BUSS, ie. as mentioned below, the pots, jacks, foot switches?  Do they need to also connect to the stars for each of the signal nodes described above, or can they stay where they are?

YES. They ALL need to be moved to their ground star.

Tubeswell's grounding worked for him, and it will work for you if you do it as he did it.

I just don't like a single star ground. Just different thinking. That's a lot of wires running across the chassis to 1 ground star point. 

[wsscott]

Thanks.

Sorry, yes the foot switches are isolated too using isolation washers.  The Reverb footswitch is grounded to the Reverb Intensity Pot, Tab 1, the far left tab on that pot looking from behind.  That pot is grounded via its connection to the chassis.

[tubeswell]

The reason for keeping the signal ground in a separate wire (as Willabe notes) is to only reference this to the chassis at 1 point (via the loop blocker).


We do this at the input end of the signal ground buss because that is the most noise-sensitive gain stage, and we want the filter cap for this stage to be riding as close as possible to the chassis ground reference point so that any micro voltage changes that might be in the buss which emanate from higher-current parts of the ground return (or even from radio wave interference) result in the least amount of interference with the most noise-sensitive stages in the circuit. (Think of the signal ground return as carrying all the return currents - and changes in current).


If you have multiple chassis contact points, you create multiple loops which exponentially increase loop current feedback/hum. In a standard guitar amp, this usually isn't much to complain about, but because this is a separately powered effects unit in the signal chain, which has to have it's own mains safety ground chassis connection, you have to work harder to eliminate sources of ground loop feedback.

[pullshocks]

Noting that you report that the hum occurs when the reverb is activated (but isn't present on the dry signal) it seems that the problem is grounding of the signal returns for the (highly noise-sensitive) reverb recovery stage is getting muddled/overwhelmed by higher current returns. As this is a stand alone effects unit that is intended to be plugged in in front of a guitar amp, you might want to consider regrouping the ground returns and installing a signal ground hum-loop blocker between the ground buss and the chassis of this stand alone unit. This should significantly quieten/dampen loop hum.


Here's an example of how I wire 6G15 ground returns to eliminate loop hum - I can vouch that this yields a quiet noise floor. The same principle of galactic grounding around filter cap nodes with a daisy-chained buss to a hum-loop blocker can be applied to your unit - but it may require a bit of re-work.


Or maybe you're happy to live with it

Tubeswell, got any gut shots of the reverb unit with that wiring?  Really curious how you solder up that many ground wires to a single point....

[wsscott]

I was going to start working on the grounding changes today, but wanted to clarify something.

If you look at my Layout, what I understand I should do is to disconnect all of the components attached to the "ground bar" at the top edge of the circuit board, and wire them to individual "star points", which would then be wired to their corresponding filter cap power source.  These components that I would be disconnecting are associated with Nodes B, C, D, and E.  Are these the only "signal grounds" that I need to deal with?

Or do I only need to deal with the grounds that are shown in the Reverb Driver/Recovery section of the Schematic starting with the Ground at the Tank's Input jack and going through the Ground on the V7B 2.7K Cathode resistor?

Since the unit has no hum when the Vibrato is ON, or when the Vibrato and Reverb are OFF, it seems like only the Grounds in just this section need some changes.

Is there a way I can use my Oscilloscope to identify where the hum originates?

Thanks for your patience.

There are many more grounds shown on the Schematic, but I'm assuming that this is what is being suggested I do.

[wsscott]

I connected my scope to the output jack, with no input signal, AC coupling.  It showed a slightly fuzzy sine wave of about 10hz with about 16mVRMS.  The scope was set to 50mV/div which is the smallest it will go.

I was expecting a 60hz signal.

[shooter]

Is there a way I can use my Oscilloscope to identify where the hum originates?

yes n no


If the source is a "stage" along the signal path, ya;   signal good, signal good, oh, that's ugly...
If the "source" is the PS, either the B+ side or ground side, then NO, it will be "riding" on that "line" everywhere you look.


the important part you should "see"; What is the frequency and amplitude, and which "line" is it most prevalent... ground buss, B+ buss

[shooter]

It showed a slightly fuzzy sine wave of about 10hz with about 16mVRMS.

the frequency is within the "tremolo range"
monitor the 10Hz, turn trem on/off...   any change??

[wsscott]

I just realized I didn't have the foot switches plugged in.  The tank is not connected. Now the foot switches are connected.

When I turn On the reverb I get a 120hz 6mVRMS sine wave.  When I turn it off, I get a straight line.  I connected the tank and get the same readings.

When I turn On the vibrato I get a much larger signal, that moves.  So I think that's normal.  Its 4hz at 101mV.

[shooter]

I turn On the reverb I get a 120hz 6mVRMS sine wave.

When I turn it off, I get a straight line.

now move scope to the REV driver (before tank) and repeat.

[wsscott]

Do I connect the probe at the Reverb Input Jack or on the V6's 1uf coupling cap negative side, or where?  Thanks.

[wsscott]

I connected the tank, and then connected the probe to the Neg. side of the 1uf cap that comes off the V6 plates that are connected together.  With the reverb ON the signal is fuzzy and jumping all around but it looks like it's closer to 60hz or less, but those numbers are continually changing.

[wsscott]

I'm probably not doing something correctly with the scope.

[shooter]

VERIFY your scope probe is rated 350VDC or more.
IF so;
probe each of the PS TAPS B-D measuring VAC (do this with BOTH trem n VERB OFF) record values
Repeat with VERB ON
post

[wsscott]

Shooter--My probe is rated up to 200VDC in 1X mode and up to 600VDC in 10X mode.  So I'll set the probe and the scope to 10X.

I'll set the scope to AC coupling.

Then connect the GND lead to chassis.

Then use the probe to check AC voltage at the + side of the Nodal caps, B, C & D.  Correct?  When you say "PS Taps" I assume that's what you mean?  I'm not experienced with the scope at all, so that's why I'm asking all these questions.

[Willabe]

..... what I understand I should do is to disconnect all of the components attached to the "ground bar" at the top edge of the circuit board, and wire them to individual "star points", which would then be wired to their corresponding filter cap power source.  These components that I would be disconnecting are associated with Nodes B, C, D, and E.  Are these the only "signal grounds" that I need to deal with?

Yes, disconnect all the components attached to the (wired) ground buss at the top of the board.

And disconnect all the components attached to the ground buss above all the pots.

Then wire ALL the components grounds to their own individual ground star on the  buss either above the pots, or on the board. Just 1 buss, not 2. Gather the stars ground leads close together on the buss, right next to each other, touching. Look at the drawing attached, 15.10 b. All grounds in the star brought very close together then buss wire separating that star from the next ground star. 

Then disconnect the ground's on the filter caps from each other. They should not be daisy chained together. That would cause a ground loop together with the ground buss. 

Then run a single wire from each ground star over to it's own corresponding filter cap ground lead. Should be 4 wires total, nodes B,C,D,E. 

Leave node A as it is, with the PT's CT going right to node A's caps ground lead. Then run a single wire from node A's ground lead/PT CT over to the very end of the ground buss closest to the filter caps. Make sure there's some distance (wire) between that ground connection to the buss and the ground star closest to it. Do not connect it to any ground star.

Leave the ground buss connected to the chassis at the input jack. That will be the only chassis ground. (Except for the power cords 3rd safety wire ground. That stays grounded at the other end of the chassis.)

I'd try it like this and use a 3 prong to 2 prong adapter and see it the humm is gone. If not then I'd add the hum loop blocker tubeswell showed.

Or, just go ahead and put the hum loop blocker in now.   

Or do I only need to deal with the grounds that are shown in the Reverb Driver/Recovery section ....   Since the unit has no hum when the Vibrato is ON, or when the Vibrato and Reverb are OFF, it seems like only the Grounds in just this section need some changes.

Ok, but how will you do that? The grounds are all tangled up with 2 busses and the daisy chained B+ filter caps. You'd have to figure how to untangle the verb section and add it back in to the ground buss somehow.

I'd just redo the whole thing. To me it will be easier.

[Willabe]

VERIFY your scope probe is rated 350VDC or more.
IF so;
probe each of the PS TAPS B-D measuring VAC (do this with BOTH trem n VERB OFF) record values
Repeat with VERB ON
post

He has an E node too.

[wsscott]

Willabe-Thanks for your detailed explanation, and the time it took for you. 

[shooter]

oops, E also

[wsscott]

Shooter-When you have a chance could you answer my question about the scope setup in #70 above?  Thanks.

[shooter]

 
when I "liked" it that was affirmation, add TAP E, it was early

[wsscott]

Thanks.

[wsscott]

Shooter, all the Nodes read 2.04Vrms when reverb is On and Off.

[shooter]

60hz, 120hz, some other frequency?


sorry didn't specify that earlier.


once you have that;
re-check with your DVM, AC volts(milli-volts)


THEN;
move the NEG probe to each "ground" point for every tube, you can leave the RED + lead on any TAP, what i'm looking for is a "bump" in the otherwise "constant" readings

[wsscott]

It was showing 0Hz on all of them.  What should the scope’s settings be? Mine doesn’t go below 50mv/div.

[shooter]

move on to the DVM test for now.

[wsscott]

Ok, for the DVM test, I connected the red lead to Node D+, set the meter to VDC not VAC, and then used the black lead to measure each connection at the Ground BUSS.  They all read the same 282VDC.

The Cathodes of the tubes get to the ground BUSS through some component, ie. their cathode resistors.  None go directly from their cathode to the BUSS.

Everything is grounded at the BUSS, except for maybe a pot that is grounded to the chassis through its direct connection to the chassis, and the Green Mains ground which has its own separate ground to chassis.

[Willabe]

Everything is grounded at the BUSS, except for maybe a pot that is grounded to the chassis through its direct connection to the chassis, .....

Which pot?

If it's for the verb output, that could be the problem.

[shooter]

They all read the same 282VDC.

DVM, AC volts(milli-volts)

testing a ground noise problem is "finding" a point where "resistance" is higher than the rest of the grounds


ways to do that;
ohming chassis to each stages ground point, sketchy if you're trying to measure milli-ohms with a typical meter
measuring the "AC" creeping along the ground path.  also sketchy with a typical meter but yields better results than measuring ohms.

[wsscott]

Wallabe-The Output Level pot has its Tab 1, far left tab looking at rear of pot, connected to the BUSS.  Of course the pot is screwed to the chassis.  Would that be the cause of the HUM?  How else would that tab go to Ground as shown in the Schematic?

The Reverb Intensity pot has its Tab 1 connected to the Reverb F/S Jack's ground tab, and that Jack is isolated from the chassis.

[wsscott]

Shooter, when I had the DM in AC millivolts setting I was getting OL, so that's why I switched over to VDC thinking that's what you meant since I couldn't get a reading otherwise.

Looks like I'm running up on a dead end with the equipment I've got. I don't think my scope is sensitive enough for the job either.

[Willabe]

Wallabe-The Output Level pot has its Tab 1, far left tab looking at rear of pot, connected to the BUSS.  Of course the pot is screwed to the chassis.  Would that be the cause of the HUM?

Wallabe?   


All 3 solder tabs on a pot are isolated from the pots metal back cover. That cover is just for shielding of the pots internal resistive track.

So no that wont cause humm, it helps stops any humm. 

[wsscott]

WILLABE!!  Yes, that's what I thought. I did dis-connect it though just to see if the hum was affected and it wasn't.

I got a Spectrum Analyzer App to check the frequency and it 120hz.  There's about a 4db difference between the levels when the reverb is Off vs. On.

At least now I can tell if its 120 or 60hz!!

[wsscott]

Willabe-I'm going to work on your suggested re-wiring this weekend and need some clarification.

Node B has 3 components; C has 5; D has 1; and E has 3.  So that's 12 wires to connect to 4 filter caps.  That's a lot of wires to be added to the chassis, and I'm trying to figure out the neatest way to do it.

I've thought of adding a terminal strip to an isolated mounting screw at the center, top edge, of the circuit board.  I would use 4 of its terminals.  Then 1 wire from each of those 4 filter caps would be connected to each terminal.  So those 4 wires would be coming from left to the center of the chassis where the terminal board is.  Easy.

Now the hard part is getting the wires neatly from the 4 sets (B,C,D, & E) of components to the terminal strip. Do I have to run separate wires from each component to its corresponding terminal, or can each group of components be soldered together and then 1 wire run from each group to the terminal? If they can be soldered together, can that be on the Ground BUSS above the circuit board?  I can't just have them dangling. Or is there some better way to do it?

Then Filter Cap A remains an issue.  The CT for the PT Secondary is grounded and mounted on the same screw as the Mains Safety Ground wire.  It is located under the circuit board.  It will be a real PIA to take out the circuit board to get access to that CT wire. (I didn't realize at the time I installed it that it would be covered by the board).  Can I just wire that cap directly to the Ground BUSS wire?

Finally, the C Node has a ground from Tab 1 on the 1M Intensity Pot and a ground from the Reverb F/S, as well as 3 other ground connections.  That F/S is isolated from the chassis.  Can the F/S ground lug be wired to Tab 1 on the Pot, and then a wire run from Tab 1 to the other components for that C Node?

Hopefully this will be all I need.  Thanks again.

[Willabe]

This is all still pretty much the same question. Answer;

Use a ground buss with multiple ground stars on it.   

We have shown you several pictures how to do this from Merlin. (See below at the bottom.)

Node B has 3 components; C has 5; D has 1; and E has 3.  So that's 12 wires to connect to 4 filter caps.  That's a lot of wires to be added to the chassis, and I'm trying to figure out the neatest way to do it.

I've thought of adding a terminal strip to an isolated mounting screw at the center, top edge, of the circuit board.  I would use 4 of its terminals.  Then 1 wire from each of those 4 filter caps would be connected to each terminal.  So those 4 wires would be coming from left to the center of the chassis where the terminal board is. Easy.

No, not easy. Use a ground buss, not a terminal strip.

Looking at the schematic drawing in reply #6, your #'s are way off, low.

Do I have to run separate wires from each component to its corresponding terminal, or can each group of components be soldered together and then 1 wire run from each group to the terminal? If they can be soldered together, can that be on the Ground BUSS above the circuit board?  I can't just have them dangling. Or is there some better way to do it?

I'd use the ground buss on the back edge of the board. The buss stars cuts down on wires and their length.

Then Filter Cap A remains an issue.  The CT for the PT Secondary is grounded and mounted on the same screw as the Mains Safety Ground wire.  It is located under the circuit board.  It will be a real PIA to take out the circuit board to get access to that CT wire. (I didn't realize at the time I installed it that it would be covered by the board).  Can I just wire that cap directly to the Ground BUSS wire?

Disconect the A node filter cap ground from the chassis ground. You should be able to do that from the circuit board and not from the chassis below. And the power cord safety green wire should be bolted to the chassis by itself, nothing else should be grounded with it.

Finally, the C Node has a ground from Tab 1 on the 1M Intensity Pot and a ground from the Reverb F/S, as well as 3 other ground connections.  That F/S is isolated from the chassis.  Can the F/S ground lug be wired to Tab 1 on the Pot, and then a wire run from Tab 1 to the other components for that C Node?

Yes, that should work.

Yes, disconnect all the components attached to the (wired) ground buss at the top of the board.

And disconnect all the components attached to the ground buss above all the pots.

Then wire ALL the components grounds to their own individual ground star on the  buss either above the pots, or on the board. Just 1 buss, not 2. Gather the stars ground leads close together on the buss, right next to each other, touching. Look at the drawing attached, 15.10 b. All grounds in the star brought very close together then buss wire separating that star from the next ground star. 

Then disconnect the ground's on the filter caps from each other. They should not be daisy chained together. That would cause a ground loop together with the ground buss. 

Then run a single wire from each ground star over to it's own corresponding filter cap ground lead. Should be 4 wires total, nodes B,C,D,E. 

Leave node A as it is, with the PT's CT going right to node A's caps ground lead. Then run a single wire from node A's ground lead/PT CT over to the very end of the ground buss closest to the filter caps. Make sure there's some distance (wire) between that ground connection to the buss and the ground star closest to it. Do not connect it to any ground star.

Leave the ground buss connected to the chassis at the input jack. That will be the only chassis ground. (Except for the power cords 3rd safety wire ground. That stays grounded at the other end of the chassis.)

[Willabe]

When you do a layout from scratch, you make it with the grounding in mind. The grounding should not be an after thought. 

[wsscott]

Thanks for the clarification.  Helpful.

[wsscott]

Willabe-I started out with re-wiring the Green Power Ground and the R/Yellow Center tap which were previously sharing the same connection to chassis.  Now, the Green Power Ground is connected by itself to a lug on the chassis.  Then as you suggested, I wired the R/Y Center tap to the Node A Filter Cap's Ground lead, and then connected that lead to the Ground BUSS at the end by the other filter caps.

I previously had connected the ground leads of the two 100 ohm resistors for the filament 6.3VAC wiring artificial center tap to the A Cap's ground lead.  Where should those ground leads for the resistors be connected now?

[Willabe]

You can put the ground leads of those two 100 ohm resistors about a 1" or 2" further away/apart from the R/Y Center tap to the Node A Filter Cap's Ground lead on the ground buss. Just a little distance away so the caps charging current can cycle without disturbing anything else.

[wsscott]

Willabe-I'm finishing up the ground stars.  I need some clarification on your comments at #71 above.  I understand to disconnect the daisy chain between the filter caps.  Am I correct that the ground stars don't connect to the Ground Buss but they connect only to the respective Filter Caps' ground leads, and that only the Filter caps' ground leads connect to the BUSS?

[Willabe]

Am I correct that the ground stars don't connect to the Ground Buss but they connect only to the respective Filter Caps' ground leads, and that only the Filter caps' ground leads connect to the BUSS?

No.          You have it backwards. 

What do you see in the ground buss drawing from Merlin?

You use a buss because it runs the length of the chassis, passing by ALL the different circuits and their ground leads. You form the ground stars on the buss.

You gather together ALL the ground leads from a single B+ nodes circuit that it feeds. They go to the buss, gathered tightly together. Then you run a wire from the B+ nodes filter caps ground lead over to the buss ground star for that nodes cap. 

I've written this out several times times for you, it's not hard to do or understand.

Look at the drawing below, the bottom example, b. Wire it just like that. Just imagine that the filter cap is further away, but it's the same thing, just a longer filter cap ground lead wire.

There's 2 separate ground stars in that drawing in the center, see the 2 blue filter caps in the center? Their both on the buss.   

[Willabe]

Here this drawing from Merlin, that tubeswell posted way back, might be clearer for you. It even has ground stars labeled on it. Says local star.

The ground stars are connected by what's labeled as daisy chain, but a buss is pretty much the same thing, both do the same thing. 

[Willabe]

I need some clarification on your comments at #71 above.  I understand to disconnect the daisy chain between the filter caps.  Am I correct that the ground stars don't connect to the Ground Buss but they connect only to the respective Filter Caps' ground leads, and that only the Filter caps' ground leads connect to the BUSS?

It seems pretty clear;

Yes, disconnect all the components attached to the (wired) ground buss at the top of the board.

And disconnect all the components attached to the ground buss above all the pots.

Then wire ALL the components grounds to their own individual ground star on the buss either above the pots, or on the board. Just 1 buss, not 2. Gather the stars ground leads close together on the buss, right next to each other, touching. Look at the drawing attached, 15.10 b. All grounds in the star brought very close together then buss wire separating that star from the next ground star.

Then disconnect the ground's on the filter caps from each other. They should not be daisy chained together. That would cause a ground loop together with the ground buss. 

Then run a single wire from each ground star over to it's own corresponding filter cap ground lead. Should be 4 wires total, nodes B,C,D,E.

[wsscott]

It seemed to me, apparently incorrectly however, that if you connect Filter Cap ground node A, to the Node A star ground which is attached to the BUSS, that when all the Filter caps get connected on their separate stars which are all wired on the same BUSS, that the grounds would all be picking up all of the nodes of the star points since everything is connected to the same BUSS.  And that's what I assumed was causing the ground loops.  Sort of like if you have 4 shots of liquor: bourbon, gin, vodka, and scotch, in separate shot glasses (the star ground points), and then pour them all into a big glass of tequila (the BUSS). You've then got everything blended together, and it will be bad (ie. Ground loop). 

But I'll stop thinking about the "logical" theory of this methodology, and just do it.  Maybe after I have a bourbon--by itself!!  Ha Ha.  Thanks.