Funny grounding issue

[Deepdark]

Not really funny, but quiet weird loool. I made a LA2A recently with Cinemag transformers. The issue is that: When cranking a 12 o'clock and more (I know it's not supposed to be cranked that much), there is a really high rise in Hiss and RF noise. So, I suspected some shielding trouble. I tested the continuity of all my ground connection and shield (all my wire shield are connected to teh chassis). The only place where I got trouble is around input and output transformer. If I test all wire shield, mounting standoff, chassis panel, to the main bolt, I got good continuity (it bips). But, on the input transformer can or the output transformer channel frame, I got a higher resistance reading (like 0.5ohm) and it does not bips. If i test the continuity between the input and output transformer bodies, it bips, but between transformers and any other groudn connection, I got this higher reading value. So, I guess my trouble is around the input transformer, amplifying ambiant noise? It would be a kind of grounding/shielding issue? My chassis is full aluminium. Maybe it just don't make a good contact with the chassis, possible?

 

[Audio1Man]

I think that it oscillating.  Look at the output with a scope.
Duke

 

[Deepdark]

Thanks. I guess I should move the transformers then? (they are quiet close to the xlrs)? in other words, it's oscillating because of the high gain?

 

[CJ]

does the cabinet have a door on it? run a wire from the door to the main chassis,

 

[Deepdark]

does the cabinet have a door on it? run a wire from the door to the main chassis,

It's a standard front panel, hold to the chassis by 4x 6-32 Screws. You think it could be a shielding problem on the chassis?

 

[CJ]

try grounding the pots on the front panel to the main chassis and see what happens,

watch for paint underneath the pots, use an internal lock washer, maybe solder a black #22 gauge wire onto the internal lock washers and run it back to a ground point on the main chassis,

 

[gridcurrent]

The issue is that: When cranking a 12 o'clock and more (I know it's not supposed to be cranked that much), there is a really high rise in Hiss and RF noise. So, I suspected some shielding trouble.

I suspect oscillation.
LA-2 and LA-2a's have a tendency to break into oscillation especially if the wiring has been "upgraded".  Not common but it happens.
The easy fix is a grid stopper in series with the input 12AX7a.  Suggesting 22k.
Rather ironic we were discussing stoppers in another thread,
with 12AX7a/ECC83, no question there is a pronounced effect.
Attached is an inside photo of a LA-2, serial number in the low 100's.  With stopper installed, has floated around Hollywood for several decades, considered  to have excellent sonics
Pertaining to grounding: ensure low resistance between the front panel and main chassis.

 

[johnheath]

Grounding issues in a LA2A is often recognized as hum and my method of getting rid of that is to have one grounding point. The original seemed to be grounding little all over the place… I do not know why really because it is not that hard to make one single grounding point.

The original have lots of gain and the 12AX7 is a monster in this case and therefore layout is critical in a home-brew.

Use shielded cables and keep leads short and away from each other as much as possible.


Best regards

/John

 

[Deepdark]

Thanks guys. Hum is not a trouble. Gain at max, input shorted with a 600ohm resistor, at 60hz, i  got about -110dbfs. The unit break into oscillation past gain at 12 oclock. The only components i have a higher resistance value at ground is the input and output transformer. Must be a weak contact i guess. If i pickup noise from the input tranny and amp,ify it, i guess it could explain what i'm experiencing. I use shielded cable everywhere by the way. All shield are connected to the chassis, and signal ground go to the main bolt. Maybe one shield cable hasn't a stromg contact to ground, too.

 

[Deepdark]

I suspect oscillation.
LA-2 and LA-2a's have a tendency to break into oscillation especially if the wiring has been "upgraded".  Not common but it happens.
The easy fix is a grid stopper in series with the input 12AX7a.  Suggesting 22k.
Rather ironic we were discussing stoppers in another thread,
with 12AX7a/ECC83, no question there is a pronounced effect.
Attached is an inside photo of a LA-2, serial number in the low 100's.  With stopper installed, has floated around Hollywood for several decades, considered  to have excellent sonics
Pertaining to grounding: ensure low resistance between the front panel and main chassis.

About the grid stopper, you mean a 22k resistor between the wipper of the gain pot and pin2 (grid) ? Why is that?

 

[Deepdark]

I'm thinking about it, the grid stoppers you mentionned wasn't on the 12bh7? Likea 1k resistor in serie at pin 2 and another one at pin 7.  You mentioned a 22k at the input 12ax7. On both grids?

 

[Deepdark]

I suspect oscillation.
LA-2 and LA-2a's have a tendency to break into oscillation especially if the wiring has been "upgraded".  Not common but it happens.
The easy fix is a grid stopper in series with the input 12AX7a.  Suggesting 22k.
Rather ironic we were discussing stoppers in another thread,
with 12AX7a/ECC83, no question there is a pronounced effect.
Attached is an inside photo of a LA-2, serial number in the low 100's.  With stopper installed, has floated around Hollywood for several decades, considered  to have excellent sonics
Pertaining to grounding: ensure low resistance between the front panel and main chassis.

Hi. I just drawn the grid stopper on the schematic. Is it what your were talking about? I add a capacitor to ground too.

 

[Deepdark]

and one last things, too. Because I use Cinemag transformers, they have a really flat frequency. So I removed C4 to eliminate the high boost. Is it ok, or I should remove other component of the negative feedback? I think only removing C4 should be ok.

 

[AusTex64]

About the grid stopper, you mean a 22k resistor between the wipper of the gain pot and pin2 (grid) ? Why is that?

The grid stopper forms a RC filter with the Miller capacitance and other interelectrode capacitances of the tube. It filters out RF, including the HF oscillation you're having.

I see no need for the 470pf cap from grid to ground. Where did that come from? It's not on any of my LA2A schematics.

I strongly suggest following gridcurrent's advice. I suspect he's seen and worked on more LA2A's than most everyone on this forum.

 

[mutterd]

Yes - GridCurrent will chime in here sooner or later I'm sure - and he is truly the expert - he's too modest to admit it - but he's responsible for a pretty famous LA2A mod that people still go nuts about...

but as is that 470pf shunt forms a low pass filter with the series grid stopper - 22k and 470pf alone gives us a 15.4kHz corner, but the thing is the 470pf is in parallel with the miller capacitance (tube gain *  grid to anode capacitance + grid to cathode capacitance) so that 15kHz -3dB point is gonna get even lower...

as for the removal of C4 - seems like a bandaid and may be unnecessary - if the grid stopper doesn't help, id wait till you've sorted out the grounding and shielding.

 

[Deepdark]

Thanks guys. In fact, I think I will always put the grid résistors haha. I'm not gonna take any chances. Is it recommended to put another 22k grid resistor at 12ax7 pin 7 (second grid) ? And, while putting grid all around, I guess it's not a bad idea to put the famous 1k grid stoppe rto the 12bh7 lol

 

[AusTex64]

I think the 22K resistor between the wiper of the pot and pin 2 of the tube will do the trick. Do one thing at a time and see what happens. You don't have to think too much when you have an expert like gridcurrent doing the heavy lifting for you. 

 

[Deepdark]

haha! Does the resistor must ideally be as close as possible from the tube socket pin? I have a run of about 4 inches to make between my pot connection on my pcb and my tube socket.

 

[AusTex64]

haha! Does the resistor must ideally be as close as possible from the tube socket pin? I have a run of about 4 inches to make between my pot connection on my pcb and my tube socket.

Yes, ideally mounted to the pin directly. Look closely at the pic gridcurrent posted.

 

[mutterd]

In theory - the 22kΩ resister adds to the source impedance of that wire  - so is 4" of high(ish) impedance wire a problem?

it might be - especially if you have instability in the circuit already - if you can park the 22kΩ resister it as close to the tube socket as possible you're eliminating the chance of any noise  forming across the wire...

haha! Does the resistor must ideally be as close as possible from the tube socket pin? I have a run of about 4 inches to make between my pot connection on my pcb and my tube socket.