Dual Pultec Build Ground Noise

[SmokeyJoe]

put a little piece of solder braid between pin 1 and the shell pin, solder it on and done. Don't connect pin 1 to anywhere else in the system

So, based on what you’re saying, it’s better to avoid making a connection from xlr pin 1 to the pin 1 connections on the board shown below?

Presumably both the input and output connectors need their pin 1 connected to their shell.

Since there is no other ground connection to the pcb, should a ground be taken from one of those pin 1 connection points on the pcb and run to the mains chassis ground?

 

[Janalex]

Thank you for all the info. I have so many things to consider after the last posts. And many things to try. In terms of grounding on this exact unit and I am also following the direction of a prior post awhile back where someone actually measured the noise reduction each step of the way. I would have never guessed upon some of these grounding schemes but worth a try. I don’t know what “normal” connection is in the first line here… is he connecting the ac inlet to star point and xlr grounds on the pcb to the xlr pin 1s? I was gonna connect those grounds to star ground each with its own wire. One thing is for sure and that’s that one ground wire from each srpp pcb to star ground is not enough.


-63,4 db
normal connection and 0v to star point ground
------------------------------------------

extern rectiviering (only with the 350v wires)
- 63,6 fb

grounding the eq filters to starpoint
-71.4

grounding the eq filter to starpoint AND Starpoint the psu "Filter Ground"
-75db

Distancing the 0v StarGround from the toroid...
-75,7

0v StarGround cable under the potentiometer - front section.
-76,5

Rotating the toroid vertical to the right site of the inside chassi.
-81 db
Hum is gone!

- Rotating The toroid did a hum difference of -76 to -95db on one channel. This was the most hum.

 

[RoadrunnerOZ]

Thank you for all the info. I have so many things to consider after the last posts. And many things to try. In terms of grounding on this exact unit and I am also following the direction of a prior post awhile back where someone actually measured the noise reduction each step of the way. I would have never guessed upon some of these grounding schemes but worth a try. I don’t know what “normal” connection is in the first line here… is he connecting the ac inlet to star point and xlr grounds on the pcb to the xlr pin 1s? One thing is for sure and that’s that one ground wire from each srpp pcb to star ground is not enough.


-63,4 db
normal connection and 0v to star point ground
------------------------------------------

extern rectiviering (only with the 350v wires)
- 63,6 fb

grounding the eq filters to starpoint
-71.4

grounding the eq filter to starpoint AND Starpoint the psu "Filter Ground"
-75db

Distancing the 0v StarGround from the toroid...
-75,7

0v StarGround cable under the potentiometer - front section.
-76,5

Rotating the toroid vertical to the right site of the inside chassi.
-81 db
Hum is gone!

- Rotating The toroid did a hum difference of -76 to -95db on one channel. This was the most hum.

Did you check both boards separately to see if they are each quiet with the 12V to one board only?

 

[Janalex]

Yes I checked each board working alone and they are equally quiet when running as a single. With all these star point ground connections I’m measuring 0.1ohms at every ground point. Unfortunately I’ll be away from the build until Sunday but processing all this info. Will also check the ac polarities as mentioned as soon as I return and other stuff mentioned. I bought the diodes for external rectification if it comes to that as well. My toroid is already vertical which seemed to work for him. Seemed that the filtered board grounding scheme made a huge difference.

 

[RoadrunnerOZ]

Yes I checked each board working alone and they are equally quiet when running as a single. With all these star point ground connections I’m measuring 0.1ohms at every ground point. Unfortunately I’ll be away from the build until Sunday but processing all this info. Will also check the ac polarities as mentioned as soon as I return and other stuff mentioned. I bought the diodes for external rectification if it comes to that as well. My toroid is already vertical which seemed to work for him. Seemed that the filtered board grounding scheme made a huge difference.

Once you determine that both boards have the AC inputs in phase you may also want to try reversing the wires from the toroid, on one pair only, in case they are out of phase. I have serviced a lot of gear with dual winding power transformers whose separate rectifiers for differing DC voltages are common grounded and these always have the AC windings of the outputs in phase. Without an oscilloscope it’s a bit tricky to check - there are no markings on the toroidal to indicate what the polarity is.

 

[ccaudle]

So, based on what you’re saying, it’s better to avoid making a connection from xlr pin 1 to the pin 1 connections on the board shown below?

Right, you do not want the interference currents from the shields flowing through your audio circuitry. Keep them on the cable shields and chassis, not in your audio circuit.

 

[SmokeyJoe]

Right, you do not want the interference currents from the shields flowing through your audio circuitry. Keep them on the cable shields and chassis, not in your audio circuit.

Thanks, I did some reading around it after your last comment and that makes a lot of sense.

 

[RoadrunnerOZ]

Audio ground and cable ground should both go independently to star ground where mains ground is terminated. Cable ground can go via the chassis link pin jumpers from pin 1 at the connector using the chassis as a path to ground without any wire to star ground and audio ground from the boards carried via wire to star ground. The two audio boards have no reason to have a ground link between them (A to B) and should go to star ground independently.
Some XLR cables come with the XLR body link jumpered to pin 1 already - if this is at both ends it can cause problems.

Some studios have their patch bays wired grounded and pin 1 lifted at the equipment end of any balanced XLR connection to the patch bay, (others lift ground on inputs at the patch bay and installation cable XLR’s and connect pin 1 at source outputs but this can cause problems). It can also be helpful to disconnect pin 1 on equipment to equipment direct connections at the inputs. This precludes the possibility of ground loops between audio cabling ground links and power grounds as no two pieces of equipment are connected except via the ground at the mains distribution - usually centred at the console.
I have just finished an install of a Neve 5105 36 channel console with wall plates up to 25 metres away and no hum loop problems. All power centres at a mains switchboard with the console ground 2 metres from it.

 

[Janalex]

So cable grounding then is just a direct link to ground between chassis and each xlr connector. No actual wire to the star ground point. Please confirm

I got the chance to mess with the polarity of the ac lines going to board B and had no effect.

I grounded all the audio and filter boards to star ground in the bottom center. It is not the same point as mains ground but for the time being I will bring a ground cable from the ac inlet to the center bottom for testing.

 

[Janalex]

Btw I am testing with nothing plugged into the input. I here almost no noise with one channel that way so figured the test with both channels should be the same and equally quiet if done right

 

[RoadrunnerOZ]

So cable grounding then is just a direct link to ground between chassis and each xlr connector. No actual wire to the star ground point. Please confirm

I got the chance to mess with the polarity of the ac lines going to board B and had no effect.

I grounded all the audio and filter boards to star ground in the bottom center. It is not the same point as mains ground but for the time being I will bring a ground cable from the ac inlet to the center bottom for testing.

Yes that is correct as the XLR chassis’s are bolted to the case they don’t need a second ground. You can confirm this by disconnecting the ground wire and use your meter to confirm proper ground connection from XLR body to Star ground point.
There should be 1 ground point for everything including the mains. Did you try identifying/swapping the AC coupling wires for the 12V and the HT? Do them 1 at a time.
There is a way of checking phase of the two transformer windings using an AC voltmeter but it’s just as easy to swap wires.
Also check the leads you are using going out don’t have case to pin 1 connected - they should just get their shell/case ground from the chassis.

 

[Janalex]

Just had time to test one more time. Board A and Board B alone are dead silent when powered solo, maybe minimal from the board closer to the transformer but nothing to be concerned about. Shields are not connected anywhere. I’m not even using the chassis mount xlrs. I’m alligator clipping my balanced signals cables to the xlr hooked up to the powered speaker with 2 alligator clips at a time. If I want a signal I’ll connect the input to the generator but I don’t have to cause I can hear the noise set in with no input connected the moment both boards are powered. Same noise level from both channels. All the boards are grounded to the star point independently. Still have to bring a cable over from the ac inlet. I did try reversing the power rails and didn’t hear a difference. I don’t know where a ground loop can still possibly exist but I’m putting all my eggs now in the external rectification basket.

 

[RoadrunnerOZ]

Just had time to test one more time. Board A and Board B alone are dead silent when powered solo, maybe minimal from the board closer to the transformer but nothing to be concerned about. Shields are not connected anywhere. I’m not even using the chassis mount xlrs. I’m alligator clipping my balanced signals cables to the xlr hooked up to the powered speaker with 2 alligator clips at a time. If I want a signal I’ll connect the input to the generator but I don’t have to cause I can hear the noise set in with no input connected the moment both boards are powered. Same noise level from both channels. All the boards are grounded to the star point independently. Still have to bring a cable over from the ac inlet. I did try reversing the power rails and didn’t hear a difference. I don’t know where a ground loop can still possibly exist but I’m putting all my eggs now in the external rectification basket.

Try reversing ONE of the wire pairs from the transformer - say the 12V pair. See if there’s a difference in the noise or if it goes away. Don’t reverse both the red and the blue pair or it’ll just be the same.
Did you use a meter to confirm pin 1 going to pin 1 on each board for both AC rails? (you will get a resistance reading through the transformer but the 0 ohms one is the match)
When you reverse the transformer wires is a good time to check the board to board wires.

 

[Janalex]

I just checked. The ht voltage rails have a point that measure 0 ohms. The 12v rails have a point that measures 0.2ohms. It is the lowest noise setup for sure but the overall noise is still there. Even though it’s mostly the same in both channels I have to go on the premise that the long AC rail crossing over to board B is somehow bleeding interference into both boards. While some prior posters have shown a modest improvement with external rectification others I have emailed have had larger success and definitely are strongly advocating for it. So I’m gonna start wiring that tonight. Wish me luck

 

[RoadrunnerOZ]

I just checked. The ht voltage rails have a point that measure 0 ohms. The 12v rails have a point that measures 0.2ohms. It is the lowest noise setup for sure but the overall noise is still there. Even though it’s mostly the same in both channels I have to go on the premise that the long AC rail crossing over to board B is somehow bleeding interference into both boards. While some prior posters have shown a modest improvement with external rectification others I have emailed have had larger success and definitely are strongly advocating for it. So I’m gonna start wiring that tonight. Wish me luck

Did you swap the secondary wires of the blue (or the red) from the transformer to the first board?
By that I don’t mean swap blue FOR red.

 

[ccaudle]

I did some reading around it after your last comment and that makes a lot of sense

The term used to describe the incorrect connection is often "pin 1 problem," so if you search that term you will find a lot of material (quite a bit even on gdiy forums).

 

[Janalex]

I’m swapping wires on the second board only. I thought the point is that a blue wire on the transformer for example goes to the same spot on both boards, etc. today I got the bridge rectifier so that seems to be my only hope now.

 

[RoadrunnerOZ]

I’m swapping wires on the second board only. I thought the point is that a blue wire on the transformer for example goes to the same spot on both boards, etc. today I got the bridge rectifier so that seems to be my only hope now.

What I’m driving at is not just getting the two boards the same, first, but then swapping the wires of one of the winding outputs from the transformer at the first board to see if the windings are out of phase and causing a ground current problem. This is not the same as making sure both boards are in phase, but making sure that what comes in is also in phase.
In the long run this may make little or no difference, but the coupling of AC coming into the boards to ground or DC rails can cause ground currents with two identical boards having out of phase AC inputs as they all share a common ground.
The real fix would be one single heater supply and one single HT DC supply feeding both boards and only the supplies providing ground for the boards.

 

[ccaudle]

Board A and Board B alone are dead silent when powered solo

cause I can hear the noise set in with no input connected the moment both boards are powered.

You mean if you apply power to one channel it is quiet, apply power to only the second channel it is quiet, but when both channels are powered they are noisy?
You said you do not have access to an o'scope to check the power stability, correct?

I don’t know where a ground loop can still possibly exist

The term "ground loop" typically refers to the case when different pieces of equipment have current flowing through the cables, and because of incorrect connection of shields and power earth connection some of that current flows through the audio circuitry. If you do not have other pieces of equipment connected when testing it is not likely that the source of the noise is a traditional ground loop. You seem to be inviting problems by still insisting to connect all the different boards to the chassis, but without some voltage difference to drive current from the chassis connection and out a different location it should not cause problems.
I should note that it is not entirely clear what exactly you are referring to when you state that "All the boards are grounded to the star point independently."
Do you have schematics for these boards, so you can point out exactly what node you are connecting to the chassis? I suppose it is still possible that there is something that could drive circulating currents between the two buffer boards, but

Given your latest information that (if I understand correctly) with no other equipment connected, powering one channel is quiet, powering two channels is noisy, it seems power supply problems are the only likely explanation.

Since each channel has rectifiers and filter capacitors on the same board as the buffer amp it is not a case of the power supply capacitors being undersized for the current and causing a lot of ripple when two channels are connected, so the next likely cause would be exceeding the current rating of the transformer. That would usually just cause a lower voltage and increased transformer temperature, but perhaps the extra magnetic field strength inducing a voltage into some of the wiring.

I looked back through the earlier posts, but I did not see anywhere a specific model number of transformer used. How confident are you that the transformer is actually appropriate for the current load of both channels?

I’m putting all my eggs now in the external rectification basket.

You don't have to necessarily go to external rectification, leave the diodes on the power supply boards, but move the transformer out of the chassis as a test. Put on some temporary extension wires from the transformer secondary windings so you can move the transformer away from the chassis.

today I got the bridge rectifier so that seems to be my only hope now.

What are you going to do with an additional rectifier? What specific problem are you attempting to solve with that, and why do you think it will help?

There are some cases where an external power supply can help by getting the magnetic fields from the transformer farther away from the circuitry and wiring, but in that case you would need rectifiers and capacitors (and possibly chokes and resistors, depending on how you were splitting up the filtering).
If the capacitors are still going to be in the chassis you might as well just run AC to the chassis (from the secondary windings) and leave the rectifiers on the power supply board.

One big thing which just jumped out at me is that I cannot see where the input transformers are in your original pictures. I see two transformers on each buffer board (SRPP and power supply PCB) both labeled "Output Transformer." Is one of those being used as an input transformer? I zoomed in on the pictures, and the wiring is confusing to me, I see four terminals near the transformers, but one of them has two twisted pairs connected. I would expect one twisted pair to each terminal, a pair on the transformer primary, and a twisted pair on the transformer secondary. Really need a schematic to see what is going on.

 

[Janalex]

Thank you for the replies. What you circled in the pic has been disconnected. The unbalanced connections you circled were used with an extra wire added under each pcb to get 6.3vdc to power the LEDs and relays. The extra pcb wire and red interconnects are all disconnected. Each srpp just has a balanced in and out and filter connect. Yes I am not plugging anything into the input and still hearing the noise which is unchanged if I do plug in a signal. Yes each board is quiet on its own.

I will try the other phase suggestion mentioned. The point of the external rectifier adjacent to the transformer would be to only run dc high voltage across the chassis to board B. Someone stated it helped them. Would only do if every other option was exhausted

The transformer specs printed on it say 230v 0.1A red and 12v 2a blue
50VA


Don’t see a brand on it. Could I benefit from getting this one from Don Audio? He sells both stereo and mono versions for this pultec. He has 9v 2.5A to the filaments. Who knows what I’m really getting on mine..?

https://www.don-audio.com/g-pultec-power-trafo-sec-220v-9v-5V