Dual Pultec Build Ground Noise

[Janalex]

Hey all,

I built this pultec some time ago when this guy was selling these boards and enclosures. I only later learned that he had ripped off Mr Gyrafs design. At this point it’s built and a really useful tool so all props to the designer. I am trying to tackle a noise issue though and I have read that it’s not uncommon however some have fixed it. Although I read the threads on it I can’t seem to spot were to start on this build. Can anyone see something that would benefit from correction to reduce noise? The noise is exactly the same on both channels. Sounds like a grounding issue.

 

[living sounds]

If it's exactly the same on both channels this does sound like a grounding problem.

However, first check for correct DC power ratings and low AC (hum) residual.

If that part is OK look at the grounding. This is a fully balanced built (four trafos), so ideally pin 1 on the XLR inputs and outputs goes to chassis only (not to audio ground or anywhere else). The only connection of audio ground to the chassis (and thus to the cable shields via pin 1) should be from the PSU ground via a sufficiently thick wire to the chassis bottom.

Also look at your cabling, the problem might not be in the unit but the way you connect to your audio interface or other equipment.

 

[Ike Zimbel]

Hey all,

I built this pultec some time ago when this guy was selling these boards and enclosures. I only later learned that he had ripped off Mr Gyrafs design. At this point it’s built and a really useful tool so all props to the designer. I am trying to tackle a noise issue though and I have read that it’s not uncommon however some have fixed it. Although I read the threads on it I can’t seem to spot were to start on this build. Can anyone see something that would benefit from correction to reduce noise? The noise is exactly the same on both channels. Sounds like a grounding issue.

I would start with the green wire that is leaning against, and therefore not connected to what is presumably the chassis ground lug, next to the fuse holder in the first pic. That was probably meant to be connected there. Having said that... a few notes:
1) BE CAREFUL!!! There will be lethal line voltage on the fuse holder terminals, so UNPLUG the AC cord first!
2) This is a kit build, so all bets are off as to whether that is the optimal place for that connection.
3) On closer inspection, that green wire looks like it was meant to be a ground tab somewhere, possibly on one of the XLR boards, and is just floating around in there. This is a bad thing. Remove it.
4) There is a heavy black wire that appears to run directly from the ground connection on the AC inlet to a central ground point between the two boards. Safety dictates that this wire should connect to the chassis, on a separate #6 screw (iow, not a mounting screw for any other part) directly adjacent to the AC inlet.
5) The XLR marked "A" is not fully assembled in that the board holding the pins is not seated in the shell properly, so its Pin-1 may not be making proper contact with the shell, or your cables.
6) The frequency of the hum in question is ALWAYS worth including in your inquiry as different frequencies indicate different faults. To wit: 50/60 Hz indicates a ground connection issue while 100/120Hz usually indicates a power supply filtering issue.

 

[Janalex]

Thank you!

So I disconnected the shield from the terminal block on all 4 xlr pcbs. The noise went away. The 4 green wires in the pic of the xlrs connects pin 1 of the xlr to the chassis via the pcb. Can I leave that as is? And now what to do with the 4 audio shields that I disconnected. Ground them to chassis I presume elsewhere..?

Quote
The only connection of audio ground to the chassis (and thus to the cable shields via pin 1) should be from the PSU ground via a sufficiently thick wire to the chassis bottom.

I guess I need guidance how to implement this where I’m at now with 4 floating audio cable shields disconnected from the xlr board terminal blocks.

 

[Janalex]

The floating green wire mentioned near the AC outlet was a ground wire that I disconnected temporarily. I instead used that thick wire to go from AC outlet to the central star ground hoping to reduce noise.

 

[RoadrunnerOZ]

Hey all,

I built this pultec some time ago when this guy was selling these boards and enclosures. I only later learned that he had ripped off Mr Gyrafs design. At this point it’s built and a really useful tool so all props to the designer. I am trying to tackle a noise issue though and I have read that it’s not uncommon however some have fixed it. Although I read the threads on it I can’t seem to spot were to start on this build. Can anyone see something that would benefit from correction to reduce noise? The noise is exactly the same on both channels. Sounds like a grounding issue.

First up your input/output board “A” is sitting loose off the XLR bodies, the socket bodies need to be locked firmly into their housings - I would make sure that the sockets are both lined up to be in the same plane as the PCB - it looks like they are tilted backwards so your board will sit at an angle so you may want to re-solder them so they sit true.
Second, try lifting the grounds on only 1 pair of the I/O PCB’s terminals ie. either the input ground or the output ground on boards A and B and see if the noise goes.
It sounds like you’ve got a hum loop somewhere so you need to also check that the audio device feeding into the Pultec and the receiving audio device are both plugged into the same power outlet - some studios I have been servicing gear in have different wall power circuits with their grounds running back to the switchboard and this creates a massive hum loop. Using a”Star” ground system solves this problem - usually the recording console (or the audio interface) sit at the Star centre and power distribution is done from there. This way there is one ground from the main switchboard for all devices.
Also another system to prevent hum is to have only one end of shielded XLR CABLES grounded - that is pin 1 is floating at one end (except for mic cables - installation cabling only). It should be the same end for all XLR installation cabling - either all the male ends or all the female ends. This prevents the audio cabling ground connection between two (or three) devices and power grounding of the same from creating hum loops. Long audio cables with shield connected at both ends and short power cables, even star grounded, can cause hum loops. This is an auto transformer effect. The grounded noise current in a short cable is multiplied in the audio ground by the ratio of the total cable ground length, power + audio/power for short power leads and long audio leads or power + audio/audio for long ground runs back to a mains switchboard that is longer than the audio cable - so in effect you can get a step up in voltage of a small hum signal either in the audio cables or the equipment chassis. The more devices and audio channels with common grounding at all ends, the more hum loops possible.

 

[Janalex]

Hi all,

Please disregard the above post. After disconnecting the shields from the I/O PCB terminals there is NO change. This isn't surprising because the other end of the shied is already grounded at its own point in the corner of the SRPP board. I have the EQ getting pink noise from the computer and then directly into a powered speaker. All three devices are on the same electrical circuit.

Not sure how to test for this and if its still relevant. The hum is 60hz so ground connection issue.
"However, first check for correct DC power ratings and low AC (hum) residual."

I fixed the issue with the XLR connector in the pic. The XLR Pin 1s are connected to the chassis adjacent by way of the green wires in the pic via the PCB boards. But this may not be adequate if I understand correctly. Sounds like PIN 1 should be connected to the chassis ONLY. The PCB on which the connectors are mounted as well as the green wire to the chassis connect PIN 1 to the chassis AS WELL AS the BYPASS switches. Should I start by removing PIN 1 from the bypass PCBs?

The current grounding scheme is as follows:
Each SRPP has its own connection to chassis bottom to which the in and out audio shields connect and EQ board in and out shields connect. The SRPP boards are also grounded to a STAR point in the center of the chassis bottom to which the AC inlet ground also connects. The XLRs are grounded to 4 adjacent points by way of the bypass PCBs.

 

[living sounds]

Check this document detailing grounding/shielding best practice:

https://www.ranecommercial.com/legacy/note151.html
As Ike wrote above, it's crucial that the safety ground from the power inlet is bolted securly to chassis. Then you need to make sure there are no ground loops and correct shielding, IMO this is best done by connecting all pins 1 only to chassis at the point of entry/exit and connect audio ground only to chassis through the PSU ground.

It may well be that the PCB stands connect to the chassis or even the pots, make sure with a tester that there is no connection and then apply the single connection as described above (with the unit powered off, of course, and make sure there is no high voltage left in the caps for you to get zapped before you take anything apart).

Having the power trafo so close to the relatively sensitive EQ circuitry is not ideal, but since you have the same hum on both channels the problem appears to originate somewhere else. Again, make sure the cabling with your interface is actually all properly balanced.

 

[Janalex]

Ok. Thanks for the reply. So I’ll start by disconnected these bypass pcbs with the relays and connect pin 1 straight to chassis adjacent.

The thick wire in the pic is carrying AC inlet ground to the bolt with ground tab in the center of the chassis. The audio cable shields currently have one end grounded at their respective srpp boards and the other end should be grounded to the center bolt I understand. Two distinct ground locations ?

 

[living sounds]

I would connect the safety ground from the inlet to chassis right next to the inlet, not with a long wire to some other place on the chassis.

Try to have one thick wire from each srpp boards 4700 uf negative side (ground) to the chassis ground bolt in the middle. No other ground connections. Nothing connected to the pins 1 on the pcbs - these are probably where the ground loops originate. Your only ground/chassis connection from the PCB is through the one wire each to the ground point in the middle. That should do it.

 

[Janalex]

Question for those familiar with the schematic. After installing the input transformer I get continuity between the input the to the EQ board and ground. Unless I’m reading the schematic wrong there’s an eq input going to a hi boost pot, an eq output going to a 220n cap, and a separate ground. Should there be continuity ?

 

[gyraf]

Is it perhaps the dc resistance of transformer secondary?

 

[Janalex]

Well the good news is that I have hum free operation when the second srpp board is disconnected from power. The moment I connect power to the second board I get the noise. I have included pics. It’s a bit of a mess in this region of the chassis now cause all the testing I’m doing but I had better wire separation initially when I started and noise was there. Also I used to get the same noise from either Channel suggesting not proximity of these power lines to the audio wires nearby on board A. On board B the eq I/o is removed and that board is grounded to same location central chassis and board A.

The noise is solely induced by the 12v 2a line.
Any ideas?

 

[ccaudle]

I disconnected the shield from the terminal block on all 4 xlr pcbs

What signal are you shielding, and from what are you shielding it?
In general signals should run in a shield, but one shield should not be inside another (barring specific need) to avoid bringing interferring currents inside the outermost shield (i.e. conducting noise right inside the chassis, which is what you are trying to prevent with shielding in the first place).
The cables have shields, the case is a shield, so the cable shield should connect to the case (as close as possible to where the cable connects to the case), and then everything inside the case is shielded as long as the case is conductive and does not have gaps in connections (seams, holes, etc.).

The 4 green wires in the pic of the xlrs connects pin 1 of the xlr to the chassis via the pcb

Ideally you want a low resistance and low inductance connection from pin 1 to the chassis, which means a short, wide connection. A small tab of metal which wrapped around from pin 1 to the shell of the connector would be ideal. The green wires are the right idea, but much longer than needed.

One potential problem I do see is that the chassis is black everywhere I can see in the picture, and I do not know of any conductive metals which are black.
That likely implies that your entire shielding chassis is covered by an insulator, which is a problem when you want all the interference currents to flow on your shield (i.e. chassis) and not be forced onto your audio circuitry. When I put some circuitry into a pre-painted chassis I had to go at the paint with a sanding drum on a Dremel tool to make sure I had bare metal wherever connectors and wires needed to connect, and where the various chassis pieces overlapped. You will likely need to do something similar if you really want to use the metal chassis as shield.

now what to do with the 4 audio shields that I disconnected.

Why do you have shields inside the chassis shield? The signal is between the hot pin (pin 2 on modern equipment, sometimes pin 3 on older US designed equipment) and the cold pin. Once inside the chassis the chassis is the shield, so just twist the audio signal wires together and take them where they need to go (presumably to an input transformer on this design).

4 floating audio cable shields disconnected from the xlr board terminal blocks

Just use twisted pair wires with no shield inside the chassis. Once on the secondary side of the input transformer one side will presumably be tied to circuit reference, just let that follow the signal, so where you have to go from one circuit board to the next the circuit reference connection will go alongside the audio signal connection.
I assume this unit has output transformers as well, so the output transformer is the same as the input transformer, but in reverse: the primary has connection to audio signal and circuit reference, the secondary should have a twisted pair connecting to pins 2 and 3 of the output connector; the output connector pin 1 should have a short direct connection to a (conductive) spot on the chassis.

 

[Janalex]

Thanks for the last reply. I think my last post suggests it’s not the shielding since one powered board seems to be working fine and essentially hum free. The second board is just a mirror image of the first. Yes all the ground to chassis location were cleared of the insulation material.

Im noticing Don audio sells 2 different toroidal transformers, for mono and for stereo builds. The specs of the mono one are not provided to compare the 2 however is it possible I need to upgrade the transformer to get rid of the hum?

220v .1A
12V 2A
Are the current secondaries. My transformer doesn’t have 3 secondaries as it probably should. And isn’t hooked up like this. Although it pretty much works in mono could this be the issue in stereo?

 

[ccaudle]

is it possible I need to upgrade the transformer

If the transformer is undersized you should see a pretty large drop in voltage when the second channel is connected. It sounds like you don't have a 'scope, but presumably you have a voltmeter to check. Best would be an o'scope to see how much ripple you have on the power rails. Power issues are more commonly 2x the power line frequency, so 100 or 120 Hz instead of 50 or 60 Hz.

And isn’t hooked up like this

It is kind of hard to tell exactly what you are talking about since you only include partial screen shots without enough context, but that two transformer scheme looks like a way to generate B+ when you can't get a transformer made for tube gear.

he current grounding scheme is as follows

That sounds like a good way to have no way to control where the power supply currents flow. Why do you have so many connections to the chassis? What were you trying to accomplish?
To reiterate: the chassis is the shield for the circuitry, it is not part of the circuitry. It is not part of the audio circuit, it is not part of the power supply. The circuitry should connect to the chassis at one point. That doesn't mean multiple places of the circuitry should converge at one location of the chassis , it means one location of the circuit should connect to the chassis.

 

[Janalex]

Sincerely appreciate the reply. To address the above points.

1. Currently I'm getting 312 V DC after the power resistor on both boards. With one board connected no noise I get 317V on that board. I don't have a scope. This is high and should be brought down. I don't think the tubes are damaged yet though since one board is pretty much silent when working alone. Can just the elevated voltages at the tubes be amplifying hum excessively? I ordered the resistors arriving tomorrow.

2. I scoured the web for pics of this build and seems have just 2 secondary voltages is adequate, 220V and 12V. People have 2 twisted pairs hooked up like mine so non issue. Also like I said a single board is working fine. Also as stated the hum is identical from both outputs when both boards connected.

3. I have reduced ground connects to chassis to simply one green wire from each SRPP board to one point in center chassis, one point near AC inlet for AC ground, and XLR grounds in proximity to each XLR. There was a relay based bypass circuit in place (all the twisted red wires) which is currently disconnected. Also the twisted red wires supply the LEDs (all 6.1V). This is all disconnected. I also tried flying the twisted power black power rails from the trannie to the second board above the boards away from the chassis like in this pic but it didn't help. I tried reversing the tubes to see if board B had a bad tube. Maybe my transformer needs to be laying flat against the bottom of the chassis like in this pic. I don't know.

 

[RoadrunnerOZ]

Thanks for the last reply. I think my last post suggests it’s not the shielding since one powered board seems to be working fine and essentially hum free. The second board is just a mirror image of the first. Yes all the ground to chassis location were cleared of the insulation material.

Im noticing Don audio sells 2 different toroidal transformers, for mono and for stereo builds. The specs of the mono one are not provided to compare the 2 however is it possible I need to upgrade the transformer to get rid of the hum?

220v .1A
12V 2A
Are the current secondaries. My transformer doesn’t have 3 secondaries as it probably should. And isn’t hooked up like this. Although it pretty much works in mono could this be the issue in stereo?

Do you have a full schematic for the unit you have? The segment you posted showing dual transformers does not appear to be for yours.
In the Gyraf schematic there is a 220:12V step down transformer - this provides 12V AC creating heater voltage by converting to 6.3V DC and also feeds AC 12V to a second step transformer 12:220Vup for the HT DC which according to that schematic is 250V DC - the one partial schematic you have posted (although showing a typo?? of 15V??) just shows the interconnection of the two transformers via a fuse.

Your build has only one power transformer with two secondaries from the first photo - what are the AC voltages from the two windings? If the HT AC is too high the obviously your DC will be the same.
Does the 12V AC drop when you connect the second board?

For a dual build running off the one transformer the connections of the transformer windings should match pin for pin on the connector at each board - not swapped over - you’re running twisted wires so check which wire is connected to pin one on board A 12V and make sure it connects to pin one on board B - just disconnect the supply from board B and use your meter to identify which way round they are. Same with the HT.
Does the same no noise scenario occur if you connect the 12V line only to board B and not to board A? You showed disconnecting & connecting board B with A connected in your vid.
I’m not sure why you are getting over 300V on your HT line - what does it say on the schematic you have for the unit you built? DC wise and AC as well.
Did the transformer come with the kit or was it purchased separately?
Where are you measuring that HT voltage where you got 317V? Both boards? The rectifiers are separate for each board but they should both have the same DC output.
Anyway the way the circuit works there is only 1/2 rail voltage across each triode in the tube so if they are 6922 or ECC83 they are rated at 220V, can handle higher at low wattage and are unlikely to be damaged by your higher voltage, but if the circuit runs at the wrong voltage it won’t perform quite as it should.
If the heater voltage is low or the heater supply voltage regulator fails you will get hum - you should be getting 6.3V DC on pin 2 of the LM317. I’d check both boards heater supply voltages.
Sounds suspiciously like the transformer is wrong for the job.
Really needs a scope to check al the supply rails for excess ripple.
Edit:
Realistically both of the transformer outputs should be in phase - Heater and HT windings plus the couplings to both boards as you may be generating ground currents as the Heater and HT Rectifiers on both boards share a common ground.

 

[SmokeyJoe]

3. I have reduced ground connects to chassis to simply one green wire from each SRPP board to one point in center chassis, one point near AC inlet for AC ground, and XLR grounds in proximity to each XLR.

So you still have three separate ground connection points to the chassis? Have you tried taking everything to one single point?

Could also try re-orienting the power transformer (on all three axes).

Edit: I’ve just checked my (unfinished) G-Pultec and the board does not have a separate ground. The schematic shows it should only be connected to ground via the XLRs, which should have their grounds daisy chained and only connect to the chassis at output XLR pin 1.

Edit2: I’m starting to question whether the XLRs should have their grounds daisy chained, as they already ground via their connection to the PCB. Perhaps someone else can weigh in here?

 

[ccaudle]

you still have three separate ground connection points to the chassis?

schematic shows it should only be connected to ground via the XLRs

only connect to the chassis at output XLR pin 1.

OK, this is a pet peeve of mine, and part of the reason that a lot of old designs are difficult to get to work optimally in modern installations (i.e. with a lot more RFI than there was in the 1950's).
The standard for XLR connector wiring is standardized in Audio Engineering Society standard AES14, with references back to IEC 268-12.
Pin one is defined as "Screen (shield)" connection. Using the vague word "ground" misleads people into interconnecting all kinds of circuit nodes together which have different purposes and should not just be connected together thoughtlessly.
A metallic chassis acts as an overall shield for the circuitry inside, so to prevent noise and interference currents from getting into the circuitry inside the chassis, the cable shield should be connected to the shielding chassis as directly as possible (optimal would be a connector shell which wraps around with pin 1 swaged into the shell, but I do not think there are any current connectors made that way).

If the circuitry inside the shielding chassis is completely floating with respect to chassis you have undefined common mode offset from the circuitry to the shield, and you will have currents flowing through parasitic capacitances into the chassis at various places, so the power supply common should be tied to the chassis at one point to define the circuitry common mode voltage to be the same as the chassis potential (typically close to AC power input grounding/earthing point, but if the chassis has decent conductivity it doesn't have to be super close).

There are some slightly different considerations for phantom power because in that case the shield pin also becomes the return pin for phantom power, but as far as I am aware there would be no phantom power in a typical Pultec EQ build (i.e. not some kind of channel strip version that had mic amp and EQ together). That makes this really simple, so no need to try to make it more complex than it needs to be. It looks like those connectors used have a pin connected to the shell, so 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.

That won't necessarily fix every hum problem you have, but it is a necessary starting point to keep the shield currents out of your audio circuit.

This advice is based on basic electrical physics, and has been widely discussed since 1995 when the journal of the audio engineering society published the special issue on grounding and shielding. Whether other particular implementations of similar circuitry have used less optimal schemes in the past doesn't change the physics of the situation.

I have also highlighted two questionable areas from the original pictures you posted. I do not find it easy to follow the wiring paths in the original pictures, but these are two areas that jumped out at me.