Custom toroidal PSU +/-48V enclosure buzz..

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

Eliani

Well-known member
Joined
May 27, 2011
Messages
244
Location
Belgium
Hey Guys,

So I've been working on this project for a while now and some of you know how much work it takes to design an enclosure, get it coated and printed and have it looking the way you want it. So when I received my finished enclosure it looked great and I couldn't, wait to finish my project. So I bolted in the custom PSU, custom wound toroidal transformer, the circuits PCB and I closed the lit only to end up with a distinctive high pitched buzz on the output....  I was hoping you guys could give me some insight in what I might try to get rid of this annoyance ,  so here are some details.

The project entails the racking of some channelstrips from my ATI-Paragon 2 console,

The PSU circuit was designed with some help from a couple of geniuses on this forum and I did the PCB design myself. It's a typical dual LM317 and LM337 PSU with +/- 48V DC (with HVK versions of the transistors), +/- 20V DC  and 5V DV...
I've had Begis from the White Market wip me up a triple voltage Toroidal transformer specially for this project, and he delivered very nice work!

some remarks on the design:
- In the PSU design I've got all 0v ground lines separate so I can tie them together with audio ground coming from the channel strip
to a star ground point in the case. it's been done the same way in the full console.
- there was supposed to be a sixth voltage of +24VDC driving all LED's, I tied that one to the +20V rail since in my previous testing that worked fine and all levels shown on the VU or gain reduction metering were the same as they would be on 24V.
- I've got the 48V schematics in the attachment...

Schermafbeelding%202019-10-12%20om%2014.54_edited.jpg


I drew a custom enclosure in fusion360 and had it laser cut and folded by 247Tailorsteel and powder coated in my local coating shop. The entire enclosure is made from 1mm stainless steel!

In the next picture you can see I've got the PSU PCB mounted in the middle of the enclosure, I know I probably should have turned it 180 degrees and have the secondaries on the side of the power transformer, and this will be so in the next enclosure I'm having made. I did twist the secondaries together and turning the toroidal secondaries exit point away from the PCB did help a lot fighting the louder noises. I also added a Mu-metal ring around the toroid...

So I'm left with an enclosure that will output a buzz on the output when I close the lid, the noise becomes worse when I screw the lid down... I'm fine if I only screw down one side of the lid, but when I secure the other side with screws (making chassis ground contact) the noise becomes troublesome...

a few thing I noticed:
- when I hook up a microphone to the channel strip everything works, including the gate, compressor, alle ins/outs, all leds.
- when I put that microphone close to the toroidal transformer there is a large electromagnetic field around it picked up by the mic resulting in loud noise.
- when hooking up a passive DI I have to put it on top of the enclosure making contact with the uncoated lid or I get a load noise...
- when I put a jack in that passive DI all noises, including the buzz go away but turn in a very low hum...

I've tried everything I think might be the problem, I'm reading up on toroids and learning that they don't have a magnetic field outside their core, I'm learning about inductive loading etc, but can't bend my mind around this problem... I'm sure It's my fault, but I could really need some insight from you guys...

12-3-19-0104_bewerkt_edited.jpg


IMG_20191008_134313_edited.jpg


at least I hope you guys think it looks good...

IMG_20191008_141453_edited.jpg


 

Attachments

  • Screenshot 2019-11-24 at 18.38.53.png
    Screenshot 2019-11-24 at 18.38.53.png
    151.7 KB · Views: 34
You might need to tie the power supply pcb with a 1ohm 1 watt resistor to ground. At the stand of take a 1r 1w resistor and solder some wire to it then attach that wire to your ground.
 
One is a general principle that should be considered in the context of each circuit component which in a few words is: ground follows signal. When an AC signal is transmitted over a wire or a trace, whether it be an audio signal or AC power, it radiates a magnetic field. This can induce noise in nearby circuitry. The best way to combat that condition is to match that magnetic field with an equal and opposite magnetic field. Fortunately that is usually pretty easy since whatever AC is delivered to a circuit must return to ground somehow. So if you just run the return trace or wire directly next to the supply trace / wire, the fields will cancel. This is why, for example, heater wires in tube amps are gently twisted together.

Another general principal is the star ground arrangement. Specifically, you have several "separate" ground which are:  1) the chassis ground 2) the earth ground from the mains supply 3) the power supply ground 4) the amplifier circuit ground and 5) the phantom power ground. You want all of these to converge at the "star" point which should be at the ground side of the filter caps on your power supply. Although there are two special exceptions here: One is that pin 1 of the XLR mic in should be connected to the chassis with the shortest possible length of wire (many XLRs now have a metal spike next to the screw hole that stabs the chassis metal and connects pin 1 directly to the chassis over only a cm or two). Two is that usually there is a grounding bolt on the chassis next to the mains cable in where the earth wire is attached with a spade and toothed washer. Then, one wire goes from the bolt to the PS filter cap ground. Meaning phantom power ground return, chassis and mains earth wire have there own "star" and then they connect to the PS "star". So the whole "star" thing is actually somewhat recursive (each circuit in the channel strip is almost certainly further divided into separate ground regions that converge into one point somewhere).

So consider these two general principles in the context of your project and maybe you'll find the issue. My guess would be that there's a 90% chance, your issues are related directly to one or both of these principles. Usually the hard problems are actually multiple problems (meaning just because you perform some change and it doesn't fix the problem, that doesn't mean it did need to be fixed - you may need to fix multiple problems before the noise will actually go away).

There's zero reason why it should not be possible to make your project completely noise free. Go slow, take your time, have fun and then it will be that much more satisfying when your done and it works properly. And bear in mind that one-off projects never work 100% on the first pass.
 
@ Ian, I know it doesn't show in the pictures, and at first there wasn't... but I have a star connection close to the mains input connecting all grounds (0V's and audio grounds), The unit is at a friends house right now so I can't shoot a new picture for you. I'm receiving new enclosures soon so I will be building in a week or so...

@squarewave,  I am going to look into the XLR ground to chassis distance, and the rest of what you wrote, I'm trying to digest that bit!! ;-)  I'm keen on learning this stuff, I have had stuff like this happen to me before in a NE5532 circuit where the negative voltage rail was hooked to the virtual ground by accident, and that thing received radio signals better than my radio...!!

@Rocinante, is there a theory behind your statement? I've seen people connecting resistors between left and right unbalanced signals on youtube where noise just disappears ...

I'm still daunted by the massive electrostatic field I've created... could it be that the stainless steel have something to do with it? because it's not magnetic, and will not stop magnetic fields, or have I miss interpreted that somehow...?
 
Eliani said:
@ Ian, I know it doesn't show in the pictures, and at first there wasn't... but I have a star connection close to the mains input connecting all grounds (0V's and audio grounds), The unit is at a friends house right now so I can't shoot a new picture for you. I'm receiving new enclosures soon so I will be building in a week or so...

I can understand you connecting all the 0Vs and analogue grounds to a star close to the mains input but did you also connect the mains earth to the same star point?

Cheers

Ian
 
Looking at your transformer wiring, nothing is twisted together on a per winding basis. So you will be getting stray noise from this.
However, are you referring to an acoustic buzz from the transformer, or an electrical buzz? The two are very different. It is not uncommon for transformers manufactured in USA to have too little metal in them, and hence high circulating currents. Also you state this is a custom wind. There is potential there for things to be bad. If you are racking these strips to use them in groups, why not put the supply in an external box? That makes the process much easier.
 
@radardoug

I did end up twisting the secondary leads as you said and that helped to reduce the noise. I don't have pictures of this because I've got a buddy of mine testing the device. (with flaws and all...)

The noise is an electrical buzz, not acoustically
I thought about the idea of having an external PSU but that isn't going to work for me because I will be using the channels separately, and I'm building some more for my buddies and maybe even sell some if I can get it to work out. I've got 96 channel strips I could possibly end up racking.
 
Move the transformer and power supply outside and away from the enclosure as far as possible. Obviously you will need to unsolder mains, re-establish mains externally somehow, rig a temporary longer wire between the chassis bolt and PS "star" ground, etc. Put the steel cover on. Try your listening test again. Does the noise change? How?
 
One other thing. That looks like a very big transformer for only one module. Are you sure it needs to be that big? Bigger transformer means more magnetic field.
 
radardoug said:
One other thing. That looks like a very big transformer for only one module. Are you sure it needs to be that big? Bigger transformer means more magnetic field.
Actually I would think it would radiate LESS than a smaller transformer with the same load. If it was accidentally loaded too much that would be a problem. Does it get warm?
 
squarewave said:
Actually I would think it would radiate LESS than a smaller transformer with the same load. If it was accidentally loaded too much that would be a problem. Does it get warm?

That is a good point. How did you specify the secondary windings in terms of the dc supply currents required?

Cheers

Ian
 
"@Rocinante, is there a theory behind your statement? I've seen people connecting resistors between left and right unbalanced signals on youtube where noise just disappears ..."

Yeah. It's establishing a star ground and this method is used in several projects including the ez1290 to prevent noise. It has nothing to do with balanced signals.
 
ruffrecords said:
Are there any inductors on the audio board?

Cheers

ian

No there are not, this is a standard ATI Paragon 2 Channelstrip, lot's of encapsulated smd's...
 
ruffrecords said:
That is a good point. How did you specify the secondary windings in terms of the dc supply currents required?

Cheers

Ian

I kind of cheated on this one and took the same currents that are in the Pro6's service manual. The pro6 is basically a Paragon channels strip so that should pan out, and after having looked at the internals of a pro6 they look very similar in size as the original...
 
Eliani said:
I kind of cheated on this one and took the same currents that are in the Pro6's service manual. The pro6 is basically a Paragon channels strip so that should pan out, and after having looked at the internals of a pro6 they look very similar in size as the original...

So did you specify the ac current rating of the transformer to be the same as the dc current drawn?

Cheers

Ian
 

Latest posts

Back
Top