Power transformer radiation in chassis?

[electric_diaries]

Hi guys, firstly thanks so much for all the help and information to date.

I have come up against a bit of an unusual transformer interference problem with some mic preamps that I made. I have used the Edcor MXL5cs input transformers which seem to do the job BUT the copper shielding around the outside doesn't get rid of as much interference as I would like. After some transformer rotation and moving transformer 1 from position X (see attached picture) to it's current location, I was able to get rid of most of the hum from the power transformer in the input transformers.

HOWEVER when I put the cover on and close it all up, the input transformers start to hum again! I was always under the impression that the exact opposite would happen! I have never had this problem where adding more shielding results in more induced hum.

1) Does anybody have any suggestions as to what I should try next? (eg mu-metal on the input transformers? or maybe brass mesh around the toroidal? a smaller toroidal?)

2) Can anyone explain why this might be happening so that I can design to prevent it in the future?

Thank you all so much for you're time. I am forever in debt to this forum and it's regular gurus! Ice cream is on me if any of you visit Berlin.

 

[isophase]

hi,
make sure you have good contact/continuity within the chassis it self (usualy involve scratching some paint)
have you tried moving the power transformer around and not just rotating it on its axe ?
very nice build btw!
cheers

 

[electric_diaries]

make sure you have good contact/continuity within the chassis it self (usualy involve scratching some paint)

I have scratched the paint around the bolt holes and everything has good continuity.

have you tried moving the power transformer around and not just rotating it on its axe ?

Yes. I initially had it mounted flat on the bottom which produced too much noise in input transformer 2. Then I tried a few different positions before settling on it's current position. I rotated it until I found the quietest spot and it all seemed to dissappear....until I put the cover on top 

very nice build btw!

Thanks here is the front almost finished.

 

[Kingston]

1) Does anybody have any suggestions as to what I should try next? (eg mu-metal on the input transformers? or maybe brass mesh around the toroidal? a smaller toroidal?)

2) Can anyone explain why this might be happening so that I can design to prevent it in the future?

the forum nickname "who" sells PSU transformer shields on white market. They can cut some 10-20dB of that magnetic hum when you shield the whole PSU transformer. In your case you would first have to reorient the transformer sideways.

I've used them on some critical places and they are handy. I've heard there are mu-metal sheets like that available somewhere that are even better, but I haven't come across any.

As for the lid increasing magnetic hum, there's no bomb proof way to prevent that. Steel does that, bounces the magnetic field back in an unpredictable way. You can always use an all-aluminum case, but then you install it in a rack and the gear on top and bottom are steel lid and you have achieved nothing. Best remove the hum at the source (PSU).

Also next time don't skimp on input transformers. They are just about the most critical part on a project and as you now found out the hard way, they really need mu metal shields. (and then use edcors on output stages instead)

 

[tv]

"And now to something completely different":

If you want a "budget" solution, i.e. to fix it on the cheap, relocate the power transformer to an external enclosure - and keep it well away from the preamp's inputs.

This can also "fix" the "cheap input trafo" hum - on the cheap.


Or else, you will most likely have to go with "transformerless" preamps.

 

[JohnRoberts]

The magnetic field is being altered by the metal chassis.

Shielding can be used on both the input and power transformer. Back with my old kits, I actually added a simple metal shield by the power transformer. My last low noise phono preamp used an external power transformer and many consoles use external power supplies.

it can be done... I've had to design small installation amps with power transformer, high voltage (100V) transformer audio outputs, and mic input transformers all in a smaller than a shoe box package, but it was a PIA.

Mu metal is your friend.

JR

 

[electric_diaries]

Thanks guys, all good advice here.

As for the cheap Edcor trafos, I bought a few of them a while back to test out since I had been pleasantly surprised by some of their output transformers on other projects. I gave them a listening and sweep test and even threw a few squares at them and they actually worked out quite well so I decided to use them in this project since I was trying to use mostly stuff I already had in stock.

I usually go down the road of an external power supply to save this kind of hassle but I thought that I would try an all in one box for once. It's actually a bit of a pain now to give it an external supply so I'm just going to keep digging and hope that I can get rid of it with mu-metal. I'm just going to wrap everything.

Wish me luck!

Thanks for the link Kingston!

 

[tv]

... a bit of a pain now to give it an external supply ...

NOT the whole PSU, but ONLY the power transformer - you can use some XLR connectors (or similar) to connect it with the main box.

The simplest solution if you just want to start to use it.

 

[nielsk]

Steel, baby, steel. Lots of ferrous metal, find some cans to cover the input trannys. Copper won't do much for EMF

 

[Kingston]

Steel, baby, steel. Lots of ferrous metal, find some cans to cover the input trannys. Copper won't do much for EMF

Plain steel is actually a bit unpredictable and usually fails at shielding magnetic fields. I've found this the hard way having tried an internal steel box around a PSU transformer in a few projects. It did almost nothing. Distance is a much more powerful "tool". Use that steel box for external enclosure instead, and take that thing away from the criticals.

Now grain oriented steel is a more expensive thing, but it's nearly as good as mu metal in shielding magnetic fields.

 

[alexc]

I've found that when adding the steel lid makes the noise floor worse in the mains harmonics, it usually indicated an instability in the grounding scheme.

If you have come to an impasse with moving the psu transformer around, and also moving the signal transformers around, then I would review the grounding scheme and try some alternative connections.

I'm assuming you are using quite heavy gauge grounding wire - like mains earth wire wherever possible for ground wires.
Trial and error of finding some alternative ground points usually shows up a scheme which is low and more stable, with the effect that the steel lid no longer influences the outcome by more than a fraction of a dB.

Virtually everytime I've had this issue I've found an alternative ground point(s) which makes the whole thing perform as expected.
And usually it tends to be an unexpected place.

Things I've tried and found have little or no effect are steel shields and also toroid 'mu metal' or similar shields.

The only things I found work are i) moving signal traffos  ii) moving/rotating psu traffos  iii) DISTANCE  iv) alternative grounding nodes  iv) making sure continuity around case back/bottom/sides/front is good\

PS - I would ditch unshielded input traffos i a nicely constructed build such as this    Not worth the grief!

Good luck

 

[Kingston]

I've found that when adding the steel lid makes the noise floor worse in the mains harmonics, it usually indicated an instability in the grounding scheme.

Don't mix electromagnetic and electrostatic radiation on random. These two are very different things.

When you have the case affecting a "grounding scheme", it's most definitely electrostatic radiation at work. In other words 230VAC mains (or rarely lower voltage VAC from PSU secondary) getting into your circuits. Shielding electrostatic is trivial, and so is rerouting VAC wires to less sensitive areas of the case.

And as you have found out, mu metal shields around transformers have absolutely no effect on electrostatic fields since all of these come from the wires themselves.

 

[electric_diaries]

Plain steel is actually a bit unpredictable and usually fails at shielding magnetic fields.

Yes plain steal is more unpredictable than I would have thought. However I wrapped the power transformer in a couple of layers of plain steal and that seems to have taken it to a very respectable noise level.

Thanks for you're help everybody!

 

[JohnRoberts]

I find it logical that a ferrous metal cover would alter a magnetic field. While a conductive metal cover could also alter the path resistance of a brute force chassis ground.

Several moving parts at the same time to account for.

JR

 

[alexc]

Don't mix electromagnetic and electrostatic radiation on random. These two are very different things.

Not sure I ever encounted 'electrostatic radiation'. Surely there are electrostatic fields, but they are static. And so can have no induction effects whatsoever. (assuming no moving conductors!)

Purely electrostatic fields can be influenced by dynamic electric or magnetic fields, in which case they are ususally understood as basic EM fields.

I am simply saying the unexpected noise in an audio system is the result of a complex mix of grounding schemes, gain staging as well as induction effects from electrodynamic fields, magnetodynamic fields and electromagnetic fields.

A complex interaction of all those things within a unit as well as nearby external field sources.

Those things can combine to find some equilibrium and so be reasonably stable or not. If not stable, then just about any peturbation can make for some large impacts on the noise of the system.

A 'tipping point' if you will. Or a strange attractor of chaotic behaviour.

Like putting a lid on or taking it off. Or moving the unit. Or any 'reseating' of wires. Or switching on nearby units.

Just about anything can alter the noise floor, particularly when that noise is unstable.

I've only seen improvements with simple metal shields and so on in those unstable cases.

Thats why when I finish a build, I spend time 'jiggling' stuff about to find the sensitive areas and focus on optimising the arrangement of those. When I find the least sensitivity, the system is a lot more stable and less sensitive to other changes.

You can try to analyse them by deconstructing components in order to optimise which in my opinion is a fairly fruitless exercise, particularly in the context of 'one off units' or you can experiment to find the most stable arrangement.

In other words, start out with your best practice then alter the available variables to find the best mix.

Other people may have different approaches - that is mine.

 

[Mike Cleaver]

The best solution always has been and continues to be (for high gain equipment) a separate power supply, connected to the unit via shielded cables.
Back in the day, RCA (and other professional equipment manufacturers) did put power supplies in their consoles and other gear for radio and recording but they were located far away from the high gain (up to 100db) preamp/program amp combination.
Also, there was nothing unbalanced in those consoles, all transformer in, inter stage and output.
Most professional consoles today come with separate power supplies, meant to be rack mounted away from the console.
Placement of internal wiring and grounding is very important, both then and now.
Shielded wire is best for anything below line level.
If you must build something in one case, mount the power transformer(s) at the rear on the outside, as far away from the sensitive electronics as possible.
Play with the orientation before drilling the holes measuring for the least amount of noise before bolting them down. 
Sometimes right angles between transformer cores work best, sometimes it's 45 degrees or some other angle.
They cost more but properly shielded transformers are best.
There's always some voodoo involved as well.
Try anything and everything before committing yourself to component placement.

 

[Kingston]

re: electrostatic radiation

got lost in translation. it's a field of course.

You can try to analyse them by deconstructing components in order to optimise which in my opinion is a fairly fruitless exercise, particularly in the context of 'one off units' or you can experiment to find the most stable arrangement.

But I come to this forum almost solely to analyse and deconstruct components in order to optimise! Any build I undertake nowadays is purely from engineering point of view. It's much more rewarding to design and then build even one-off units with a large production run in mind. Makes things run error free and stable much longer. And never again voodoo for me. Interactions are no longer complex, but controlled and known.

You know, instead of just hacking about everything together, praying that it works for at least a year and hoping the noise floor doesn't change when moon/sun alignment is different.

[edit]

I recommend playing around with VAC wires before any transformer orientation or relocation exercise. Do you have a power switch on the front panel, VAC heaters for tubes etc. These can be shielded on the cheap. Even with a couple of layers of that cheapo kitchen utility aluminum foil, but I don't recommend this for longevity reasons. For a simple test it's good enough.