Sourcing shielded power transformers

[Deactivated-1]

I need a power supply for a tube preamp, 40VA with a 250VAC @ 100mA HV winding.
but I can relax the current requirement if there's a great option.

are torroidial transformers better for not putting out EMI?

[jonpaul]

No mains trsf is 'putting out" EMI.

The P-S capoacitance determins the passage of CM noise from mains to secondary.

Shielded trsf have a Faraday shield.
See the many fine books on noise reduction like Ott.



Toroidal trsf mfg indeed make shielded and double shielded.

see Pliotron

Expect long lead delivery and ~ $75..200 each

Jon

j

[jonpaul]

Our experience with SMD EER pulse/transmission transformers

Cps no shield 6..30 pF
Cps shield 0.3..3 pF

CMRR difference 10 MHz HP 4195A New Spec analyz)

22..36 dB improvement with shield

Just my experience

Mains isolation with shield or super low capaciatnace:
Furman Sound
https://furmanpower.com/discrete-symmetrically-balanced-power/

Topaz Ultra-Isolation transformers
https://www.surplussales.com//item/_tp/91905-68.html

OEM toroidals with shield: 100s of mfg, Pliotron (?)

Bon Chance

Jon


Jon

[RoGeorge]

are torroidial transformers better

Yes, toroidal transformers are much, much better then the normal E+I transformers.

There can be two types of parasitic coupling from a transformer to the nearby circuits, one through the electric field (capacitive coupling), the other type is through magnetic fields (induction in the nearby wires).

Against capacitive coupling, a simple metal shield or enclosure can help a lot.  However, the magnetic field will still escape through normal shielding, so special metal alloys have been developed to block the magnetic fields better, for example "Permalloy" metal, or "Mu-METAL".

Search for magnetic shielding transformers, or transformers shielding photos, and you'll find places that sells shielding enclosures for transformers.

A torroidal transformer might not need shielding at all, but the E+I usually need shielding, or at least sit the E+I transformer far away from the electronics, if the construction allows it.

I've seen even audio speakers dedicate to sound monitoring with bad design (an E+I mains transformer sited in the same enclosure with the power amp for the speakers).  It was no other way to eliminate the mains hum then to remove the transformer from the amplifier box, and sit it in a separate box on the floor (the hum was caused by the magnetic fields escaping from the E+I transformer).

No idea how bad the stray magnetic fields would be for tubes, but it can only be worst than for semiconductors, because of the direct interaction of the magnetic fields with all the flying electrons over long distances (inside tubes).

[jonpaul]

The issues discussed re toroid/EI magnetics and shielded transformers were researched and solved decades ago.

 books:   Noise Reduction Techniques In Electronic Systems By Henry W Ott
Electromagnetic Compossibility, Heinz M. Schlicke

1/ Magnetic: Gaps , winding technique and leakage flux are the cause of external magnetic fields. Toroid's are best to reduce magnetic leakage but the winding technique is of importance.

EI and EE bobbin transformers not as good but high quality fabrication and good design   for audio eg core faces are lapped, the laminations interleaved and a copper shading band applied.

Even in 1960s this was well known and used in high end  hifi like Marantz, HH Scott Fischer, etc.

The external MuMetal like Perfection Mica Netic and CoNetic are very high perm foils an sheets, used to divert the flux.

Those magnetic shields are VERY costly and connot be cut, soldered, or folded without precautions.


2/ Electrostatic: The common mode noise is a function of the P-S capacitance (interwinding) Use an interwinding Faraday shield to intercept the CM current flowing from Pri to Sec thru the insulation.

See my SMPTE/AES/NAB papers.

The CM current is diverted to return or earth rather than flowing from P to S.

In general interference reduction an noise

See the fine work of Jim Brown at http://audiosystemsgroup.com/

Also fine audio transformer app notes like Jensen Transformer.

Bon chance

Jon








http://audiosystemsgroup.com/

[Terry Bites]

Good layout wiring should reduce the need for a shield. Better to make your circuit pick up immune than wasting cash on a phile transformer.
 You can use an insulated spiral of mu-metal on a torroid. In fact any metal will work ferromagnetic or not.
Use punched to get air flow.

Gold plated mesh back lit with blue LEDs will improve the quality no end, allegedly.

[jonpaul]

A thin foil wrap of low u iron is not effective

Performance in db attenuaton depending on flux densité and freqency

Very high perméabilité Netic and Co Netic high Ni oriented are the best, see fine application notés from Perfection Mica.

Plots of attenuation vs flux density and frequency

Jon

[TimFox]

One source for toroidal power transformers, designed for vacuum-tube purposes, is AnTek.
https://www.antekinc.com/power-transformers/
I have dealt with them only once, but their delivery was prompt.

[TheMG]

Plenty of tube amps out there without fancy magnetically shielded transformers that don't have problems with hum.

A lot of it comes down to careful circuit design and construction technique. Small things, like ensuring the outer foil of film capacitors (unfortunately not marked on modern capacitors) is connected to circuit ground, minimizing lead lengths, careful routing of AC power wiring, placement and orientation of components, etc. It all can make a big difference.

[TheMG]

Toroidal transformers would have the lowest magnetic flux leakage without going too fancy/$$$$$

But I would also consider an SMPS as much of the EMI would be well outside the audio range.

[Haldor]

Toroidal transformers contain the vast majority of magnetic fields within the core which means they are much less likely to cause electrical interference with other circuitry. You could also just segregate your wiring to keep it away from the power transformer. That and use differential wiring and add noise filters (common mode chokes) to sensitive signals that must be routed near power transformer. Noise rejection is why pro audio equipment normally uses differential signaling.

Shielded transformers are more commonly found in signal isolation application than for power. Like these for example.

https://www.jensen-transformers.com/transformers/direct-box/

[coppercone2]

I measured magnetic field around a EI core and there was not much, the screws had a little nearby (1 Gauss), they concentrated flux on otherwise aluminum brackets.

It seemed rather well contained. If you have very sensitive circuits nearby it might be a problem. Don't put a Av=10000 against the transformer.

I think maybe if its loose and making alot of mechanical noise there might be fringing going on (there is fringing around the corners).

[jonpaul]

Coppercone,

How did you measure the flux, type, orientation, probe, meter, scope?

What was the geometry, and type , orientation, loading of the transformer?
Photos, sketch of setup, flux meter/ probe?
Magnetic flux difficult  to measure, position, area and frequency dependent.

In audio preamps, even milligauss of mains frequency flux can generate audible hum.
Even a toroid without a gap has a leakage flux.

Shielded power transformers are used in medical, instrumentation, high voltage
Double and triple shields are in so called super isolation transformers.

So far no one mentioned the  production solution in Hifi and instrument mains EI, EE laminated:
shading bands ( copper foil surrounding the entire winding/ bobbin and core) to reduce leakage flux
Most high end and pro audio amps, preamps in 1950s..1970s had such shading bands.
Easy to DIY! NO expensive hi mu foil!

All very old techniques,

Jon

[coppercone2]

I used a Hall effect probe made of a ic on a stick. Just waved it around the transformer like a wand rotating it’s orientation till I found the biggest number.

Forgot the Ic, it’s wrapped in heat shrink, but analog Hall effect sensor. I made a cable that hooks up to 4x nimh for 4.8v and another to read output voltage, or scope.

[uer166]

I measured magnetic field around a EI core and there was not much, the screws had a little nearby (1 Gauss), they concentrated flux on otherwise aluminum brackets.

It seemed rather well contained. If you have very sensitive circuits nearby it might be a problem. Don't put a Av=10000 against the transformer.

I think maybe if its loose and making alot of mechanical noise there might be fringing going on (there is fringing around the corners).

the fact remains that that "very little" is actually very much a problem.
if I have a power supply for a mic preamp, and the input coupling transformer has 1mV of induced hum (very optimistic if not miles away from the power transformer), and the preamp has 60dB of gain, then at the output you'll have 0dB of noise (1V)
in other words, the preamp would be useless.

even for an output transformer, -60dB of noise isnt ideal either, it would be a pretty bad preamp according to market standards.

Sounds like a crap design, improving PSRR of the sensitive circuit and making it more immune to outside influences would help more than trying to chase down the noise sources.

[coppercone2]

It’s a 49e sensor from digikey. Plenty of counterfeits on ebay.

It’s a bootleg setup just fishing for the highest number wherever you can find it.

But I think that is useful to know that I at least measured a useful number to be 1
“A few gauss”

[jonpaul]

waveform of Hall output?

[coppercone2]

maybe later because scope is repairing something right now, and I am not sure how well the home scope will do this over work, but this is relevant for a 6$ gamble
https://softsolder.com/2013/01/23/hall-effect-current-sensor-magnetic-flux-calibration/

https://softsolder.com/2013/01/24/hall-effect-sensors-from-ebay-variations-on-a-specification/

Mine is a 100nF and 4.7uF on the end of a wiring harness with four banana plugs, two grounds go to hall effect sensor ground, and power and signal wire come off two twisted pairs with respective ground, meant for a banana jack altoids battery box or a lab power supply (so it is star ground on sensor).

[Neomys Sapiens]

OEM toroidals with shield: 100s of mfg, Pliotron (?)

Formerly Plitron Pliotron, now Noratel
https://www.noratel.com/home/

Maybe youv'e read too much paleontology, where there's a clade of water reptilians called Pliosaurians?
Don't connect any of them into your circuits, they are very likely going 'CHOMP' on you!

[jonpaul]

Neomys, mille mercis pour votre information

Used Plitron isolation toroids 120/240v 1..5 kW in 1990s for testing and design of balanced power/ transient protection for Furman Sound.   So, the application was professional audio, cinéma sound, labs,  to isolate and balance mains power to a system,  usually on a rack.


The transformers are well made and highly Shielded, fine that the brand survives!

Jon


Jon

[jonpaul]

Rebonjour coppercone,

1/ thanks for  the photos.

 A DIY   cannot give a useable measurement unless calibrated, not a simple task.

I cannot  comment on the flux measurement without the  Hall sensor sec sheet, waveform of observed flux, distance and orientation of sensor to transformer.
A sketch or photo of entire setup, transformer, sensor with a ruler or tape to see the distance is needed.

Finally,  many commercial  flux meters exist, some very low cost.


Best Regards,

Jon

[coppercone2]

https://www.farnell.com/datasheets/2007294.pdf

yes the measurement is not worth while because the span is almost 70% with the data sheet, but it does still show you that things like bolts are flux concentrators, and that there is more energy near the corners of a transformer (you can still do relative tests compared to a R core or doughnut).

a zero Gauss chamber to find a null and then using the earths magnetic field might be the easiest way to determine sensitivity.

still useful to see if there was an effect before and after with something like a shield or component being installed.

[Sengcid]

Hello,

The main problem I have experienced is stray flux – mainly third and higher odd harmonics (150/180Hz is easily audible) due to high flux density (approaching saturation - not enough iron and/or turns). Toroidal transformers are definitely better than rectangular (E + I) types.

Stray flux can be mapped with a simple search coil – ten turns of thin wire about 2-3cm diameter, input to an amplifier/oscilloscope/spectrum analyser/headphones. I have one held together with cotton thread (used a darning needle to thread through and around) then three layers of nail varnish and glued into the end of a plastic tube (ex. ballpoint).

One can reduce MMF and flux levels in the core by taking load current from a secondary. If I remember correctly the load must be inductive for the current to be in the correct phase. The penalty is more heat.

With machine wound toroidal transformers look out for axial flux (50 or 60Hz) due to net turns (windings applied in the same direction). Inductive voltage dividers and ratio transformers employ “No-Net Loop” and, even better, “Balanced No-Net-Loop” winding schemes to avoid this. A simple way to reduce this effect is to add the requisite number of turns around the outer circumference and connect in series with the primary.

Another common source of “hum” is even harmonics – radiated by the pulses of current when the bridge rectifier diodes conduct. The solution is to use tightly twisted pairs from the transformer to bridge rectifier and from bridge rectifier to the smoothing capacitor.

N.B. Ideally all pairs, whether current carrying or voltage sensing, should be twisted or coaxial.

The current pulses also have an impact on the heat dissipation due to the I-squared R rule. As a rule of thumb the pulses are ten times higher than the average (PSU DC output current) for one tenth of the time, resulting in ten-fold higher I2R losses compared to what one might expect.