power transformer placement question for knowledgeable EE guys

[darron]

hi!

i saw something, probably years ago, and PROBABLY by R.G. i've searched google and the forum and just can't find it again.

i don't have the prestigious qualifications to know the answers for this, so i'd like to borrow some pair-for education from you guys!



i'd like to use many transformers side-by-side to make something like the spider. http://www.geofex.com/article_folders/spyder/spyder.htm



so what is the best PLACEMENT? in everything that i find in google images power supplies always have them running side-by-side in the same direction. do i remember reading that they should be perpendicular to each other for hum cancellation? IE side-by-side but every second one rotated 90 degrees.

or maybe completely 180 degrees around?

also, in different configurations, would it possibly actually be better running every second one with an inverted phase?

this is all to be regulated, but i'd like to know the absolute best practice to begin with.


thanks (:

[R.G.]

For power transformers, it largely doesn't matter. Chances of getting meaningful hum reduction on a row of power transformers is about nil.

I believe that the article you read was referring to the orientation of transformers in a tube amp. In that case, you want the axis of the windings for power transformer, output transformer, and choke all mutually orthogonal.

It is possible to get some small reduction in hum pickup in power transformers by careful orientation. However, this will make little or no difference in the resulting hum on the output of the isolated power supplies you mention.  This is because the transformers are being fed the identical same input signal, and it's a sine wave. For a single sine, all linear combinations and phases result in the same single-frequency sine wave, possibly phase shifted and modified in amplitude. The magnetic fields of multiple small transformers are primarily a sine wave with some small third harmonic distortion from the shape of the B-H curves inside, so what you get is primarily sine wave, same frequency. So however the leakage is added or subtracted, it's all power-line frequency.

The bigger issue is that the magnetic leakage from one transformer has a hard time coupling into the core of another. This is an issue for audio signals where the human ear is evaluating things and can pick out signals in range of over 100db. But for power transformers, whatever is inside the core M-field is going to push current into a rectifier/filter arrangement. The rectifier-filter causes its own distortions on the incoming (primarily) sine wave power signal that are likely to be much bigger than those caused by magnetic coupling to adjacent transformers through leakage fields. Even then, notice that using a single transformer with multiple windings to derive many outputs has the secondaries with near-100% cross coupling; there is still not a hum problem other than that caused by the wire routing and filtering.

A bigger issue is coupling of the M-field of the rectifier pulses. The coils of the power transformer itself are mostly shielded by the transformer core arrangement. This is not perfect, but it is there. The current loop of the secondary going into rectifiers is not only not shielded, it can be on wires which are spread apart into a loop. The M-field of this loop is proportional to the current pulses, which are big, and spikey in nature. They have many harmonics which couple better through space. And they may even overlap with other wiring loops, forming an air-core transformer. If these wire loops happen to overlap loops of the output voltage wiring, then pulses at the rectifier frequency get transformed into the output wiring, and can cause a buzzing hum even if there is a regulator on the output.

As a result, the biggest reductions in hum from a setup with multiple transformers like this come from careful wire routing.

I'm simplifying heavily (this notice is for the fact-checkers who read my work  ), but for this setup, worry rather more about keeping the transformer secondary wiring to rectifiers in low-area loops and making sure the output wires do not overlap the secondary wiring.

[darron]

Thanks R.G.!

I was so sure I read a comment in response to a similar project, but they may have just been regurgitating misinformation.


Well, the good news is that makes my life so much easier! I'll copy the footprint topology pretty much just has the schematic lays it out: Transformer > rectifier > filter > regulator > filter. It will all be PCB so the lead dress will be minimal, not crossing and tight.

I've said it before. I personally don't see the need to always have isolated supplies in small setups, but I have friends who insist. This will be a satisfying project.


Thanks for keeping your answer simple R.G. .... sorta

[Rob Strand]

If you keep pair of transformers side by side and close together, then reverse the mains side connection of one transformer I suspect you can get some hum reduction of the type when your guitar is in the vicinity of the transformers.   This type of hum cancellation is similar to humbucking.

I seemed to recall it working.   It's easy do the test with two transformers yourself!  Maybe try a single coil pickup.

[darron]

If you keep pair of transformers side by side and close together, then reverse the mains side connection of one transformer I suspect you can get some hum reduction of the type when your guitar is in the vicinity of the transformers.   This type of hum cancellation is similar to humbucking.

I seemed to recall it working.   It's easy do the test with two transformers yourself!  Maybe try a single coil pickup.

that's what i meant by possibly using a phase inversion.

i only have stupidly hot and quiet neodymium pickups here hehe :S


i half finished a pcb layout today, will only fit in a 1590BB practically. i might have to experiment now then before i finish it off, might need to adjust phases... might not make a difference too...

thanks (: