Silly PT mounting question

[EFK]

Is there any REAL difference between mounting a laydown power transformer with the laminations tight to the chassis vs. mounting it with the lamination nuts between the laminations and chassis? Obviously, old amps can be found with both methods as well as new amps. I'm wondering if there is any electrical, magnetic or mechanical reason to do so one way vs. the other?

[R.G.]

Is there any REAL difference between mounting a laydown power transformer with the laminations tight to the chassis vs. mounting it with the lamination nuts between the laminations and chassis? Obviously, old amps can be found with both methods as well as new amps. I'm wondering if there is any electrical, magnetic or mechanical reason to do so one way vs. the other?

The real answer is that it depends on the details.

First - is the chassis steel or aluminum? Steel is ferromagnetic, which means it acts like a relative "short circuit" for magnetic fields to move through. It can carry a magnetic field a good distance away. Steel is 100s to 1000s of times a better "conductor" for M fields than air is. So steel helps any radiated M-field leakage from the PT become carried to sensitive circuits.

Aluminum on the other hand is nonmagnetic, so it does not preferentially carry M-field. In fact, being a good conductor of electricity, it damps the transmission of M-field by electrically eating the energy out of any alternating M-fields trying to get through it. So it reduces the spread of any M-field, and helps with shielding it, or forcing the M-fields to not cross it by the opposing action of the eddy currents. Air is non-magnetic, but nonconductive. M-field spreads as it will. But air conducts M-field not that well. The net of all this is that steel chassis can carry M-field to other parts of the chassis, aluminum helps damp it a bit in some orientations, and air makes it harder for the field to get to other places.

Putting the transformer laminations right on a steel chassis makes coupling M-field to the steel chassis as good as it can get, so it makes any field-carrying worse - or better. Whether it's good or bad depends on how far away sensitive stuff is and how well they couple to the M-field carried by the chassis.

Lifting the laminations off a steel chassis cuts some of the coupling of leaked field into the chassis. It makes an aluminum chassis not be as good a "shorting band" to cut emissions from it.

Second: how well does the receiving end pick up any M-field on the chassis? Magnetic pickup coupling to the chassis for signal wires is largely a question of whether there is a conductor loop and what the area inside the loop is. If signal-carrying wires are twisted pair loops, this is minimized. If there are big, wide loops near the chassis, coupling is maximized. Note that the wires coming out of the PT also form current carrying loops that can broadcast M-field into the air and chassis. These should be twisted tightly to avoid large loop areas. High current heater windings broadcast power line hum; rectifier windings broadcast "buzzy" pulses at 2x power line frequency.

The other issue is whether the OT picks up M-field from the chassis, or through air. In either case, putting the OT as far from the PT as possible makes this minimum, as does orienting its axis at right angles to the PT.

Third: there are thermal issues. Laying the laminations on the chassis with significant areas of transformer laminations flat on the chassis helps transfer heat out of the laminations, so it can help keep the PT cooler. This should NOT be significant, but most "designers" have no clue about thermal issues, so they might in fact easter-egg themselves into a position where the chassis cooling is what keeps the PT from melting down. Bad Design. Bad Design. No biscuit. Heat transfer is better through an aluminum chassis than through a steel one, as aluminum is more thermally conductive than steel. Note that it would be even better if heat sink goo was put between them if that was really needed. Still - bad Design, no biscuit.

Suspending the PT one nut-height above the chassis introduces an opening which lets air rise from below by chimney-effect. I don't know whether lamination heat conduction or air rising is better. If that little change matters, or if the heat removal by chassis conduction matters, the design is IMHO too fragile for the real world, but it is a possible effect.

Fourth: Eddy current losses: The mounting bolts in a stacked-lamination transformer need to be insulated from the laminations, because they can form a conductive loop which acts like a shorted secondary for part of the M-field in the core. Tying the core to the chassis needs to be done in a way which leaves them insulated, however it's done. It may be an advantage to not muck with the insulation on the bolts by removing the nuts, inserting it into the chassis, then retightening. That's one way to introduce a shorted turn with the mounting bolts/nuts if you're not careful.

Fifth: mechanically, whether the PT is flat on the chassis or not makes little difference other than it sticks up a bit higher.

There's probably other stuff I'm forgetting.

I would not bother removing and replacing nuts on the lamination bolts in the transformer to bolt the lams flat to the chassis unless the bolts were too short to take the additional nuts, or unless the transformer already came with a setup that needed that. I'd probably put longer bots in if the were too short, and if I tinkered with them, I'd check for lamination shorts. If it hums, I'd check the X-Y-Z orientation of the tongues of the PT, choke and OT to make sure they're orthogonal, and that my wire dress and twisting was right, etc, etc. If it got too hot, I'd worry about the whole design and whether I'd specified too small a PT.

[EFK]

Thank you! That was one well-written, easily-comprehensible and extremely informative answer! I very much appreciate it.

[defaced]

Fifth: mechanically, whether the PT is flat on the chassis or not makes little difference other than it sticks up a bit higher.

Funny you bring this up. I was helping a guy on a different forum pick a chassis for a small amp build. He was asking about the new(ish) Hammond sheet metal chassis and the differences in strength between aluminum and steel, both 0.040" thick, and which one would be better for his build. I asked about the transformers he was using and upon looking at them, noticed something.

A transformer laying down where half of one end bell is through the chassis mounts with a wide foot print and a low center of gravity. Basically, when bounced around, it exerts less stress on the chassis than a transformer standing upright which typically has a narrower footprint and a higher center of gravity relative to the chassis mounting face. Significant? Depends on the chassis, iron, and load conditions. But if you already know you're up against something like a marginal chassis, it might be worth considering a flat transformer orientation. For a build that needs to survive true rigors, I'd just get a thicker chassis or change materials to a stronger alloy.

[R.G.]

If I was worried about the stress of the transformer weight on the chassis, I would get myself some aluminum or steel angle, cut off a couple of chunks and fit them under the chassis so that the transformer bolted them to the underside of the chassis, and also drill a couple of holes for bolts and nuts outside the PT outline. This would distribute the stress (a) outwards from the PT mounting holes right at the mounting holes and (b) away from the transformer mountings to the rest of the chassis, making the chassis much stiffer and less subject to bending from PT weight.

[Gregg]

R.G. nice and informative explanation as usual.
However I haven't seen or heard of a single issue flat mounting PTs on the chassis. Theory is good thing but practice is even better.
Also I've seen PT's that have the laminations welded/shorted working long time without any problems whatsoever for example this one:

http://marshallamps.com/product.asp?...pageType=SEEIT

Bottom line is you can flat mount a PTs and tighten it to a steel chassis with steel screws shorting all laminations and it will work without any problems.

[R.G.]

However I haven't seen or heard of a single issue flat mounting PTs on the chassis. Theory is good thing but practice is even better.

Sure - it works fine to do it that way. I probably should have been clearer: it works fine to put them flat on the chassis as long as your circuit is not otherwise sensitive to chassis conducted M-field. It is (IMHO) not fine to **count** on that to solve thermal problems. And it can cause eddy current losses, depending on the construction of the transformer core.

Also I've seen PT's that have the laminations welded/shorted working long time without any problems whatsoever for example this one:

Yep. That *can* work. Welding the lams across one side is a common thing to ensure conductivity to all the lams. But that weld doesn't form a conductive loop. And even when bolts are shorting, it doesn't necessarily kill things with an eddy current loop. But it can, depending on the stacking and jogging the lams got. But insulating the bolts always works.

I've seen people live to over 100years. However, while I can hope, I don't use the centenarians as a gauge for where I'm likely to get. In fact, given my demonstrated luck, I'm likely to be on the short end of the distribution.

Bottom line is you can flat mount a PTs and tighten it to a steel chassis with steel screws shorting all laminations and it will work without any problems.

I would say it as: you can flat mount a PTs and tighten it to a steel chassis with steel screws shorting all laminations and it MIGHT work without any problems. There also may be a counterexample that will throw you for a loop. Knowing what might happen helps if the details stack up against you.

[NateS]

If I was worried about the stress of the transformer weight on the chassis, I would get myself some aluminum or steel angle, cut off a couple of chunks and fit them under the chassis so that the transformer bolted them to the underside of the chassis, and also drill a couple of holes for bolts and nuts outside the PT outline. This would distribute the stress (a) outwards from the PT mounting holes right at the mounting holes and (b) away from the transformer mountings to the rest of the chassis, making the chassis much stiffer and less subject to bending from PT weight.

Was looking for info on this very topic. Changed horses in midstream and decided to build a full up JCM 800 / 2204 clone on the aluminum smallbox chassis I intended to build a 15w EL-84 2204 variant on. It's going in a head, and I think the chassis is marginal for the transformers for the 50W'er. Since this is going in a head, my plan was to build up a spacer / stress distributor out of standoffs and circuit board, with spacers going clear to bottom. Something like: