End-bells on chokes and SE OTs

[Swedish Chef]

I have an amp in for repair at the moment which is used by a touring band, and on the road at some point, the choke has been bashed and the ears of its clamp have sheared off.

On writing to the transformer manufacturer to request a replacement clamp I was told that the choke would be rendered useless as soon as I took the clamp off. The clamp had already been removed and as the manufacturer had kindly fitted the choke with what I can only describe as 'keepers' and done an admirable job of varnishing the whole thing, it was sturdy enough to stay happily together until I could bolt through the lams and secure it.

After a few more emails, the up-shot of which was, I had ruined the choke, I canibalised an old SE OT that I had lying around and used its end-bells (with mounting feet 2 per side) to re-fit the choke to the amp, and it is much sturdier and more able to withstand size 11 steel-toes or whatever belted it before.

You'll be surprised to hear that the manufacturer then told me that they do not use end-bells on chokes or SE OTs as:

FYI end bells can interfere with the field and inductance on chokes and
single ended gapped output transformers.
That is why we do not use them.

I appreciate that the gap is to prevent the core from saturating due to the DC flowing in the windings, and if the bells are not isolated from the lams they will couple the two parts to a small extent, but surely a clamp will do a similar thing?

How much science is there in that statement?

I have both a huge choke and enormous SE OT both with end bells, from Hammond in a prototype amp and these sound amazing...

Perhaps I am missing something... :sad:

chef

 

[Larrchild]

Seems to act like a shorted single turn

Steel or any metal container will increase heating and primary current due to it acting as an external shorted turn. It will not decrease the hum field only divide it up around the perimeter of the metal container. Rated primary current must be keep correct if transformer is to be in a metal container. The use of nonmetallic containers is suggested.

http://www.electra-print.com/custompower.php

 

[CJ]

"end bells can interfere with the field and inductance on chokes and
single ended gapped output transformers"

By how much?
This is the key question.


So you do a blind test.
Have a friend move the bells around the field then place them around the choke or transformer and see if you can hear a difference.

Fringing can occur at a gap.

This is where the flux hits a road block, so it tries to go around the gap.
The flux will bend around the road block in order to get to the other side.
When you put an end bell on a choke, you can make it easier for the flux to jump the gap?
Maybe, I do not know.
I would want to see it on a scope or visually befor I made a conclusion.

Also, what frequency are we talkin about?
Flux will jump the gap easier at high freqs, but then your min iductance required is already way down by then because of 2 pi f L.
The f is so high, that the L can drop almost to nothing, which it does, because the core starts giving up at around 1 K, depending on the alloy.

Do they make aluminum end bells?
I will have to ask ol Doc Hoyer about this one, could be interesting, although there are a heck of a lot of transformers and chokes out there with endbells.

It might be a minor audio phool tweak, like taking the shields off your 12AX7a preamp tubes in the back of your Fender, because they deflect the beam, even though the are Al, ok whatever.

What I try to keep in mind is:

If the wire does not pass through the metal as a complete turn, then coupling of the magnetic field to the metal object shall be minimal.

If anybody argues, you just whip out Maxwell and Farady on them, that usually shuts them up.

At Line Level, most of the flux never leaves the center of the core, so outside influences are negligible.

I did notice a slight difference on the HA 100 X, with the 3 layered shield, plus the outside can. More Leakage Capacitance.
But I think all it did was move the peak down a bit.
Out of the audio range, but maybe an influence on upper harmonics of high notes, if anything.

 

[Swedish Chef]

Additional taps will make the transformer inefficient, decrease regulation and heat up, due to the winding space increase to via the tap leads, and this decouples the winding. If a variable voltage is needed there are other means.

http://www.electra-print.com/custompower.php

Am I missing something....

Probably.

CJ! Thanks for that exposition. I was leaning towards your 'audiophool tweak' opinion, but I would still be really interested to hear the opinions of others...

I ran the amp for a couple of days and it sounded fine and didn't smoke, so if the electrical 'performance' of the choke is less than 10% poorer, for this job, strength trumps finesse.

chef

 

[CJ]

One thing i forgot to mention is that usually the transformer coil/core assy. gets dipped without the end bells, then baked in an oven maybe overnight or a few hours, depending on the stuff used.
So when you bolt up the covers, they are pretty well insulated from the core.
You can use sleeves for the hardware if you are worried about it.

The taps he is talkng about are on big honkin transformers.
Usually audio types do not ever handle that much power, even the power trans only has to drive a 12BH7 at worst.

 

[PRR]

> How much science is there in that statement?

Too much.

Yes, the inductance will change. Almost anything changes inductance.

It won't change a lot.

If you want to sell end-bell iron, you test with bells and take a few turns on/off to hit your nominal spec.

And the exact inductance is not critical in OTs or power chokes. Only minimum impedance at a frequency.

The impedance will drop, typically a few percent, because the bell iron bridges the gap. Until you get near maximum rated current, then the bell iron saturates and hardly conducts flux.

If that un-named designer prefers not to use endbells, fine. That does not mean another designer building endbell iron is a scoundrel.

> Seems to act like a shorted single turn

No. It is parallel to the laminations and has a similar effect. A bit worse because mild iron is a better conductor than silicon iron. A bit better because it is not in the center leg.

The quote refers not to separate endbells but to a total-container. That does act as a shorted turn... but >99% of flux stays in the iron, <1% is available to excite the container. And it is a known design problem. It was very carefuly worked-out for radio IF coil cans. Canning your power iron is a simpler problem, because the iron restrains the flux, and because a little loss can be tolerated.

> Additional taps will make the transformer inefficient

Another case where a sharp designer "knows too much". It's true, but.... taps can be SO darn convenient that we will PAY for the extra space (marginally larger core) and accept the higher electric bill. He's making his best transformers. You are supporting a touring band. Another $5 and half-pound scandalizes him. For you, when needs change (town has 250V walls, or lead gitarist buys a wrong-impedance speaker), it saves the gig.

 

[Swedish Chef]

Thanks PRR!

So they are not splitting hairs, but for the-show-must-go-on factor, the end bells win.

Thanks so much for that in depth analysis! :thumb:

chef

 

[CJ]

Smaller audio transformers use a high perm material for shielding.
It is almost the same material as the laminations.
But it does not suck up a lot of power from the signal.

It does need to be insulated from the transformer with wax, paper and press board insulators, maybe some type of fiber board insulator for a core pad.
Here is a shot of a transformer, you can see the hi-perm material being used as shielding:

 

[Larrchild]

PRR:

> Seems to act like a shorted single turn

No. It is parallel to the laminations and has a similar effect. A bit worse because mild iron is a better conductor than silicon iron. A bit better because it is not in the center leg.

The quote refers not to separate endbells but to a total-container. That does act as a shorted turn... but >99% of flux stays in the iron, <1% is available to excite the container. And it is a known design problem. It was very carefuly worked-out for radio IF coil cans. Canning your power iron is a simpler problem, because the iron restrains the flux, and because a little loss can be tolerated.

Combined with CJ's comment about the core being dipped before assembly, are the fiber or plastic washers I see on certain transformer bolts to float the bells?

 

[PRR]

Through-bolts are big trouble, so get tubes and washers.

End-bells are fairly benign. Not no-effect, but it is an imperfect world and endbells (and their insulation) are part of dealing with imperfection. A tranny which stays on the chassis and burns-up without releasing fire is sometimes preferable to a more-perfect imperfect design.

 

[CJ]

You can use Stainless Steel core bolts like they do at the transformer shop.

Yes, the record player motor works by way of a shorted turn.
That large copper wire. Saturation, heat, it's all good.


There is capacitive coupling from the coil to the core, and then from the core to a case, if you use one.
The case, which if grounded, could bring in the high end a bit. on the freq chart.