Hum in homebuilt valve amp

[Radio Wrangler]

A copper band forming a shorted turn around the whole transformer does indeed act to confine flux to the transformer core. It's usually shaped to conform to the shape of the outside of the bobbin and its windings. This is why it's done.

They also work at excluding external fields.

So in this case, one could have been applied to the mains transformer, and a second one could have been added to the output transformer as well.

This is what I'd have suggested IF raising the transformers a little and insulating one end of every mounting bolt hadn't worked. The copper bands would have acted against mag field crossing the fresh air between the transformers, but not acted against the mounting bolt linkage via chassis currents. The mounting bolts being genuine turns added to the transformer and intercepting about 50% of the legitimate core flux.

But with the no-valve hum now negligible, the job's done. Copper bands would have been an alternative to the steel shield. These sorts of jobs can often be done in more than one way.

Davd

[AdrianH]

OK thank you for the answer.

Adrian

[Sideband]

Just to show what the 'big boys' do (in this case Orange), I have been safety-testing a 200 watt bass amplifier (4 x KT88) and one thing I have to do is check the transformer construction. Here, I'm just removing the shrouds and you notice that the through-bolts have plastic collars fitted. They haven't fitted insulators to the chassis....the transformers are bolted directly to the chassis so in this case just the insulating washers are enough to reduce hum (which is barely noticeable).

[Radio Wrangler]

The Orange valve amps were originally badged versions of Matamps. Matt was a thoroughly competent engineer and one of the good guys (and a regular customer of dad's).

David

[G6fylneil]

I've always assumed the through-bolt head insulation washers were to stop the bolts forming a shorted turn, on a transformer I have to hand, one bolt has a steel washer for earth continuity.

[Radio Wrangler]

Yes, a single one doing some earthing is a good idea.

Some HP instruments had transformers mounted to the rear panel with an end bell on the outside. The shorted turn issue had been found some time in the past and plastic bushes on the end of the screws going through the panel and bell were fitted. The folk on the line just fitted them without wondering why.

It was discovered that the bells and the screw heads didn't pass earth bonding tests when such tests came along with new safety standards.

Some instruments had 2-pack paint on the rear panel while the steel bells had a varnish on them. Even deleting the plastic bushes didn't create reliable earthing. I got pulled into the matter, worked out why the bushes were there, and found that in most cases, the paint and/or varnish provided insulation, but not reliably. So the bushes, well at least three of them had to stay because of circulating current issues, but one screw with plain stainless washers still didn't give reliable earthing. I ended up getting our own bells pressed out of aluminium and alochrome treated. I changed the rear panels to silk screen black text on bare alochrome. So the aluminium bell was directly onto the aluminium panel, just with our standard, conductive, alochrome finish on both surfaces. Saved the cost of painting the panel and made our family of instruments look more uniform. Transformer core earthing still wasn't certain because of varnish on the lams from vacuum impregnation. Bushes were still a good idea but could be done on the inside end of the mounting bolts. The transformers had mummy-wrapped primaries and an electrostatic screen, but most importantly they were on the inside and fully covered so didn't have to be earth bonding tested. The aluminium bells passed just fine.

We did leave one bolt without a bush, but that was usually not too important.

Earthing transformer laminations is a difficult thing. They do their work by breaking up circulating current paths transverse to the magnetic flux paths. They work by having insulating coatings on them, well, not very conductive surface treatments. So you can't rely on all the lams of a core being connected together. Electrically, there is no singular item as 'A core'. A crinkle washer went on the bolt sans bush. That theoretically earthed the lam on the inward face of the core. The lam on the outwards face of the core was pressed against the earthed, aluminium rear panel.

So to fix circulating current issues AND have all externally accessible parts meet an earth bond test requirement, the solution needing to be found is highly dependent on the arrangement of the transformer and how it's supported. No one solution can fit all the possibilities, but bits of solutions can be used... with care.

David

[kalee20]

That's a fascinating post, David - thanks!

[kalee20]

Originally Posted by GrimJosef

... the main cause of things going quiet when the amp warms up is the appearance of the negative feedback. The transformer-to-transformer hum is introduced inside the feedback loop, so it will be suppressed (by 26dB if I remember rightly in the Leak) when the amp's gain turns on.

If you are hunting for hum it's often worth disconnecting the NFB while you do it. It makes the problem much easier to hear and it doesn't cause its source to be confusingly relocated by the action of the feedback loop.

Cheers,

GJ

That's a subtle point, GJ! And shows the complexities of a feedback amplifier. It may be quiet with no external signal applied, but it's working all the same, as 'scoping the output valve grids would show.

Unsuspected signals like this could easily 'use up' some of the internal signal headroom and lead to overloading of intermediate stages, even when you think they shouldn't be.

It emphasizes your point (and Radio Wrangler's point in the Quad II thread) to get the amplifier free from problems before adding feedback.

[murphyv310]

Hi.
I had the very same issue with my Mullard 3-3 stereo amp build. Slight hum with the Rectifier valve removed but when inserted absolutely no hum at all. I later substituted the mains transformer for a one with higher heater current of a different format with full screening only to cure the issue. I had already deduced it was magnetic leakage but when warmed I thought wrongly it may have been cancellation between the induced hum and out of phase hum from the HT, so it seems I was wrong.