Currently I'm building a replica of a 1960 stereo phono preamp - originally made by Shure to go with their magnetic cartridges.

The original unit is reputed to suffer from excessive hum levels. A stated hum and noise spec of -55dB WRT maximum rated output is pretty shabby by today's standards.

The circuit has a couple of 12AX7s with lots of negative feedback (good) but with unbypassed cathodes and AC on the heaters (not so good).

So my replica has the heaters powered from DC with the circuitry hidden under a stainless steel cover which also hides the modern power transformer.

Imagine my chagrin to find 10mV P-P of hum on the outputs!

I had used passive filtering for the heater supply, with the heaters in series and a ripple of about 7mV P-P on the 24 volts.

"I'll soon fix this" I thought, and added a linear regulator that reduced the heater ripple to zero, connecting the heaters for 12 volts now. The 7812 was quite happy to sink the extra 3 or 4 watts to the chassis.

BUT

Now my hum levels had increased to 25mV P-P on one channel, and 15mV P-P on the other!

It didn't take too long to confirm that the problem was actually the effect of the core leakage field from the E-I lamination transformer. And the reason it got worse was those extra 4 watts going into the regulator was hammering the transformer a bit harder.

Being a replica, I can't move the transformer any further away from the valves, but doing so temporarily certainly gets rid of the hum, so we know where it's coming from.

And, you guessed it, the channel with its valve closer to the transformer has the worst hum.

What would you do?

A copper flux shorting strap might help, still trying to lay my hands on some sheet copper to try it. A toroidal transformer won't fit in the available space.

I think I might have to fit a switched mode supply - 240 volts AC to 24 volts DC. It can hide nicely under the metal cover.

Perversely, the thing I though I'd have trouble with, the little MC34063 boost converter that takes the 24 volts up to 125 volts for the B+, has not been contributing any noise to the output, despite it running in (noisy) discontinuous mode because it's so lightly loaded.

It never quite turns out exactly as you planned!

As you have identified the transformer as the culprit, I guess some form of experimental (additional?) shielding may be worth trying.

Parts placement is often a design consideration, however, it seems that you are constrained by the need for originality, or proximity to that.

A photo of the chassis would be useful.

the reason it got worse was those extra 4 watts going into the regulator

If you reduce the power -- by whatever means -- is the hum level attenuated significantly?

I think I might have to fit a switched mode supply

You may end up with worse noise.

Switched mode PSU's and tubes can be a match made in hell.

Probably a good idea to see a circuit so that we can actually see what we are dealing with. Placing tubes within proximity of the transformers magnetic field is always, looking for trouble. That also applies to modern strong speaker magnets near transformers. Transformers tend to growl, get hot and create hum when overloaded.

One might consider a salvage transformer with a bit more attitude. One can always use an LR8 & TIP 50 to bring the volts down and act as a choke. You can put a smoothing cap before it. Lowering the capacity of C1 on a choke input filer will lower the output voltage.

An option that I have been forced to use when a speaker & transformer don't get on, as well as heat. Is a piece of sheet metal between the waring parties.

DC is a way to reduce hum from cathode leakage. I did find on an American set, that there were less issues with hum & better performance with Metal output tubes. It has a wholly transformer coupled push pull driver pair driving another transformer coupled push pull stage, with four 6L6's.

Hi Ian, three ways I have cut down valve hum from the preamp valves in guitar amps may be applicable.

1. find best orientation for the transformer, the leakage field varies in strength around the body.
Just turning the tranny at right angles may help.

2. Fit thick like 1/16" metal shields to valves.

3. Run heaters from DC.

Cheers, Fred.

The chassis material may be a factor.
Non magnetic stainless being used ?
JJ

Don't see the chassis being anything other than a corrosion, or bonding issue some SS can get a skin on it. Happens on a lot of rechargeable phone handsets, causing issues.

If you have a shielded transformer, its pretty much mandatory to ground the set a they can put a charge onto the chassis and has been known to cause hum.

On the older AC radios and phone lines we see "twisted pairs" & tubes with staggered feed on the heaters. e.g. in octals there was a tendency to have pin 2 grounded an on the next tube it would be hot. It is damn near impossible to solder to the SS with no iron any grounding should be with toothed washers and a dedicated earth with like used by Astor with Aluminium chasses.

Marc

I was meaning that the filament power transformer being mounted within a stainless chassis.
So, direct pickup by the valve or its filaments has not got that normal steel shielding and orientation effect.
From the magnetic fields produced by the transformer. See original post.
Nothing to do with corrosion or soldering ability.
JJ

Stainless steel can be soldered if a phosphoric acid flux is used. A common rust converter does the trick too. I have done this to make extra contacts for a voltage divider resistor.

Do be aware that there are different Stainless Steels. 2xx, 3xx and 4xx

I would suspect that chassis to be 302. 302 is an Austenitic stainless steel and is actually able to have a magnet attach to it. It's therefore only a minor bit lower than a steel one. Exhaust pipe is commonly another grade, viz 304. GP cutlery is 316 and magnetic whilst that marketed as marine grade 316 SS is not magnetic. SS is slightly lower in electrical conductivity. It is vulnerable to electrolytic corrosion with dissimilar metal and I have seen that in the chemical industry.

Could the chassis be made from mu metal which is a 75-80% Nickel material with very high magnetic field permeability which makes a good shield for magnetic fields? With this high nickel content it would look like stainless steel and would be nonmagnetic.

A lot of the automotive trim is martensitic 4xx which is more crystalline and the first one. Mu metal was common on Oscilloscope tubes.

I have used normal sheet metal as a flux shield.

Marc

Nickel is expensive and always has been - present price over US$20,000 per metric ton so the chassis might be marked as mumetal. Though on reflection, why make the whole chassis from it. A box around the transformer would be enough.

Magnetic fields can be a problem with torroidal transformers as well. EA in the 1980s designed the Playmaster Series 200 2x100watt stereo amplifier. They had to drop the ceramic cartridge option because of stray magnetic field. I built this amp, and hum was still perceptible on high volume settings with the magnetic cartridge option (would only hear the hum if there was no programme signal so not a problem for normal use). The power supply was built onto the bottom cover of the amp and if you removed this cover and moved it away from the rest of the amp - no hum.

A very good amp and still going well, with a remote volume control added. If you pick one up replace the dinky little 240v mains switch which blows up (literally) from start-up in-rush current taking out the speaker relay control transistor. I used both poles of a much meatier switch.

One does at times wonder at the sacrifices made and incredible placement of parts. If one wants hum you place the output transformer close to the mains transformer when it would have better off inside the pan. The same goes for the tubes placed so close to the mains transformer that its bad news. Even that the tube is internally shielded.

It actually goes from one extreme to the other. I had a 1930's AWA R301 Factory variant in for repair. It had a bigger transformer than the production model, an extra PP Transformer, & 6A6 plus another choke. At full volume no signal: That's what you heard, no signal: No hum.

Then one gets HMV chassis 42. 1st AF, Output & Rectifier as close to the transformer as was possible. The radiated heat alone was enough to cook the rubber jacketed primary & secondaries, wire to the point it had to be cut out.

On that one it ended up with a piece of 0.55mm sheet-metal between the three tubes and transformer, in an effort, not to act as a flux shield but to try & take some heat away from the wires.

Well, an external flux shorting strap made from 1.6mm copper sheet, lapped and soldered over 20mm made next to no difference.
Maybe about 1 or 2 dB.

The tiny improvement was more noticeable on the channel furthest from the transformer.

Lifting the transformer off the chassis does not help, so it's not magnetic field through the chassis. Not surprising, the inputs are wired with single point earth return to chassis to eliminate earth loops.

If you move the transformer 40mm away from the 12AX7 the problem goes away and all that's left is shot noise. If you ground the inputs most of that goes away too.

In the process of doing this I was having a phone conversation with the other stakeholder in this project. Phone connection was by WiFi.

It was fascinating to watch, with the phone less than 60cm from the preamp, the way the WiFi avoids interference from microwave ovens by operating only within a certain part of the mains waveform. The activity spikes were solidly locked to the mains!

I have some switch mode supplies arriving any day now. Bugger the authenticity, there will be no (audible) magnetic field and it's hidden under a metal cover anyway. All it needs to supply is 24 volts DC at just under 200mA, the boost converter does the rest. Nice, clean, regulated, smooth DC and no 50Hz!

For the sake of the audiophools, I will try a small toroidal transformer. I'm not expecting a complete elimination of the hum with this.