Humbucking Reverb Tank?

[jimboyogi]

Thanks Helmholtz. I'll have a look at the wring from reverb jack to gain stage grid.

I did some trials on the PS to see if I could reduce the buzzy noise pickup (pictured ^ thread a bit) by cleaning up the diode switching noise.

I tried -

Between first diode/PT secondary junction and ground, for both phases -
10nF and 1K in series - no change to noise signal voltage when reverb on and tank close to PT
10nF and 1.5K in series - no change as above (these are the values suggested by Merlin in his Hi Fi preamp book)
10nF and 2.2K in series - no change as above

Between the two PT secondary phases -
100nF cap - increased the noise signal substantially as measured on my oscilloscope!

So from these experiments I was not able to reduce the noise pickup. And the cap across the secondary windings made the noise pickup worse. Either my component values are not optimised for effective snubbing, or else I'm barking up the wrong tree!

[Helmholtz]

the noise signal on the oscilloscope reads the same value with the shorted jack with reverb ON at footswitch as with reverb OFF at foot switch without the short.

Is the footswitch cable (as well as the switch itself) well shielded? Also the amp internal wiring to the footswitch is critical.

[jimboyogi]

The foot switch cable is a microphone grade shielded cable. I just did the experiment of measuring the noise signal on my 'scope with the footswtich plugged into the chassis jack with Reverb ON, then unplugging the foot switch from the chassis (reverb is ON when footswitch cable unplugged). There was no difference in noise signal voltage or audible noise between these two modes. I think that this indicates that the foot switch is not contributing to the noise pickup.

I also measured the noise signal with reverb on, with and without a mu-metal shield around the reverb tank at the "output" transducer end. Without the mu-metal = 1.1Vpp. With Mu-metal = 0.4Vpp, So the mu-metal does provide a big shielding effect for the transducer.

WRT my theory of diode switching noise, I also scoped the HT rail in the amp at each of the filter capacitor nodes. There was no sign of any HF spikes on any of the nodes. So I think that this, combined with the failure to get any PS snubbers to improve the noise signal pickup as outlined in previous post, points to this diode noise theory being incorrect.

[Helmholtz]

It might make sense to look for external sources of magnetic interference.

You can use a high impedance reverb transducer coil (or a single coil PU) connected to the amp's input as a "sniffer".

[Lemonhead]

I also measured the noise signal with reverb on, with and without a mu-metal shield around the reverb tank at the "output" transducer end. Without the mu-metal = 1.1Vpp. With Mu-metal = 0.4Vpp, So the mu-metal does provide a big shielding effect for the transducer

Hello folks,

sorry to resurrect an old thread, but jimboyogi's heroic efforts are of particular interest to me. This is the first documented use, that I've seen, of Mu-Metal to shield a reverb spring tank's output transducer from magnetic PT AC hum in a guitar tube amp.

I'm having the same problem with a Princeton Reverb clone. The spring tank has nowhere to hide from the PT's field, so the only options I see are to change the PT and hope for a more forgiving field or try to shield the tank.

jimboyogi, can you please describe how did you go about shielding the transducer? How much material did you use? Was the attenuation enough to make the reverb more usable? Is there a specific part of the transducer (coil, magnets, lamination) that is responsible for the hum and can be targeted more precisely, or should every part treated the same? Any information would be of great help for the many people afflicted by noisy reverb.

Thank you.

[Enzo]

The output transducer can be sensitive to power transformer fields. I can see no way to shield the lamination without also doing the winding and so on. I mean the whole thing is the size of a garbanzo bean. Metal shielding over the open side of a reverb pan is nothing new. I have seen the whole side covered, and I have seen a plate that covers about half the open side, down at the output end.

[Helmholtz]

Completely covering the open side of the tank with a steel plate should work best.
The steel tank shell itself is a magntic shield but leaves the bottom unshielded.
Aluminum would have little effect.

[jimboyogi]

I can't find the notes I made (or may not have made!) about this, and the amp is long gone. From memory I used a 0.1T mu metal band, either 10 or 15cm wide, wrapped around the output transducer side of the tank, and slightly overlapped. I think I just used tape to secure it. It reduced the reverb emf noise to quite acceptable levels in that amp.

I did try covering the open side of the reverb tank with a steel plate, I can't remember how much effect it had on emf noise shielding. But I didn't end up using it on the finished amp, which suggests that it didn't help enough to be worthwhile in this case.

[jack briggs]

Completely covering the open side of the tank with a steel plate should work best.
The steel tank shell itself is a magntic shield but leaves the bottom unshielded.
Aluminum would have little effect.

What about copper? I'd think that would afford better shielding than steel, albeit more expensive.

cheers,

[Lemonhead]

Thank you for responding, everyone.

I tried to cover the open side of the tank with all sorts of metal objects found around the house (a wok, tin can, scrap steel plate, tin foil, etc.) and nothing even registered. This relative permeability table on Wikipedia shows that copper and aluminum have about the same effect as air. Permalloy/Mu-Metal have significantly higher permeability and can be wrapped around the transducer, but they are expensive. I have no way to measure the field and estimate chances of attenuation, but it can be an interesting experiment.

The plan is to use 8"x12" of Mu-Metal, wrapped around the transducer assembly inside the tank and then try to cover the entire open side with a plate of some kind.

[Helmholtz]

What about copper? I'd think that would afford better shielding than steel, albeit more expensive.

cheers,

LIke aluminum, copper is an excellent electrical conductor but a poor conductor for magnetic fields (doesn't stick to a magnet).
So these 2 metals can only provide shielding against electrical fields.

Low frequency magnetic shielding requires high permeability.

Mu-metal (actually µ-metal) is a nickel-iron alloy having extremely high permeability (= µ), so works best for magnetic shielding..

At high frequencies copper provides some magnetic shielding due to Eddy currents.

[Lemonhead]

Another material that could be useful is G-iron. It has a lower permeability compared with Mu-Metal, but higher saturation. Since I have no way to measure the field's intensity, I'm not sure which one of these materials is better suited to this specific task.

Does anyone know what should be the Gauss/Tesla reading of a tube amp PT at a distance of ~50cm?

[Helmholtz]

Another material that could be useful is G-iron. It has a lower permeability compared with Mu-Metal, but higher saturation. Since I have no way to measure the field's intensity, I'm not sure which one of these materials is better suited to this specific task.

Does anyone know what should be the Gauss/Tesla reading of a tube amp PT at a distance of ~50cm?

Forget saturation B. The field will be very small, probably milli-Gauss.
High permeability helps but typically regular magnetic steel (not stainless) is sufficient.

[Lemonhead]

Forget saturation B. The field will be very small, probably milli-Gauss.
High permeability helps but typically regular magnetic steel (not stainless) is sufficient.

OK, thanks. Just ordered some permalloy. Should be here in about a month, so I'll try to report then.