reducing hum in SET amps

[pjpoes]

I was wondering what sorts of things I could do to reduce the hum in my Meixing SET amp. Its an 805 type amp with 6l6 drive tubes and 12ax7 and 12au7 before that. It produces noticable hum, though I don't consider it intrusive. One thing I have read about is using dc for the filaments instead of AC, I haven't looked inside to see what this amp uses, but would assume ac. Would switching to DC fix some of the hum? What else can be done? Thanks.

[Joel]

Unless you're doing really low level signals, like phono, etc, AC filaments is not an issue with those tube types.

The two big culprits are elevated cathodes, and ground loops.

Elevated cathodes can be solved by raising the heaters a certain voltage above ground. Ground loops and chassis currents are more difficult to deal with - although adding an input transformer will completely eliminate that problem with one easy (somewhat costly) step.

I run all my amps and preamps fully balanced, with input and output transformers, and I no longer have any noise issues. They don't even make noise when I pull the input cables out and hot swap in other components.

Joel

[Klimon]

First take a listen and decide if the hum is 60hz or 120hz.

Hum from the power supply is 120hz and can be reduced by redesigning the power supply. Downside: first you'll have to know what you're doing and second, replacing by a ps with a choke (very effective for killing ps ripple) will probably be impossible without rebuilding the chassis (not enough space).

Second possible culprit is magnetic interference from the AC-carrying heater wires = 60hz. Those wires should be twisted very carefully (as many twists per inch as possible) and should never run parallel to other wires (be it signal or DC-carrying). This type of AC-interference can also generate other strange low-level noises.

Personally I wouldn't worry too much about the hum, unless it really bothers you e.g. if you have very high efficiency speakers or listen at very low volume levels.

though I don't consider it intrusive

Hmm. It seems to me that the tweak-fever has bitten you... Try swapping coupling caps if you can't keep your hands out of the amp

Simon

[pjpoes]

Yes the tweak fever has definatly bitten me and no I can't keep my hands out of the amp. It has an overall very nice design I think, plus I have always wanted a high power SET amp using a single output tube. 805's, 845's, and 211's have been on my wish list, so when I finaly bought this thing, it was one my dreams coming true for once. Overall I love the amp, I just feel happier when I listen to music now.

I think the noise is 120hz not 60hz, but to be honost, I am not sure. I will listen to some test tones at those frequencies to get a better handle on which I think it is. Oh and my speakers are 92db efficient at 1 watt, so not bad, but by no means a high efficiency design.

I would like to replace the power supply caps, which are currently a bank of small caps rather than a few large ones, which I generally prefer. I want to replace all the film caps with better ones, some of the signal path wire with better wire, etc. Also, though the binding posts look very sturdy, they are very poorly made insulated types that have stripped a little, so I may replace those with better ones. Is it worth maybe replacing all the DC lines with sheilded wire? Also I think the amp uses choke regulation, but could be wrong. I will ook sometime.

[billr]

most of the other guys have covered this, but the main issues are:

1. earth loops, star point earthing is good, also, look at ground lifting resistors to the mains ground.

2. SE amps are very sensitive to ripple on the HT, two stage filtering is a good idea. I've been going through valve amps by morgan jones again, cant recc this enough i have issues 1 & 3, anyway, he has a Choke substitute that uses a couple of transistors, one BJT and one MOSFET, and they pull the ripple down by 100db.

3. seriously consider dc heaters too.

4. mount the output and power trannies at right angles to each other to minimise inductive coupling.

5. make signal and power lines cross each other at right angles,

6. keep heater wiring in the corners of the chassis.

can't think of anything else.

sorry if i am repeating anything that has come up elsewhere

bill

[Klimon]

Is it worth maybe replacing all the DC lines with sheilded wire?

I wouldn't bother (never saw this solution either). First try to find out if it's the B+ rail or filament lines that are causing the hum. I consider some hum very normal; transistor amps have more hum/noise than tube-amps with a good power supply and wire layout so it's not a tube-related issue inspite of 'public' opinion.

Simon

[Jeb-D.]

All good adviced, one thing I didn't see is the suggestion of using soft recovory diodes. Hexfred's are good, www.tubesandmore.com carries them. They have worked wonders for me before. But in anycase SET amps are prone to hum because there is no CMRR and with the low power outputs it causes you to use high sensitivity speakers. So you can't cheap on power supply filtering, Soft recovory diodes, capacitor ,inductor, capacitor (C,L,C) scheme works for me. For low power stages capacitor, resistor, capacitor (C,R,C) works fine too. Also using constant current sources as plate loads for the input/driver stages, kills hum in that stage, even if you have a 10V ripple on the supply rail.

This all assumes there is no ground loops. Ground loops generally cause a pretty loud Hum. A faint hum is usually insufficient power supply filtering.

[cotdt]

would using differential balanced drive lower the hum problem?

[Eli Duttman]

Originally posted by Jeb-D.
All good adviced, one thing I didn't see is the suggestion of using soft recovory diodes. Hexfred's are good, www.tubesandmore.com carries them. They have worked wonders for me before. But in anycase SET amps are prone to hum because there is no CMRR and with the low power outputs it causes you to use high sensitivity speakers. So you can't cheap on power supply filtering, Soft recovory diodes, capacitor ,inductor, capacitor (C,L,C) scheme works for me. For low power stages capacitor, resistor, capacitor (C,R,C) works fine too. Also using constant current sources as plate loads for the input/driver stages, kills hum in that stage, even if you have a 10V ripple on the supply rail.

This all assumes there is no ground loops. Ground loops generally cause a pretty loud Hum. A faint hum is usually insufficient power supply filtering.

Jeb,

Once a choke is present in the B+ filter, even crappy 1N4007 diodes' switching noise is killed. The B+ will be clean. However, that NASTY switching noise can get into the amp via the power trafo and heater winding.

A cost effective SS solution is a combination of UFnnnn diodes and "Buddha" Camille's RRSF (reverse recovery spike filter). With a RRSF in place, even CRC filtration has proved satisfactory and the switching noise is kept out of the power trafo. Hop over to AA and scan the Bottlehead Forum archives for RRSF details.