Hello--
I borrowed an oscilloscope from a friend recently. Actually, not much of an oscilloscope at all--a dinky DIY kit. Minimum sensitivity is 8 mV, maximum input voltage is only 50V. That leaves me able to test the signal outputs of my components. I've consulted a number of sources (Radiotron handbook, https://dalmura.com.au/static/Hum article.pdf, Neurochrome's articles, Eli Duttman's site), I'd like to bounce a few measurements and some questions off of anyone whose willing to lend some of their time.
On my integrated tube amplifier, for example, I measure 8-10mV pp with the volume knob at 0. So I am at the minimum measurable threshold of the oscilloscope. It shows a 120Hz waveform there. Through my ~88dB monitor speakers, I can hear this with my ear near a cone. It does not increase with volume. I can swap in other tubes and measure anywhere from 12-30mV pp with them.
When I get the volume control at about 75%, or 12 o'clock, a 60Hz waveform appears. This increases with volume and by 100% can bump the measured voltage to about 50-60 mV.
The 120 Hz waveform is the ripple in my rectified DC, correct? The amplifier is SE 2A3 transformer coupled output with two gain stages. Each 2A3 has its own DC supply.
So the point of the post is to try and understand the sources of the noise and how much I can expect to do about each.
Obviously the answer varies with speaker sensitivity, but within reason, what mV (or sub-millivolt?) measurement of the noise floor should I be aiming for?
Is the 120Hz waveform from the B+ going into the output transformer and/or the B+ ripple on the two gain stages and/or ripple from the filament DC?
With regard to the 60Hz noise, which I am 99% sure is from the input stages AC heaters (all indirectly heated triodes): Is it possible that a humdinger on the 6.3V transformer secondary (artificial) CT would be able to "null" this noise source? Obviously, switching from AC to DC supply for the heaters would eliminate that noise source, so perhaps it's best to pursue this approach. I am assuming that none of the 60Hz noise is coupled from, say, the heater or primary TX, but I have reason to believe that any 60Hz coupling is minimal. Both TX are toroids; one is mounted on top with an iron shield; the other is mounted underneath but very far away from OPT or signal circuitry.
We recently moved our audio room and discovered that in the new location, if we have our dining room dimmer-switch canister lights on, we can hear a very harsh buzzing noise in the amp. It's not a big deal for us to have the lights off in another room when we listen to our amplifier, but for my own edification, is this a noise source that adding a snubber to the power supply would mitigate?
Lastly: We live in an old house. Unless I run a line from our breaker box in the basement (which I did), there is no grounded 3-prong outlet available to us. Before I installed this electrical line, the noise floor of my components was much higher. On tubed gear, I could put my hand near input tubes and generate substantial hum in the output. What exactly causes this? Is this the potential voltage difference between myself (standing in my socks on a wooden floor; components mounted on a wooden shelf) and the voltages present in the amplifier interacting via capacitance? (I could be way off on this last point, but would really like to understand!)
Thanks for your time!
I borrowed an oscilloscope from a friend recently. Actually, not much of an oscilloscope at all--a dinky DIY kit. Minimum sensitivity is 8 mV, maximum input voltage is only 50V. That leaves me able to test the signal outputs of my components. I've consulted a number of sources (Radiotron handbook, https://dalmura.com.au/static/Hum article.pdf, Neurochrome's articles, Eli Duttman's site), I'd like to bounce a few measurements and some questions off of anyone whose willing to lend some of their time.
On my integrated tube amplifier, for example, I measure 8-10mV pp with the volume knob at 0. So I am at the minimum measurable threshold of the oscilloscope. It shows a 120Hz waveform there. Through my ~88dB monitor speakers, I can hear this with my ear near a cone. It does not increase with volume. I can swap in other tubes and measure anywhere from 12-30mV pp with them.
When I get the volume control at about 75%, or 12 o'clock, a 60Hz waveform appears. This increases with volume and by 100% can bump the measured voltage to about 50-60 mV.
The 120 Hz waveform is the ripple in my rectified DC, correct? The amplifier is SE 2A3 transformer coupled output with two gain stages. Each 2A3 has its own DC supply.
So the point of the post is to try and understand the sources of the noise and how much I can expect to do about each.
Obviously the answer varies with speaker sensitivity, but within reason, what mV (or sub-millivolt?) measurement of the noise floor should I be aiming for?
Is the 120Hz waveform from the B+ going into the output transformer and/or the B+ ripple on the two gain stages and/or ripple from the filament DC?
With regard to the 60Hz noise, which I am 99% sure is from the input stages AC heaters (all indirectly heated triodes): Is it possible that a humdinger on the 6.3V transformer secondary (artificial) CT would be able to "null" this noise source? Obviously, switching from AC to DC supply for the heaters would eliminate that noise source, so perhaps it's best to pursue this approach. I am assuming that none of the 60Hz noise is coupled from, say, the heater or primary TX, but I have reason to believe that any 60Hz coupling is minimal. Both TX are toroids; one is mounted on top with an iron shield; the other is mounted underneath but very far away from OPT or signal circuitry.
We recently moved our audio room and discovered that in the new location, if we have our dining room dimmer-switch canister lights on, we can hear a very harsh buzzing noise in the amp. It's not a big deal for us to have the lights off in another room when we listen to our amplifier, but for my own edification, is this a noise source that adding a snubber to the power supply would mitigate?
Lastly: We live in an old house. Unless I run a line from our breaker box in the basement (which I did), there is no grounded 3-prong outlet available to us. Before I installed this electrical line, the noise floor of my components was much higher. On tubed gear, I could put my hand near input tubes and generate substantial hum in the output. What exactly causes this? Is this the potential voltage difference between myself (standing in my socks on a wooden floor; components mounted on a wooden shelf) and the voltages present in the amplifier interacting via capacitance? (I could be way off on this last point, but would really like to understand!)
Thanks for your time!
Your body acts as an aerial ( antenna ) picking up your mains electricity and transmitting its signal to your input tubes via body capacitance .
In the UK my old website has 100,s of complaints about those light dimmers ,its well know and the cheaper they are the very low grade ( usually Chinese ) types have inadequate suppression of the pulses which --yes I and another checked out by pulling one apart -very low grade components.
I don't know where you live but its not the UK where an earth is compulsory--- if you don't live in a flat install an earth spike outside and if you live in a dry area ---keep it watered.
In the UK my old website has 100,s of complaints about those light dimmers ,its well know and the cheaper they are the very low grade ( usually Chinese ) types have inadequate suppression of the pulses which --yes I and another checked out by pulling one apart -very low grade components.
I don't know where you live but its not the UK where an earth is compulsory--- if you don't live in a flat install an earth spike outside and if you live in a dry area ---keep it watered.
The absolute value of the 60Hz and 120Hz ripple is, on itself, not a useable metric. What is important is how much 60Hz/120Hz there is relative to the signal. 10mV 60Hz on 10V signal is 60dB signal-to-noise, which is not bad for your average tube amp.
Jan
Jan