Here are some photographs illustrating the gain of the DHT Pre-amp with the type 31 vacuum tube. The first photograph is the oscilloscope trace of the output from a 1.0V/1Khz sine wave input. The vertical scale is set at 1 volt/division. The volume control was all the way up for maximum gain. The gain is ~3.8V, which is exactly what the datasheet states! The next photograph shows the output trace from an input of a 1.0V sine wave at 300Hz, which shows a gain of 3.8V as well. The last photograph shows the output trace from an input of a 1.0V sine wave at 10kHz and shows a gain of 3.8V. The gain is uniform throughout the audio bandwidth at different input voltages and volume settings.
Here are some photos showing oscilloscope traces of noise from the pre-amp:
With inputs grounded and set to full gain (volume control all the way up). The vertical scale on the scope is set at 2mV/division. Based on the intensity of the phosphore trace, most of the energy is within ~3mV peak to peak. I would call this the noise floor of the pre-amp, which is not super quiet at all. The noise is mostly in the audio bandwidth. I looked at the noise after passing the signal through a low-pass filter (set at 30kHz) and the noise spectrum did not change. What is interesting is that I cannot hear any noise from the speakers at normal listening levels with pretty efficient speakers (97dB). I have to put my hear right up to the speaker to hear the noise. Also, there is not a significant difference in noise when the volume control is turned all the way down.
With inputs grounded and amp turned off. Not sure what to think, maybe just background noise. Anyway, not significant.
With inputs grounded and amp set to full gain, but no Mapleshade Grounded Brass Tube Halos. There is noticeable increase in noise, which you can hear at the speaker in the form of hum. I also tried full steel tube shields and the noise floor is the same as it is with the Grounded Brass Tube Halos.
Here is the output from the filament power supply (set at 2.0V for the 31). The vertical scale is 2mV/division. You can clearly see the 120Hz ripple (twice the mains frequency from a full bridge rectifier) along with some modulated noise. The ripple is ~2mV peak to peak, which gives ~60dB (2000mV/2mV=1000) of ripple reduction. I think most voltage regulators are going to give some ripple on the output. You could try some larger capacitors after the regulator to knock down the ripple even more. I have some ideas that I think will result in a significant reduction in the ripple (on the order of <0.1mV) and I will report the results in the near future. In the meantime, I tried a capacitor from the voltage divider to ground and the noise increased in the circuit. I also tried using a larger value capacitor (1000uF) at the output of the regulator (before the common-mode choke) and it did not make any difference. However, I added a 1000uF capacitor to the output (after the common-mode choke) and there was a major improvement in the ripple (measured at ~0.8mV, which is not bad). I would recommend using a 1000uF cap on the output, still bypassed with a 1uF or larger film cap.
Here is the output from the Glassware Audio PS-7 Power Supply. I calculated the 120Hz ripple to be ~0.035V (see Power Supply for DHT Pre-amp. The actual 120Hz ripple is ~32mV peak to peak, which is pretty close to the the calculated ripple! The second photograph is the output after the second RC stage (ripple reduction factor of 17). The calculated ripple is 2mV peak to peak and the actual ripple is ~1.5mV peak to peak.
The first photo shows the output after the LR8 regulator. No more ripple, but ~2mV of noise. The second photo shows the output after a 4.7uF electrolytic capacitor is placed from the voltage divider to ground. The noise is less than ~0.4mV peak to peak.
Here is the trace from the plate of the 31 without shielding. You can clearly see the ~33mV 60Hz signal with some noise modulation. My guess is the tube is picking up the mains frequency from the transformers.