I can’t really hear a noise floor with mine regardless, but the lamps and muting relays all work, ABX works. The input switch also triggers the relay. The issue is that I only have USB source, so it’s possible the issue is with that. I need to test AES.
That seems unlikely to me.
The start-up state machine in the FPGA waits until it gets about the right frequency from the crystal oscillator before starting anything else. The ECC81 is the active part of the crystal oscillator, it would be strange if the oscillator kept oscillating at about the right frequency without ECC81.
If there is still about -0.5 V DC between the test point of the crystal oscillator and the -300 V, that's direct evidence that the ECC81 still works normally.
The start-up state machine in the FPGA waits until it gets about the right frequency from the crystal oscillator before starting anything else. The ECC81 is the active part of the crystal oscillator, it would be strange if the oscillator kept oscillating at about the right frequency without ECC81.
If there is still about -0.5 V DC between the test point of the crystal oscillator and the -300 V, that's direct evidence that the ECC81 still works normally.
Last edited:
I mean the DC voltage across R52, between P3 and P4. If without retrimming, it is still between -0.4 V and -0.6 V, then the ECC81 does precisely what it used to do. If the voltage is negative, but not quite -0.4 V, then something has drifted. If there is no DC voltage at all, then the crystal oscillator doesn't work, which could be due to the ECC81 and various other things - and then it's a mystery why the FPGA module works as well as it does.
@SonnyMarrow When your valve DAC still worked, did you ever play back DSD files with the USB interface in raw DSD mode?
With raw DSD, it is much easier to determine up to where there is a useful signal than with PCM, hence the question. All you need is a makeshift low-pass filter to convert raw DSD into (poor quality) analogue audio.
I probably asked this before, but do you have an oscilloscope?
With raw DSD, it is much easier to determine up to where there is a useful signal than with PCM, hence the question. All you need is a makeshift low-pass filter to convert raw DSD into (poor quality) analogue audio.
I probably asked this before, but do you have an oscilloscope?
I've been working on a second Valve DAC build for a friend and I'm currently listening to it having nulled the noise using the trim pots - it's sounding good. Incidentally, the revised values of components around the noise trimming arrangement are much more effective than with my original build, when I had to mix up the valves in order to have sufficient adjustment to null the noise.
The build basically replicates my original build with respect to power supplies, a Maida regulated supply for the B-, regulated filament suply and low-noise regulated supplies for the digital circuits, the exception is the use of a JLSounds USBoverI2S on the input (instead of a Beaglebone with Pavel's reclocker on my original build).
I had one or two issues to solve but Marcel has been fantastic with his assistance as usual.
The build basically replicates my original build with respect to power supplies, a Maida regulated supply for the B-, regulated filament suply and low-noise regulated supplies for the digital circuits, the exception is the use of a JLSounds USBoverI2S on the input (instead of a Beaglebone with Pavel's reclocker on my original build).
I had one or two issues to solve but Marcel has been fantastic with his assistance as usual.
Somewhat disappointingly, when listening to the DAC yesterday it developed a low-level noise, like static, in one channel, though it still works and plays music through both channels. The noise can be heard in the background during quiet passages. The noise seems consistent and doesn't change when the trimpot is adjusted. I also noticed that it changes in level when I press on one of the valves (U3). There is also a change if I press on the adjustment knob of the trimpot but that may be about the internal mechanics of the pot? I have tried different valves in the socket and it makes no difference so I don't believe it is a valve problem. I'm guessing at the moment, but perhaps the thermal stresses from several power-up/down cycles have revealed a dry solder weld? Anyway, work to do.
Last edited:
After some investigation this morning it seems the problem was one of the 6922 valves - switching them into different places and substituting a spare helped me to isolate the one introducing the noise so I feel as though I have a firmer footing now to go forward with tuning up the DAC. I still have some tuning of noise to do though as I have found the valve-noise was masking a low-level warbling' type noise that I can hear on quiet passages - I'm working through different playback scenarios to see if I can isolate the specific conditions it occurs in as it seems to not be present on all playback.
If you turned them back and forth, ended up with the same setting as you started with but with improved sound, that's weird.
If you found that there is some optimum setting or range of settings, that's as expected. In fact it's why the trimming potmeters are there. They trim the delay of the data going into the DAC valves and should be set such that you have the least amount of artefacts in the noise floor. The valves then presumably work as return-to-zero DACs, like they are supposed to.
In either case, please check what happens at all DSD rates that will be used, particularly the highest DSD rate. The timing is most critical at the highest rate.
If you found that there is some optimum setting or range of settings, that's as expected. In fact it's why the trimming potmeters are there. They trim the delay of the data going into the DAC valves and should be set such that you have the least amount of artefacts in the noise floor. The valves then presumably work as return-to-zero DACs, like they are supposed to.
In either case, please check what happens at all DSD rates that will be used, particularly the highest DSD rate. The timing is most critical at the highest rate.