I appreciate it, but it's probably cheaper for me to do it. It's only $215 here in the US plus the $60 cable.
I'm planning on having the 300VDC supply, the 6.3VDC heater supply, and the 6.3VAC transformer for the digital components in a separate enclosure.
If I use a 6-pin umbilical to connect the PSUs to the DAC, can the 6.3VAC run with the DC lines? I considered doing the conversion to 5VDC off-board, but noticed that the 6.3VAC is needed for the relay section starting at R107 and R108.
If I use a 6-pin umbilical to connect the PSUs to the DAC, can the 6.3VAC run with the DC lines? I considered doing the conversion to 5VDC off-board, but noticed that the 6.3VAC is needed for the relay section starting at R107 and R108.
Yes, I did a quick calculation and I don't expect that combined umbilical to cause any problems.
Is a hybrid approach for the reconstruction filter with one passive LC stage and a stack of two active Linkwitz-Riley 2nd order filters (realized with unity gain Jfet buffers) out of the question? The LR are not exactly linear phase...
Speaking of the filter, has anyone in the US found where to get the ferrite cores? All the other components to assemble the inductor are available, but I cannot find P26x16 250 nH/turn^2 cores. The recommended ones are out of stock everywhere.
Indeed, I see Digi-Key now has a minimum order quantity of 600 for them, and a lead time of 12 weeks...
If you can't get them anywhere, you could take the 630 nH/turn^2 N48 version instead, B65671T0630A048, and reduce the numbers of turns to 63 % of the original values (sqrt(250/630) ~= 0.63).
If you can't get them anywhere, you could take the 630 nH/turn^2 N48 version instead, B65671T0630A048, and reduce the numbers of turns to 63 % of the original values (sqrt(250/630) ~= 0.63).
Last edited:
So that would mean 35.595, 29.925, and 16.38 turns. How precise do these need to be? Would 35.5, 30, and 16.5 be good enough?
Is a hybrid approach for the reconstruction filter with one passive LC stage and a stack of two active Linkwitz-Riley 2nd order filters (realized with unity gain Jfet buffers) out of the question? The LR are not exactly linear phase...
A combination of passive and active filters, starting with a passive one, should work fine. Why Linkwitz-Riley?
Marcel,
Do the dac output caps need to be as large as 10uF? That is serious $$ using my normal caps. What would the consequences be of say 2.2uF instead?
Then I have to think about the reconstruction filter and cabling. Preamp is a 10Y DHT (with SS mu-follower) with a 47k input resistor (no cap) but has a cap and autoformer at output (band limiting to some degree I assume). 75r BNC cabling will be used between dac and preamp but length will be 6.5m. I may end up using custom inductors and OPT's (Dave Slagle) so can have the OPT wound however is needed perhaps even to alter the primary load. Saying that, for most recordings, a nominal 1Vrms output from the dac is needed to get my music nice and loud, so we can't play around too much with the output level from the ValveDac which I think I read somewhere is about 1.25Vrms.
So, with that info, any advice you are able to offer on integrating the ValveDac into my system would be appreciated, particularly in terms of output caps, reconstruction filter and output transformers.
Anthony
Do the dac output caps need to be as large as 10uF? That is serious $$ using my normal caps. What would the consequences be of say 2.2uF instead?
Then I have to think about the reconstruction filter and cabling. Preamp is a 10Y DHT (with SS mu-follower) with a 47k input resistor (no cap) but has a cap and autoformer at output (band limiting to some degree I assume). 75r BNC cabling will be used between dac and preamp but length will be 6.5m. I may end up using custom inductors and OPT's (Dave Slagle) so can have the OPT wound however is needed perhaps even to alter the primary load. Saying that, for most recordings, a nominal 1Vrms output from the dac is needed to get my music nice and loud, so we can't play around too much with the output level from the ValveDac which I think I read somewhere is about 1.25Vrms.
So, with that info, any advice you are able to offer on integrating the ValveDac into my system would be appreciated, particularly in terms of output caps, reconstruction filter and output transformers.
Anthony
A combination of passive and active filters, starting with a passive one, should work fine. Why Linkwitz-Riley?
I was already half asleep when I typed that... What I meant was Sallen Key, which however still raises questions regarding phase response. The Sallen Key topology is simply what first came to my mind, as they're easy to implement with series Rs (say 3k5) and shunt Cs (1nF) and a Jfet buffer. I'm not entirely sure how well they're suited for this application.
You can design a sixth-order filter that has one pole pair made with an LC section and the other two pole pairs with Sallen and Key sections. I don't see any issues as long as the first filter section is the passive LC section.
Marcel,
Do the dac output caps need to be as large as 10uF? That is serious $$ using my normal caps. What would the consequences be of say 2.2uF instead?
A first-order roll-off below 12.1 Hz (instead of the normal 2.66 Hz) and a some aggravation of the noise modulation artefact for frequencies below 193 Hz. I guess you have to try it to find out whether this bothers you.
Then I have to think about the reconstruction filter and cabling. Preamp is a 10Y DHT (with SS mu-follower) with a 47k input resistor (no cap) but has a cap and autoformer at output (band limiting to some degree I assume). 75r BNC cabling will be used between dac and preamp but length will be 6.5m. I may end up using custom inductors and OPT's (Dave Slagle) so can have the OPT wound however is needed perhaps even to alter the primary load. Saying that, for most recordings, a nominal 1Vrms output from the dac is needed to get my music nice and loud, so we can't play around too much with the output level from the ValveDac which I think I read somewhere is about 1.25Vrms.
So, with that info, any advice you are able to offer on integrating the ValveDac into my system would be appreciated, particularly in terms of output caps, reconstruction filter and output transformers.
Anthony
75 ohm coaxial cable has a relatively low capacitance per unit length, so I don't expect the cable to be a big issue. The level is about 1 V RMS when you account for the transformer losses.
some aggravation of the noise modulation artefact for frequencies below 193 Hz
I am not sure what this is...is this to do with the dac architecture or is this an in-general situation for a 2.2uF output cap against a 5950ohm Rp like in any amplifier? To compare, what is this number with the 10uF output caps?
The noise of the valve DAC gradually increases with increasing signal level, see section 13 of https://linearaudio.net/sites/linearaudio.net/files/03 Didden LA V13 mvdg.pdf Normally this gets worse at signal frequencies below 42.4 Hz, with a smaller capacitor that will shift to 193 Hz.
Ok, thanks Marcel, 10uF it is. I assume that a larger value will improve LF noise performance a little. (ordering lots of parts at the moment)
Hi Marcel,
Sorry to bother. Are any of the resistors in locations which are sensitive to current noise? I was toying with the idea of utilizing some bulk foil resistors for anything related to the signal path or clock lines, but thought I'd ask rather than swap things out for the sake of it.
-Raja
Sorry to bother. Are any of the resistors in locations which are sensitive to current noise? I was toying with the idea of utilizing some bulk foil resistors for anything related to the signal path or clock lines, but thought I'd ask rather than swap things out for the sake of it.
-Raja
I'm sorry I'm asking so many questions, but I don't want to order the incorrect parts.
If I used these 400 nH cores: https://www.mouser.com/ProductDetail/EPCOS-TDK/B65671D0400A048/?qs=/sLciWRBLmCLGHayV0aCBQ==
I calculated ~79.06% from the original turn count, which would mean 44.667, 37.552, and 20.555 turns respectively? Using 0.6mm wire? These turn counts seem a little better than with the other cores because they are pretty close to half-turns.
If I used these 400 nH cores: https://www.mouser.com/ProductDetail/EPCOS-TDK/B65671D0400A048/?qs=/sLciWRBLmCLGHayV0aCBQ==
I calculated ~79.06% from the original turn count, which would mean 44.667, 37.552, and 20.555 turns respectively? Using 0.6mm wire? These turn counts seem a little better than with the other cores because they are pretty close to half-turns.
Good idea, those are more suitable than the ones I found on Digi-Key (those from Mouser have a bigger air gap, are closer to the original, and use the same ferrite material).
Hi Marcel,
Sorry to bother. Are any of the resistors in locations which are sensitive to current noise? I was toying with the idea of utilizing some bulk foil resistors for anything related to the signal path or clock lines, but thought I'd ask rather than swap things out for the sake of it.
-Raja
Hi Raja,
I've marked the resistors that I expect to be sensitive to 1/f noise(*) with "MF" or "metal film" on the schematic.
Regards,
Marcel
(*): I prefer to call it 1/f noise rather than current noise because you can also see the thermal noise as current noise when you choose the Norton equivalent of a noisy resistor. Then again, others may object against the term 1/f noise because its power spectral density is not necessarily exactly inversely proportional to frequency. Anyway, we mean the same.
- Home
- Source & Line
- Digital Line Level
- Valve DAC from Linear Audio volume 13