I played around with the UcD series modules a few years ago while working for a local manufacturer of audio gear that used them, as the owner of the company thought they were great. They sounded nasty to me. I brought my Leach amp in and after hearing that he moved away from class D. One of the other engineers there still thought the Hypex were great and couldn't hear any difference. He was not into hifi and didn't have a system at home (and needless to say thought the whole process of evaluation by listening was a waste of time). I'm not saying this proves anything, just recounting my experience.
All this discussion about the merits or not as the case may be for GNFB reminds of the discussion about tubes vs SS.
The industry is so distorted and psychophantic that one very well known and current UK manufacturer markets products using nuvistor NOS tubes. I have heard it from more that one source that the tubes are not even in the circuit. Only the heaters are wired up.
The industry is so distorted and psychophantic that one very well known and current UK manufacturer markets products using nuvistor NOS tubes. I have heard it from more that one source that the tubes are not even in the circuit. Only the heaters are wired up.
I like "psychophantic".
I have some nuvistors. I think they are 6DS4s. I tried to see if they were any better than typical triodes with high impedance plate loads, but no.
The sockets are fairly scarce.
The fact is, the whole thing is a sham. The only reason is so they can get the hi fi press going 'ooh' and 'aah what a fantastic tube sound' when in fact it's a plain good old bipolar transistor sound.
Last edited:
Why do they promote something that is not true!??!!! Makes me get In bad mood.
Sent from my iPhone using Tapatalk
Reminds of the much vaunted Audio Reseach. Zane Johnson (sp?) emphatically dissed semis and claimed tubes were absolutely superior . . . And then used a cheap Dallas Semiconductor volume control on the front end of his tube pre.
The circa 1963 Lumatron sampling scopes used the grid capacitance as a sampling capacitor. Probably the most brilliant piece of hybrid engineering ever, Nuvistors, gold doped transitors, snap diodes, and GE TD 11 tunnel diodes all in the same design. The sampling head was a bridge of schottky diodes soldered directly to the center line of a coaxial hard line through a small hole with the other side soldered to the Nuvistor, all connections GR hermaphroditic hard line.
25psec resolution in 1963.
EDIT - They were 6CW4's
Last edited:
My father designed an FM multiplex subcarrier receiver, all tube, everything conservatively rated and happy at high ambients, that found its way into many attics of establishments that had Muzak coming over the ceiling speakers. He was skeptical of transistors, and even more of integrated circuits ("those are too small to be practical").
When nuvistors came along he embraced them with enthusiasm, believing that they would prolong the culture of tube electronics, and modified the receiver to put one in the front end.
As time went on he started a company making demonstration switching equipment for audio showrooms. In time he came up with relay boards for distributed systems more appropriate for car stereo displays. They used discrete transistors, but he realized the boards didn't look modern enough and reluctantly migrated to an integrated circuit---but just one that replaced some of the discretes, the venerable CA3086. This was in the early 70's.
When nuvistors came along he embraced them with enthusiasm, believing that they would prolong the culture of tube electronics, and modified the receiver to put one in the front end.
As time went on he started a company making demonstration switching equipment for audio showrooms. In time he came up with relay boards for distributed systems more appropriate for car stereo displays. They used discrete transistors, but he realized the boards didn't look modern enough and reluctantly migrated to an integrated circuit---but just one that replaced some of the discretes, the venerable CA3086. This was in the early 70's.
I understand that some manufacturers kept a small stock of nuvistors for replacement parts for things like oil drilling head instrumentation. They got upset when audio people would try to buy them for their use and refused to sell to them.
Microphonic they may well be, but nothing else withstood the high temperatures.
As for microphonics, I noticed some of the supertriodes with very small grid-cathode seperation have a rather high mechanical Q. Once excited they decay with ringing somewhere in the 10kHz range, modulating the gain, and take a long long time to get down to less than the noise level of the circuit.
When I proposed using some for a tube line-level crossover, for someone who swears by thermionic devices, one of the Audio Note guys that he respected said that he found them to sound "metallic". Another of the AN folk said that one didn't want to use the same tube twice in a signal chain because they would introduce too much of a particular sound, which the friend translated into "sonic signature buildup". I told him it sounded like a laundry detergent ad.
Microphonic they may well be, but nothing else withstood the high temperatures.
As for microphonics, I noticed some of the supertriodes with very small grid-cathode seperation have a rather high mechanical Q. Once excited they decay with ringing somewhere in the 10kHz range, modulating the gain, and take a long long time to get down to less than the noise level of the circuit.
When I proposed using some for a tube line-level crossover, for someone who swears by thermionic devices, one of the Audio Note guys that he respected said that he found them to sound "metallic". Another of the AN folk said that one didn't want to use the same tube twice in a signal chain because they would introduce too much of a particular sound, which the friend translated into "sonic signature buildup". I told him it sounded like a laundry detergent ad.
Last edited:
Member
Joined 2009
Paid Member
I heard that some Class D designs allow for a 'Class AB' type cross-over to eliminate the issue with dead-time distortion???? I imagine that this approach may be less efficient.
With reactive/inductive loads there is often a phase difference between voltage and current so that the zero cross-over in the voltage domain may not align with zero cross-over in the current domain - I can imagine this makes the problem more subtle ????
So, what kind of crossover distortion type things happen in class d? Things that make distortion go high as signal level decreases toward zero? If there are lots of such, as stated.
It must be understood that their can be no overlap between the positive and negative switching. IF you had overlap, like we normally put into class AB or even class B amplifiers, you would get a current SHORT that would blow the unit, or at least reduce its efficiency severely. Matching the turn on times and turnoff times of two separate fets is the challenge. This creates a 'dead zone' just like analog crossover distortion does. It is quite a design challenge, perhaps the best designs have fixed this, but I doubt that the majority of class D designs have.
And that causes a discontinuity at low signal - how? Both sides still switch. Duty cycles of both polarities still change continuously. The non-conducting time will certainly affect linearity, but how is it worse at small signal?