JohnSwenson
2006-07-20, 21:31
I've started a series of measurements to see if I can measure any differences between a stock SB3 with the stock switcher and my DIY linear supply. I have a some preliminary results.
First off a little about the supply, its has a hefty triad transformer, high current Schottky diodes and a CLC filter that has been carefully tuned so it has a fast response but does not ring when presented with a current change. I have found that such supplies sound the best. The filter is followed by a LT1084 regulator.
The measurements were performed with a spectrum analyzer looking at the analog outs while playing a 440Hz tone at -3dbFS with the SB volume at full.
There were three differences I could see
1: the 3rd harmonic was 2.5 db lower with the linear supply.
2: the noise floor was slightly lower with the linear supply.
3: occasionally a "spike" would show up with the stock supply.
#3 takes some explanation. The spectrum analyzer works by slowly sweping through the frequency spectrum with a certain "filter" bandwidth, this shows up a vertical trace with a "skirt" that sort of looks like a cone. These spikes had no skirt, they are straight lines. This can only come from a transient phenomenum. The pikes did not always occur, and when they did they were at different frquencies. This leads me to belive they are really noise bursts rather than a continuous frequency.
Now comes the fun part, what causes these effects? #2 and #3 seem like they could be line noise making its way through the PS, the linear supply does a better job of filtering out junk on the line. This doesn't explain #1. The only thing that makes sense to me is that the switcher has more "signal modulation", its voltage varies with the audio signal being outputted, this could cause some sight increase in the distortion. To test this I'm going to so some measurements of the supply voltage itself and see if there is any difference in regulation.
Next comes measurements of jitter with the different supplies, I'l post that in a day or two.
John S.
First off a little about the supply, its has a hefty triad transformer, high current Schottky diodes and a CLC filter that has been carefully tuned so it has a fast response but does not ring when presented with a current change. I have found that such supplies sound the best. The filter is followed by a LT1084 regulator.
The measurements were performed with a spectrum analyzer looking at the analog outs while playing a 440Hz tone at -3dbFS with the SB volume at full.
There were three differences I could see
1: the 3rd harmonic was 2.5 db lower with the linear supply.
2: the noise floor was slightly lower with the linear supply.
3: occasionally a "spike" would show up with the stock supply.
#3 takes some explanation. The spectrum analyzer works by slowly sweping through the frequency spectrum with a certain "filter" bandwidth, this shows up a vertical trace with a "skirt" that sort of looks like a cone. These spikes had no skirt, they are straight lines. This can only come from a transient phenomenum. The pikes did not always occur, and when they did they were at different frquencies. This leads me to belive they are really noise bursts rather than a continuous frequency.
Now comes the fun part, what causes these effects? #2 and #3 seem like they could be line noise making its way through the PS, the linear supply does a better job of filtering out junk on the line. This doesn't explain #1. The only thing that makes sense to me is that the switcher has more "signal modulation", its voltage varies with the audio signal being outputted, this could cause some sight increase in the distortion. To test this I'm going to so some measurements of the supply voltage itself and see if there is any difference in regulation.
Next comes measurements of jitter with the different supplies, I'l post that in a day or two.
John S.