|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
69.40.209.98
The Boston A70's I reported on a few days ago had a short life with me. Plucked bass in the lower ranges made one woofer buzz, and besides that, they did not compare at all well to the Advents. They got swapped for a Rotel RCD-955AX CD player.I really like my RCD-971 and wanted a second good CD player, so I tried the 955AX. Once I got it home, I hooked the 955 up to another input on the NAD and dug out a pair of matched CD's. (this is the interesting part)
The two sounded very similar, but not identical, the 955 being a little lower in level in some ranges and having slightly more sibilent highs. It then dawned on me that the interconnects I was using for the comparison were not the same. I had a 12" pair on the 971 and a 1 meter pair of a higher capacitance design on the 955. I dug out a set of 1/2 meter DiskWasher Gold-ens low capacitance cables and tried again. Now the two players sounded 100% identical.
I had the same experience 5 years ago with a different model of NAD and a Sony CDP-XA1ES series player. Going from the same kind of 1 meter, 100 pF IC to a 12", 28 pF low capacitance DIY made a dramatic difference. Same effect, too; the highs got a lot smoother and cleaner. On the Sony the highs were unlistenably shrill with the 1 meter cable. Nothing subtle about it. Frequency response measurements showed nothing.
Part of the issue (I think) is the rather high input capacitance on the NAD preamp, around 400 pF. Still, if you work out the reactance of 500 pF at 10 kHz, it's around 30,000 ohms. So there should not be any loading effects. Besides, if the capacitance acted to load down the CD player's output, the highs ought to get softer, not brighter. So something else is at work here.
Oh yes, the 1 meter IC's are Tara Labs Prism 3 at 100 pF. The 12" IC was made from a a cut down Radio Shack Gold extension cable (male on one end female on the other) which had 54 pF capacitance at a 3 foot length. I added Radio Shack gold plated plugs and made 2 of the 12" IC's from the longer one. The DiskWasher 1/2 meter IC's measure 34 pF.
If anyone has any thoughts on this, I'd certainly welcome them.
Follow Ups:
Interesting post, as I have a Rotel 945Ax feeding an NAD 1000 preamp via 1/2 meter Monster M550's - no complaint with the sound quality thru my large Advents, but maybe I'll try swapping cables around to see what happens.Coincidentaly, I recently replaced both woofers in my garage systems BA 70 IIs with new ones direct from Boston Acoustics ($106 the pair, shipped) - one of the originals (1987) had the same low freqency buzzing. Best money I've spent recently on hifi - astoundingly good sound for the $181 I've put into them, including purchase.
The only explanation I can think of involves the extremely high gain of modern opamps(or differential discrete gain stages) that are likely used in the NAD to buffer the inputs. If the interconnect impedance somehow interacts with the gain controlling feedback impedance of the input stage(parasitic resonances etc.) you might see effects like yours. I remember designing a push/pull on/off VFET driver circuit once without understanding the concept of Miller effect that "amplifies" input capacitance. In my design the effect didn't let one VFET turn off before the other turned on creating a momentary direct short between B+ and ground everytime the state changed. Needless to say that was not a good thing.:)
Gary
Gary,The effect seems to follow the CD player. Some CD players don't do it with any amp. Is there a similar effect on the output stage due to capacitive loading?
I spoke to one of my engineering buddies, the "analog circuit wizard". I told him about my experience, and he thinks it sounds like oscillations. The added capacitance of the cables could be creating something like a Colpitz oscillator out of the CD player's output circuitry.I do know that John Dunlavy was a proponent of very low capacitance IC's and so was "Gizmo", but he never expressed it in that way.
Jerry, I think your “analog circuit wizard” is on the right track.It’s been a long time since I exercised brain cells in this area, but there are a few things that can lead to waveform distortion or parasitic oscillations. Parasitic oscillations are usually at high frequencies and components / sub-components can cause a high frequency resonant circuit when the device has high gain. Because oscillations can be dampened by spoiling the Q of the resonant circuit, small changes to circuit lead lengths and position can stop a resonance. Occasionally, a few pf can move the oscillation especially if the resonant circuit includes a decoupling capacitor (which may have an internal series inductance). I’ve seen parasitics on a scope / analyzer show themselves as unexplained distortion or jumps in dc offset.
Some old grad school notes also lead me to believe that if an output load has an excessive capacitance, it can also cause a mismatch at high frequencies resulting in standing waves on the output cable.
If there is a weakness with input IC shielding / ground performance / routing, it also increases the probability of picking up RF / EMI which is amplified as non musical distortions and can create a small potential. If only a small portion flows back via the input, then a frequency boost or oscillation may occur.
It looks like the interactions you saw are more sensitive than I would have previously guessed but it makes sense looking back.
The JPS Labs cables I favor have both low capacitance and inductance resulting in a very low dielectric constant of 1.16.
I'm no analog wizard as evidenced by my anecdote so I would go with your buddies theory. My only point was that small impedances can have a much larger effect than anticipated especially if some sort of feedback/resonance (e.g.oscillation)occurs which is consistent with your friends explanation. Sounds like the extra capacitance of the NAD interacts with some CD players output stage and the extra capacitance of the longer cable is the proverbial straw.
Gary
That's kind of what we pieced together as the likely scenario. But without some experimental data, it's just speculation.It does bring up an old "thing" of mine. That is, the "interface" errors are often larger than the internal errors in components. For example, the just mentioned aberration caused by extra cable capacitance at the interface between the Cd player and the preamp. Further, it doesn't show up in the bench tests on either component.
Amp and cable to speaker is one I've hammered on for years here and other places. Speaker to room, of course, generates big errors. And phono cartidge to phono preamp can generate frequency response aberrations larger than the specs for either the cartridge or phono preamp.
There are lots of others, some fairly obscure. Like having a tape deck connected to the tape out jacks, and the tape deck not powered up. Turns out some tape decks had a very low input impedance in that condition and it loads down the selected input.
This is all caused by the industry having a component orientation, instead of a systems orientation. So there are no good systems interfacing specifications, just industry standards for measuring the components on the bench.
When are you coming to Oberlin? I didn't save the e mail.
Hi Jerry,Do you have any recommendations for a good capacitance meter. I think it would be interesting to test some of the cables that I have.
Parts Express has a small stand alone Meter that has sufficinet resolution for the small capacitances in cables. Don't get the LRC meter, as the capacitance ranges don't go low enough.
Hi Jerry,Thank you for the info. I had seen and wondered about that specific metter. Parts Express is not always the cheapest game in town but they are very reliable.
Best wishes,
This post is made possible by the generous support of people like you and our sponsors: