When I am taking about the sound of the mesa being better in class a/b, I am not talking playback fidelity. Even clean a/b projects. I have an old 30's single ended a amp. Magic too, no balls but lots of harmonic when played clean. Better than knoplers tone, but 15 watts or so.
Last edited:
The previous tests I ran (there have been a few over the last few weeks) were a lot of fun. If you are all interested I might run another but this time its based on passives. And it would be run much more stringently with a view to getting real statistically valid results as to whether or not you really can detect an audible difference.
Is it the common best understanding that on system level, with a fix distance between source and transducer, keeping the speaker connection short is the optimum topology.
If so, what is the basic argument for this?
//
I can appreciate that but it is not an insight that can be practically used at all times. Rather, we sit with fix distance and wonder... long line level or long speaker cables.... I know there are a lot of "depends" but in general, on a system architectural level.. what would be the best approach - and why.
Some people put resistor at amplifier output. Even 10 Ohm does not seem to "ruin" the sound so I cannot accept that a few ohm difference in cables is 100% responsible for the difference in sound.
Is there any articles on how the cable will become a receiver for RF or such thing? I think this is more serious than resistance.
If I had to put my driver close to my amplifier, I will experiment with the effect of nearby AC wire to speaker driver (or crossover coils). From several experiences I found the effect were detrimental to sound quality. But it was very close, so I don't know if 1 meter is acceptable or not, but it is a distance I will tolerate without having to conduct experiment.
That's easily amenable to a basic analysis. In both cases, you have a voltage source and a voltage divider. The effect on frequency response of the load (the bottom leg of the divider) is trivially calculated. The main complication is the capacitance of the typical coax used between preamp and power amp, but that's just as trivially calculated- see, for example, the analysis in my Heretical article (The Heretical Preamp).
I apologize for the appearance of self-promotion, there's nothing original there, just basic circuit theory; it's just that I happen to know where to find it.
"Fixed" isn't really the right word. If you have a cable with a high DCR and your speakers are optimum with a low source impedance, you can't add a negative resistance to "fix" the frequency response and power loss. The advantage, such as it is, to having the DCR high is that it tends to reduce instability in amplifiers (almost always "high end") designed in ignorance or defiance of basic engineering.
Personally, I'd rather use a competent amp, but there's certainly a school of thought that an amp with marginal stability is somehow more "resolving" because it "reveals differences between cables." Likewise, speakers with wild impedance swings.
The LM's are fast and have a UGF of c. 55MHz and a LF OLG of over 130 dB.
You really have to decouple them well close to the PSU pins, make sure any load you drive is isolated through 50 ohm resistors. If you drive a capacitive load, it's quite easy to get them to oscillate at 20 MHz. Feedback resistor junction has to be right at the -input. Use an RF filter in the input if you are feeding if from an external source.
These are very good opamps- but you have to treat them much more carefully than 5532/4 or some of the other VFA types on the market.
My simple X-Altra Mini uses a LM4562 per channel and has a very clean open sound
I have read the PSU pin decoupling caps recommended in the data sheets is often not sufficient. How do you go about determining optimum values? Also, with regards to ocillations and output resistance or input filtering. How do determine what is needed at the design level?
SMT, is the ticket, has been for a while now, same goes for RFI/EMI, ferrite beads, leadless caps (MLC) and solid ground planes, shielding, just like any other RF type of design. AM radio these days is a PITA for removing noise sources.
If you design as if it is RF frequencies, then I think that you can be almost assured that it will work at 20KHz, except for the controlled Z stuff. I am not not going to spec TL test coupons for a audio design, am I
?
Last edited:
Actually that is not what I mean with the word "fix" (I forgot to use quotes there). When there's difference in sound between two cables of differing resistance, the sound will not be similar when resistances are made similar by adding small resistor. It must be LRC affecting phase.
- Status
- Not open for further replies.