The opamps in these soundcards, the sondigo inferno and the x-meridian are only for the input right? Is there analog out identical? That is all I will be using the card for and was wondering.

Thanks!

The Auzentuch X-Meridian does mention op-amps. Those would be on the analogue out side of things. The Sondigo Inferno may well have the same kind of circuitry but I didn't see any mention of that aspect on their web site.

How does this effect the signal? Are they only going to "color" sound or is there some technical necessity given the desired level of quality the the ak4396 achieves? In short what do they do for this circuit?

if the opamps on the analog outside of things don't color the sound then why would they want to be replaceable? Are these required for an accurate signal?

I noticed the rmaa test they had up with the xmeridian and sondigo inferno compared showed bad results for the sondigo, but perhaps this was because of the inferior input dac used for the sodigo, thus effecting the loopback. If they used a separate card for the loopback would the results have been the same?

the rmaa results I am talking about are here
http://techreport.com/reviews/2007q1/oxyge...s/index.x?pg=17

Opamps on the output are part of the analogue processing, post DAC. I haven't investigated any particular circuit but the two reasonable functions for them are amplifier and buffer. The goal is to put out a stable line level signal. "Line level" for home audio equipment, and for soundcards not intended for professional studio use, is rather undefined. It can vary quite a bit from device to device but rarely so much that they can't work together properly as long as there is a good volume control someone in the signal chain.

Easily replaceable opamps are not especially uncommon, although I don't know if there is another soundcard manufactured that way. I have a CD player with the opamps in sockets for easy replacement. Not all opamps are equal by any means, but I suspect that if what you have in the circuit is adequate to the basic needs, putting in something more expensive is pretty much the same as replacing the stock power cord with an $500 power cord: the true believer will be transported, the ABX tester will be left scratching his head. On the other hand, perhaps those in this soundcard are not adequate to its basic needs.

When two soundcards are used together with RMAA, the lesser one will determine the results. The idea behind using two is to compare an unknown soundcard to a know card of extra high quality, one known to be so good it can be a reference against which to compare the other. If neither can be assured as a superior product, there is no point to it.

ok, thats mostly makes sense too me. As usual you are very informative

So do all soundcards have opamps after the dac to bring the level to something they desire? I am under the impression that the sondigo inferno doesn't have any after the dac; can the dac take care of outputting true line level?

There is just so much opinion I guess to try and sell products that the technically superior product gets lost when I research, and I haven't had enough electrical based physics classes for this sort of thing. Since I am looking for something that just does the DAC job and outputs line level at a very accurate level, perhaps there is some science based card somewhere that they sell to universities or something! I mean honestly I don't see anywhere on either the xmeridians site or sondigos that shows the level being put out and its frequency tolerance or anything of the sort. Is there anyway I can figure out which is more accurate????

Op-amps are usually used with negative feedback to stabilise them and maintain a gain far lower than open-loop gain. With simple resistive feedback this ought to produce a very flat frequency response.
The gain-bandwidth product (unity-gain bandwidth) of most op-amps is many MHz at unity gain and surely the voltage amplification applied after a DAC is going to be much less than 10, so I'd be amazed to find the 3dB bandwidth is very close to audible frequencies or that the op-amp's noise figure is significant.

It may be that the DAC is followed by a reconstruction filter (which might affect audio quality far more) and that any op-amp present is principally used as a buffer to present the filter with a fixed impedance load yet drive line-out or headphone-out loads of variable impedance.

I'd have thought that a well-designed reconstruction filter (which might well be integrated into the DAC chip rather than made of discretes) is probably a little more important to audio quality than the choice of op-amp.

The x-meridian seems to use an AK4396 DAC. As will most AK DACs, the output filter, differential to single ended conversion, level shifting and buffering is achieved neatly with a single OP-AMP (or 3). The AK data sheet shows this as fig 16 on page 34. (There are other examples in the following figures.)

It is a very standard circuit topology for modern DACs that run from a single ended supply.
It is probably quite important to have the correct, well-behaved load on the output of the DAC to achieve the performance goals (which are extremely good, I notice).

Most of the filtering is internal, with only a couplle of relatively high frequency poles in analog. If a soundcard designer can fail to make that circuit transparent they should find a new job. The choice of OPAMP will determine such things as load tolerance (5534 being quite good, for example) and could easily influence the RMAA results, especially at high frequency (but -I would guess- not audibly, for most reasonable choices, in most conditions).

Ken


Ken

I am mainly ignorant here, which is certainly not very useful when trying to answer questions, but since I started: I know that the output of many DAC chips is differential to facilitate their use in balanced (mainly professional audio) circuits. Another common usage for opamps with DACs, perhaps really the most common, is to sum that differential output to produce the unbalanced output that most home audio equipment uses. I think this is the most likely use for them in that soundcard.

While I don't know the ends and outs of it all, I think this is an easy enough job that quite a few opamps can handle it without difficulty. So, unless the manufacturer deliberately choose from the lowest end of the line, what comes with the soundcard will perform about as well as any replacement.

There are a number of speciality audio application opamps. For certain kinds of jobs, some of them really are better than others, but my first guess would be that, more often, using a more expensive chip, rather than a more general (audio type) opamp, is mostly about producing measurements that no one can detect without test equipment.

Putting in sockets to let you change the opamps is probably not significantly more expensive for the manufacturer than just soldering the chips in place. It's only real advantage is the advertising it allows. This is the same reason some CD player manufacturers do it: sell to the audience that likes to twiddle with their gear.

It should also be noted that putting the opamps in sockets could be a source of unreliability compared with soldering them into the circuit.

That is always possible but not actually a problem very frequently.

Perhaps not, but since this is a contact between a (probably) tinned component lead and a (possibly) gold plated socket contact, the chances of corrosion are greater than a gold-gold connection, and noise due to corrosion will be noticeable at microvolt levels in an audio circuit whereas it probably won't be noticed in digital circuitry even at tens of millivolts. I just don't see that it is worth the risk.