Guide to install pre-outs on an AVR - Page 3

Have you calculated out or measured what the load-dependency is with a passive versus op-amp solution? I might suggest that it's something completely not worth being worried about.

Reviews do not comprise technical evidence, none of which so far, beyond the trivial impedance difference, you have actually supplied.
No, I'm sorry, you are simply wrong. A simple survey of the input sensitivities of amplifiers will show that they are almost always in the vicinity of 1-2vrms for rated power (see below). As a typical DAC puts out around 2Vrms for practical and historical reasons (it was the Redbook CD standard) and a convenient value on which to base sensitivity as it is close to the pro standard of +4dBu (1.23Vrms). Very old power amps sometimes have a sensitivty much higher than this reflects the old domestic standard of -10dBV (~333mV)

Part of the reason for the delay in replying was that I had a service manual for my own AVR on order and got it yesterday. Below is a snapshot of the black diagram for the audio path.



Note the 29dB in the magnifying glass, as this is the gain for the power amps. 29dB is 28.2x and the amp section is rated for 160W/6Ω which is 31Vrms at rated power/load so the input sensitivity of the amps is 1.1Vrms, right where I said it would be.

Here is the block diagram for the entire audio section.



When supplied a 0dBFS signal the BB PCM1690 DACs they put out around 5Vpp (differential biased at 2.5V). In this AVR it is then filtered and attenuated by a factor of 1.7 to 2.9Vpp or 1.04Vrms. This then goes into the VLSI chip in the dotted section where it gets buffered and passed to the power amp sections and parallel to the pre outs.

Interestingly, the analogue inputs are digitised and processed in the digital domain via the CS42516 and it appears the volume controls are implemented in the DSP as I can find no sign of them elsewhere.

Whilst one Onkyo AVR hardly constitutes universality, it is intended to be used in a vast variety of systems including lots of other hardware, so it's hardly in Onkyo's (or any other manufacturer's) interest to have it vastly different to most of the other gear out there. I found a couple of oddballs like a Krell at 4Vrms sensitivity, but for a domestic amplifier of nearly a kW, a lower sensitivity is probably a bit safer, considering the general lack of technical competence of audiophiles.

However, I took 5 mins and looked at the Rotel, Parasound, Emotiva and QSC sites and found they were all generally similar.





Note the Parasound mentioning it's sensitivty is compliant with THX specs. (1.58Vrms for full power)



300W/8Ω is 49Vrms, 32dB is 40x, sensitivity is 1.23Vrms (+4dBu, pro standard)



I cropped this to only include the class AB amps to keep the discussion simple.
Sorry, but in this instance, you simply are. I've designed a considerable number of amplifiers over many years, and read lots and lots on this subject (Cordell's recent book the latest) and am quite well qualified in linear electronics. I also spent a lot of years designing or modifying a myriad of gear to interface with other gear at a domestic, studio, broadcast and PA levels, often to ensure that the technically ignorant artistic types using the controls couldn't damage it or easily produce a poor result, eg distorted sound going to air by accident, so I have a good background in knowing what's actually in a vast range of audio gear out there. I also have an offer to go back to teach electronics at College next year, but haven't decided whether I will yet.
Edited by A9X-308 - 8/1/12 at 11:58pm

Yeah, well, I build amplifiers as well, and have researched quite a bit on forums like DIYAudio, HTGuide, this one, DIYMobileAudio, etc. I also repair (generally) car audio amps. I didn't say that gain is never in the power amp section. Clearly the gear you have worked with is not the same as what I have worked with. Can you claim that every amplifier has gain in the amp stage? Because it is, at the very least, reasonably common to be in the preamp stage. Perhaps I did overstate my argument, but you're now doing the exact polar opposite, so... Also, I can read your provided schematics just fine, thanks. You can pass your judgement while bragging about your own qualifications elsewhere because it's not required here.

For the op amp argument, to each his own. I'm not going to provide you or anyone else a 10-page research document showing that you should use an op amp and not just resistors. I've personally tried the resistor-only method before and have seen it create issues, even when the resistors were of a much higher impedance than the speakers they were in parallel with (about 100x). You say I don't have a good enough reason to explain why they're worth it, but your argument isn't too great either, and you seem to feel too strongly about it, whereas I haven't been pushy with my idea whatsoever. I really don't give a crap if you want to use op amps or not, but they ARE a better solution - period. Your only valid argument is that they may add unnecessary complexity, because they're damn sure not a bad idea.

But hey, if you need to make yourself feel smarter than me or others on this board, then you do what you gotta do man. Personally that's not why I make my posts though.
Edited by DonoMan - 8/7/12 at 7:17am

Sorry, it's not common for the majority of gain to be in the poweramp stage, and you've provided nothing so far to convince me that my long experience is wrong. Not to mention I gave examples from amplifiers in a range of applications, as well as historical reasons why. Schematics and amp specs are easy to obtain, so where are yours to back this up? Where are the low gain power amps, ie those that don't typically have a sensitivity near 1V? Opamps do create complexity and expense in this instance, but you have not yet given another reason for using them except the trivial impedance issue and your assertion that they ARE better (capitals do not make your argument stronger). What precisely, in technical terms, do they add and what are the 'issues' created without them being present in circuit?. It doesn't need the strawman argument of a "10 page research document", a paragraph should do. But as you have used more than that to actually not answer how/why, it's clear you do not have a good reason why, except it's your preference.

i haven't been following this thread, but linkwitz has a method for attenuating the speaker level outputs in order to turn them into pre-outs for his active system.

might be worth having a look at.

Don't you need a resistor divider to set the gain of the op amp stage?

Of course, but it is still then independent from the load. Personally I do believe in adhering to specs as closely as possible, and in buffers around cables. Yes, he'll probably say the power amp is its own buffer, and that's not false, but I don't generally like to put audio stuff in parallel unless it's the same - I HAVE heard distortion using just a voltage divider on a power amp output when in parallel with a speaker on that output. I have mentioned several times that they are not actually necessary, but are an improvement, especially if driving speakers in parallel with the divider. Not sure what that other dude feels the need to troll that advice so badly for as if I'm saying that they're absolutely necessary. My argument in this thread has consistently been "while the voltage dividers will do just fine in most cases, if you want to do an even better job that is totally independent of the amp you use, I recommend op amps." Again, if someone were to want to use, say, an older amp with BJT inputs, which require a lot more current than FETs, then the passive circuit would be a relatively bad idea compared to the active.

@Know-it-all dude:

P.S. I will admit this - I have spent more time with amp designs from DIYAudio (and some of my own derivatives) and somewhat with car amps than I have repairing home audio amps (I have repaired those in the past as well, but only a handful). I have made discrete amps too but not for a good 5-6 years (nothing too amazing on that front anyway). Make sense why we'd have different experiences now? Now stop being an ass. It's unnecessary.

Stuff using chip amps (typically the less expensive stuff, but not necessarily only the junk - chip amps can be good too) will generally have their gain structure like I described. And I bet those chip amps still make up the majority of amps (I mean amps in all gear, not specifically standalone amps) out there. And that's what I meant. Personally I prefer Hypex UcD Class D amps anyway.
Edited by DonoMan - 8/8/12 at 11:55am