rowuk
Been here a while!
Pain in the ass, ex-patriot yank living in the land of sauerkraut
Posts: 1,011
|
Post by rowuk on Aug 17, 2013 22:54:54 GMT
I have finished building some speakers and the next step will be multiamping. I have ordered a MiniDSP and have googled more nonsense than I am willing to put up with. Almost all of the threads get reduced to subwoofer control, 10 HZ, and 2000 watt amplifiers implemented by the brain dead. Over 50% demonstrate the inability toread the Funking documentation and when you spoon feed them, they just vomit more trash.
So turning the rant off, has anyone here done something reasonable with the MiniDSP. Any modifications? Powersupply tests? As I have one coming, I will document my experiences first hand. If there are interesting suggestions, I am willing to experiment.
My primary interests: Digital EQ Higher order crossovers (>24 dB) Possible crossover frequencies and their effect on tone Digital delay Freedom from digital artifacts Digital attenuation vs analogue Integration of xover/amps/speakers
Goal is to find out what works not what people with no first hand experience THINK that they would hear.
I have some additional hardware requirements and am not sure where I want to go. My tweeters are 93dB efficient and I really will not need more than 5 watts class A. The midrange driver is 115dB efficient and 2 -5 killer watts would be enough. The woofer is somewhere between 94 and 98dB efficient, so 10-20 watts would be fine. I really like valves, but if I could find some really transparent SS low power amps with TONE, I would be interested. The Pass First Watt amps do not sound "transparent" to me.
Stay tuned, it could be a very interesting ride...........
Robin
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 17, 2013 23:14:13 GMT
Hi Robin Not what you are after, but you may find something interesting in this new section of Computer Audiophile : DSP, Room Correction, and Multi Channel Audio. Personally, I think they should first get the best out of digital audio before further processing mediocre digital audio. As usual, the PSU area is the key.
Kind Regards Alex
|
|
rowuk
Been here a while!
Pain in the ass, ex-patriot yank living in the land of sauerkraut
Posts: 1,011
|
Post by rowuk on Aug 17, 2013 23:23:18 GMT
Hi Alex, I think that there is nothing wrong with working on other areas while waiting for the geniuses to figure out how to screw up digital audio even more.......
With the proper design considerations, the MiniDSP could become my "preamp", DAC and crossover. There are some really interesting interface options here. I think that the multiamp process could offer incredible advantages for dialing in SQ. Even if the MD is only a design tool in the end, I will have real data to move forward with. tHe PS is a major concern. A stiff, low impedance JLH could be what the doctor ordered - or perhaps not....... I will know for sure sometime VERY soon.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 18, 2013 11:42:13 GMT
Hi Robin That is a hot topic in several forums at the moment, with many preferring the use of analogue attenuation. With my old Oppo 981 for example, any more digital attenuation than 10dB caused very obvious degradation, although the use of 32 bits in some newer DACs does give a little more headway. Surprisingly, my new Oppo 103 no longer provides digital attenuation. A major problem is that the vast majority of modern amplifiers have far too much gain, possibly to help cater for those poor misguided souls who think that vinyl rules. With the vast majority these days using digital sources with 2V RMS or higher output, we are throwing away S/N if the amplifier has a typical sensitivity of around 650mV for full rated power. The problem here, is that in an average domestic situation we are more likely to have a typical level more like 1W for normal everyday listening, so we need to chuck in even more attenuation! I found the attached quote from HFC forum quite interesting, especially as I have been trying to work out ways to decrease the gain of my 15W Class A amplifier for this . However reducing the gain just with feedback resistor values can lead to the need for major changes in amplifier compensation unless the gain reduction is only minimal. I believe that some amplifiers, such as those designed by Charles Hansen from Ayre, actually provide variable gain , not simple attenuation. Regards Alex www.hificritic.com/Forum/yaf_postst1539_ESS-SABRE-DACS.aspx
|
|
rowuk
Been here a while!
Pain in the ass, ex-patriot yank living in the land of sauerkraut
Posts: 1,011
|
Post by rowuk on Aug 18, 2013 14:01:34 GMT
Hi Alex,
maybe, just maybe the secret is to modify a headphone amplifier up instead of a "speaker" amplifier down. I will get around to this when the MiniDSP is here. Burning gain through a resistor seems to be really stupid. I am even investigating a technique used in studios where the position of the analogue gain controls is "programmable" - with servo motors. I will avoid at all costs sacrificing bit depth through mismatch of gain. I still want to listen to it and describe what is different.
My current system is set up so that I can use about 3/4 of the volume control for my typical listening levels. Adjusting gain with valve amps is VERY easy.
Regards,
Robin
|
|
|
Post by gommer on Aug 18, 2013 18:50:05 GMT
Interesting thread, I'm sold. The first place I would like to try this, would be in the car, in combination with a car PC for audio.
I wonder how good integration can be for master volume control, profile switching, automatic source switching (e.g. voice announcements from Nav, RDS traffic info). This thing is better than than an Alpine PXA-H701 or PXA-H800 for much less money. Use case is off topic, but interesting nevertheless.
Thanks for sharing.
Cheers, Marc
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 18, 2013 22:27:30 GMT
Hi Robin That's an interesting thought, as the gain of my preamp is 3.12 and the gain of my 15W Class A is 16. If only it was that simple. Perhaps if I had very sensitive horn speakers, the output of my Class A preamp with a Zobel fitted could drive them to a decent level ? Kind regards Alex
|
|
rowuk
Been here a while!
Pain in the ass, ex-patriot yank living in the land of sauerkraut
Posts: 1,011
|
Post by rowuk on Aug 19, 2013 15:43:36 GMT
Hi Robin That's an interesting thought, as the gain of my preamp is 3.12 and the gain of my 15W Class A is 16. If only it was that simple. Perhaps if I had very sensitive horn speakers, the output of my Class A preamp with a Zobel fitted could drive them to a decent level ? Kind regards Alex Hi Alex, I think it is easier to take a Class A Headphone amplifier and parallel double or triple up the output devices to get the couple of watts that you need. I will try my HA on the 115dB/Watt/Meter BMS midrange driver (there are even 16 ohm diaphragms to keep other changes minimal!).
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Aug 19, 2013 21:20:51 GMT
Hi Robin My Class A HA/preamp as built by several members, uses + and -20V rails and a bias of 100mA which can be further increased if required. I haven't done the maths,(just out of bed) but it should already be capable of a couple of watts in Class A ? Kind Regards Alex
|
|
|
Post by eugenius on Sept 28, 2013 7:17:12 GMT
LOL, 93dB/W/m needs 5 watts? It's more like 90 watts to hit 100dB at 3 meters with 3dB amp headroom.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 28, 2013 10:01:48 GMT
|
|
|
Post by eugenius on Dec 28, 2013 22:14:25 GMT
I'm sorry, the bigger amplifier just sounds better because it doesn't clip on peaks. I'm sure I don't have to explain what kind of peak power a simple drum needs. 5W is laughably low, if you have a 5W amplifier with a 93dB/W/m speaker you better have some kind of solid state soft clipping implemented or a tube amp that clips nicer than solid state - and you can't speak of sound fidelity except for vocals and easy material. As a tweeter / compression driver amp it's OK for an efficient driver. I doubt the 115dB/W/m is actually a midrange. Maybe it's a low tweeter - 5W is ok for that obviously. But a 20W woofer amplifier is a bad idea. Anyway, if you want an amplifier with "tone" (bad idea in itself) and you're multi-amping, get a Nad M25. Then add your killer watts amplifier and see if you gain anything. Since we're in a headphone forum, would you drive a headphone with 93dB/mW efficiency with a 5mW headphone amplifier?
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Dec 28, 2013 22:46:46 GMT
Eugenius Unless you are living in a huge house, and want to drive your speakers at ear damaging levels, then the 15W Class A amplifier which was constructed by quite a few members , and when using the suggested dual PSUs , which then permits up to 30W into 4 ohms in Class AB is more than adequate for many domestic situations these days. Many people prefer extremely high quality over "grunt" Many like valve amplifiers with lower drive capabilities than the 15W Class A SS amplifier, for example. I have previously constructed several SS amplifiers with 100W/Ch. into 8ohms, and even a big Mosfet P.A. with 240W/ 8 ohms per channel. but due to having similar space requirements as many other members these days ,I simply don't need anywhere near your recommended power output in a large master bedroom, that has been converted to an AV room. A previous version was tested into a friend's big Ohm Walshe 4 ohm speakers with Plasma Tweeters in a larger listening room with additional speakers for surround, and never went anywhere near clipping. Are you by any chance from the U.S.A. where KW amplifiers give bragging rights ? P.S. I see you come from Bucharest, or nearby. Perhaps you have a much larger listening area than many RG members, or have very tolerant neighbours ? Are you married ? The Class A HA/ preamp project constructed by quite a few members, uses + and -20V supply rails , and has a bias of 100mA (or higher) . I will let you do the maths. It sure isn't only 5mW !!!
|
|
rowuk
Been here a while!
Pain in the ass, ex-patriot yank living in the land of sauerkraut
Posts: 1,011
|
Post by rowuk on Jan 2, 2014 22:06:33 GMT
I'm sorry, the bigger amplifier just sounds better because it doesn't clip on peaks. I'm sure I don't have to explain what kind of peak power a simple drum needs. 5W is laughably low, if you have a 5W amplifier with a 93dB/W/m speaker you better have some kind of solid state soft clipping implemented or a tube amp that clips nicer than solid state - and you can't speak of sound fidelity except for vocals and easy material. As a tweeter / compression driver amp it's OK for an efficient driver. I doubt the 115dB/W/m is actually a midrange. Maybe it's a low tweeter - 5W is ok for that obviously. But a 20W woofer amplifier is a bad idea. Anyway, if you want an amplifier with "tone" (bad idea in itself) and you're multi-amping, get a Nad M25. Then add your killer watts amplifier and see if you gain anything. Since we're in a headphone forum, would you drive a headphone with 93dB/mW efficiency with a 5mW headphone amplifier? Eugenius, you have the facts backwards. Based on inefficient speaker designs, all sorts of comprimises pop up when it comes to amplifier power. That has truly biased "common" knowledge and additionally gives the copy paste generation immediate visibility. There can be NO DISCUSSION of power requirements without considering what the source really looks like. Mini speakers are often praised by the press, but their interest is mercenary. Here are my considerations that are my "facts" used in making my audio decisions: Acoustic power required for a given "loudness" can be calculated. First you have to know what type of acoustic power is present at what frequency for live performances. The spread typically looks something like this for the classical music that I listen to: This means that our speakers must have about 20dB more output capability from 150 - 650Hz than they need at 1k-20KHz. If we consider the instruments that produce that high power, they are all LARGE in size. It would be a joke to assume accurate high output from small drivers. The electrical power that you need to get a particular output can be calculated based on the efficiency of the drivers being used. Here we cannot forget that we have 2 channels in stereo and room gain based on its size. We also have to consider that it is impossible to shoehorn a symphony orchestra into a 20 square meter living room at original volume. - the room reflections will destroy any enjoyment at those levels! All of a sudden, a real (large) 95-98db/Watt woofer/midrange (not the full range back loaded horn type!) can get by with little comprimise when you have 10 honest watts per channel where real music needs it most (95db/Watt means 105db at 10 Watts per speaker, this is 108dB for 2 speakers and with room gain, a bit more at the frequencies where the room helps). ULF like in home theater can also be calculated in the same way - of course someone that thinks that they need 10Hz usually just fools themselves into believing whatever that they want to - ported 10" subwoofers, 1000 Watts are typical "exaggerations" that only show lack of knowledge of the basic math. The typical 10" subwoofer driver is luck if it even has 70dB/watt at 10Hz. That means 80db 10 watts, 90 db 100 watts, 100dB 1000watts (if it can even survive this). The port noise alone from such "solutions" disqualify the implementation. The second problem is the "common knowledge" about amplifier power and clipping. The lies found all over the internet have only reduced the argument to the level of a casual reader with no real knowledge. The quality of sound has NOTHING to do with "power", perhaps except for the fact that many high power amplifiers are truly "sound challenged". If we have inefficient speakers (mostly the woofer being inefficient), and burn off power to the midrange and tweeter through resistors/LPADS, of course we are asking for trouble - but only when we are drunk or high. When we seriously listen to music, problems become apparent LONG before any damage is possible. Even if there is occasional clipping, so what? For the milliseconds of clipping, there is no immediate danger for anything, generally there is not, except for exceptionally poor amplifiers, even a noticable change in the sound. Basic math helps us to get an amplifier that will give us 100dB peaks in the living room (in my opinion - more than enough). I don't need 3dB headroom or anything else as this is truly meaningless if we are talking about dynamic music and not AM radio. As far as an amplifier with "tone", I guess you don't understand what I mean. Maybe you have not heard the breath of life that a fine amplifier can breath into fine speakers. Instruments are reproduced in "space" with substance (tone). The playback gains depth, width and presence - even at <80db output. The NAD is not qualified in performance for my "must have" list. In any case, I have had my listening experiments and have decided to use the MiniDSP for what it does best - as a set up tool, then turned off and completely disconnected as it did not pass any transparency test regardless if it was in bypass mode or only delay, only Xover or as an equalizer. My crossovers are now completely passive and part of the amplifier design. What was interesting - even with headphones, adding even small amounts of DSP signal correction helped the frequency domain become more linear, but sucked life out of the music. This sometimes made challenged music (128 bit MP3s) and AM/SW radio even sound "better". The better digital and formats all got worse........ So, I guess 5 watts above 650Hz is only laughable if someones playback defies the laws of physics. My tweeters (>10KHz) have a 2 watt amplifier now and that is many times times bigger than I really would need (20db below 105dB = 85dB. My 93dB/watt tweeter needs 8dB less than 1 watt for this level - that is about 0.14 watts. So I have over 10dB headroom in my situation!).
|
|
|
Post by outdoorguy on Jan 5, 2014 21:57:24 GMT
Kalman Robinson of Sterophile, January 2014, p.56-59, did a favorable short take of the miniDSP10x10hd...of course that's in the same issue as the $100k+ dcs Vivaldi digital playback system.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jan 5, 2014 22:16:58 GMT
Kalman Robinson of Sterophile, January 2014, p.56-59, did a favorable short take of the miniDSP10x10hd...of course that's in the same issue as the $100k+ dcs Vivaldi digital playback system. My own investigations into further digital processing, even copying/moving to another storage location, of high quality digital source material tells me that further digital processing degrades subtle low level detail such as low level ambience, even the very low level harmonics of a female singers voice, so they start to lose a little of what makes their voice unique compared with many other female artists. It's a little like going from Stax headphones to lesser headphones.Even copying /moving .wav files to another storage location causes a minor degradation, despite the check sums reporting the binary content is still the same. This has been confirmed by Martin Colloms and many RG and C.A. members. The quality of the PSU plays a large part in all of this. Alex
|
|
rowuk
Been here a while!
Pain in the ass, ex-patriot yank living in the land of sauerkraut
Posts: 1,011
|
Post by rowuk on Mar 29, 2014 18:16:59 GMT
I have spent more time on this and have given up. Digital can delay cleanly but not filter. Period.
Even a digital 6dB crossover really trashes the sound compared to a line level capacitor. What sounded best was the crossovers over 24dB/8va. Best was still not really "good". My MiniDSP is now a tool for experimenting. It is really easy to model a crossover and then build it analog and get the rest of the sound back.
I will say that there are many audiophiles that can't hear the difference and I definitely don't get into discussions. It is like talking about women, religion or good wine. Without a common description for the important stuff, we always end up wondering why we even talk about this stuff.
|
|