| terrorizer |
Well, I come from the stone age of the amps that were built with discrete transistors, and I haven't kept up with the latest technologies in power audio amp design for a long time.
And it's time to wake up and smell the reality of integrated amps :)
It's so hard to choose from the aboundance of integrated amps that are available. I'm looking for a chip around 30 watts, with a good price/quality ratio. It will be used on speakers with Perrless 832556/811820 drivers, if this helps.
I really don't know where to begin looking. From what I know, National Semiconductor's range of audio amps is pretty good, but I need some advice. Would someone help? |
|
|
| till |
hey, they are only chips- LM3875, LM3886, OPA549, OPA541...
do you expect much difference? |
|
|
| Morse |
Hi Till;
>>>...hey, they are only chips- LM3875, LM3886, OPA549, OPA541...do you expect much difference?...<<<
Um, yes actually, but that's because I'm an antediluvian from a valves background. I guarantee there are tonal differences between the 2A3 and 6BM8 for example, along with subtle differences in presentation (the 2A3 tends to have a very 'spacious' sound, as though you're in a larger space, and somehow the 6BM8 seems more 'crowded' - they both sound great, but in different ways). Even within a valve type, different manufacturer's versions will have subtle differences (which leads to the practice of "tube rolling", as it's called in the US) that causes all manner of pointless name-calling on some fora (only valves I genuinely detest are ones that fail early and ugly).
Maybe I'm just hoping though, and all chips really DO sound alike - there wasn't any input when I posted a specific thread asking about tonal differences in chip families, so maybe it's a really dumb question. When time permits, I'll have an LM1875 up and running and compare it to my TDA2030A. Who knows? Maybe they really do all sound the same....
All the best,
Morse |
|
|
| terrorizer |
| What do you mean? I don't think that those are the only chips available. I'm not asking only about the NS chips, if this is what you understood. |
|
|
| till |
please donīt take my post to serios.
more serios: yes, they may have tonal differences, but different than differences from different tube devices or topologies. I suspect the chips has a average lower spectrum of harmonics than the common valve (ot ZEN - like) amp, but amount of harmonics over output voltage is different. The ZEN f.e. has very low THD at low levels, and THD goes up with power. The chip has relative much more THD at low levels, going down to to a min. at maybe 50 or so% of maximal power output, and climbing again. The distortion at low power levels is heavy crossover distortion. SET or Zens has no crossovver distortion. So with very efficient speakers (you will need to reproduce dynamic range) you always listen a chip amp at power levels with a lot of crossover distortion. With a SET or ZEN at 0,01 to 0,1 W average output power you alway listen with much less distortion than with the chip....
I use a chipamp for my computer soundcard, as whats coming out there is allready ****, and a class a amp burns a little too much expensive energie for this purpose.
Over all i would say, quality of the amplifier depends mostly on the speaker it is connected to. |
|
|
| jean-paul |
| quote: | | So with very efficient speakers (you will need to reproduce dynamic range) you always listen a chip amp at power levels with a lot of crossover distortion. |
Quite a bold statement Till. Are you sure about this, it seems to me you are exxagerating ? The chipamps are biased too to avoid crossover. They are not plain class B amplifiers and certainly the National types are well designed ( hence the worldwide trend of building Gainclones etc. ).
I recall Peter Daniel comparing Gainclones with very serious equipment including Pass DIY amps in favor of the chipamps. |
|
|
| till |
hey, iīm not sure about any thing, and iīm only a mechanics student. Not EE.
But i feed sine waves of different frequencys into my amplifiers, and put a ECM8000 in front of the speaker, and measured a) the amplifiers output b) the signal from mic in each one channel of 7904 scope. Also i measured THD with spectralab, but resolution of this measurement is limited.
This told me the Hypothesis above.
| quote: | | The chipamps are biased too to avoid crossover. They are not plain class B amplifiers. |
Maybe the National chips are much better, but i canīt say something about. I did my experiments with different BBs. OPA548 and 549. The effect isnīt measureable at a few 100Hz, but at a few 1000. Speaker was TAD2001 Horn at about 110dB/1W. |
|
|
| terrorizer |
| quote: | Originally posted by till
So with very efficient speakers you always listen a chip amp at power levels with a lot of crossover distortion. |
I'll use drivers with small sensitivity then :D
Like you said, the selection of an amp is based mostly on the used speaker, and that's why I mentioned my future setup: bass-reflex with Peerless 832556/811830 speakers, 18dB/octave filters. There should be a formula to compute acceptable THD based upon speaker parameters. Just kiddin. |
|
|
| terrorizer |
| Let me put it in another way. What chips do the consecrated amplifier manufacturers use? For instance, for amplifiers up to... say $1000. A few examples would be appreciated. You may strip "consecrated" if it bothers you. |
|
|
| till |
the edit button is under your posting - buit it dissapears short time after you posted.
T+A http://www.taelektroakustik.de/ uses LM3886 or so, And the gaincard, the origin if all gainclone hype, i think LM3875
from
http://home.ca.inter.net/~cfraser/Gainbuilder.htm
"First and foremost, the ACTUAL CHIP used in the Gaincard is relatively irrelevant. Most monolithic solid State Power Op-Amp's share the same basic amplification circuit, have very similar THD/N, IMD and such measurements and so on." |
|
|
| jean-paul |
That is more a personal attack site against Kuei aka Thorsten Loesch. Some of the pics and hints that made a fool outof Kuei have been removed in the meantime. I do share *some* of his opinions though, I also never understood the hype around Inverted Gainclones. But that does not take away that hatred and jealousy aren't the best friends of ratio.
Till, it seems you have some prejudice concerning Gainclones etc. I can only advise you to listen to a wellbuilt LM3875 ( this is the one to use ) amp one day. I know one guy that sold his expensive tube amp after sceptically building one. I joined the groupbuy of BrianGT and I think a better amp can't be bought for the money ( 60 euro for a stereo kit + supply board ). Some amps require a heatsink that costs nearly the same. |
|
|
| till |
Its not i have something against chipamps, but the hype makes me think. I donīt belive there is such thing like an optimal amplifier. And a chip isnīt one also. They have a lot advantages, like they are cheap and easy to build, and you canīt make much mistakes building the power supply and the little bit you have to build yourshelf until its usefull as audio amp. I doubt they are better amplifiers than discret amps. The chip has lots of compromises, as a discret design has. But with a discret design you can descide what compromises to make. Iīm sure most important for the hype is gainclones are easy to build - not that they are better than all discret amps. This does not necessarily mean they are worse in general. The amount of designs out there speaks for "there may be better and worse discret amps out there than chips" As i said, personal choice may depend on the speakers. I donīt need 50W with speakers of low 100dB/1W and high 110dB/1W. A Zen lite at 1W max. is just fine, and a ZEN is overpowered.
Another idea is: choice may depend on music you play. Iīm sure electronic music is allread generated in a way it does not cause to much "bad" distortion orders and IMD for the ear, as all these products are allready generated by the instrument. So the instrumentalist avoid what sounds bad. (Also "feedback) Other way a recorded choir: to degenartion of sounds made by human voice the ear is highly sensitive. But the voice allready contains loads of lower order HDs. The device that produces other than these HDs (IMD) or higher order HDs from the complex waveforms generated by the choir disturbs the sound you listen much more.
Only some thoughts....
wg.:| quote: | | Some amps require a heatsink that costs nearly the same. |
My Zen V4 heatsinks alone was more than the double...
You are absolutely right, what you get for the money is hard to beat.
wg.| quote: | | I can only advise you to listen to a wellbuilt LM3875 ( this is the one to use ) amp one day. |
I only had some Burr Brown Op chipamps until now, because National doesnīt want me to test the LM3875.... |
|
|
| jean-paul |
| quote: | | Its not that I have something against chipamps, but the hype makes me think. |
Don't like hypes either and the thought of a cheap chipamp beating class A amps did not make me really happy too. They are easy to build and give great pleasure to their proud owners. I can live with that, hype or not. I just wish people build their amps a bit more careful and esthetical. After seeing the 382nd pic of a bundle of wires around a Pentium II heatsink with some chip on it drowning in compound I have had enough.
But facts are facts, they do sound wonderful. |
|
|
| Ropie |
It never ceases to amaze me how quickly the same old arguments rear their heads again and again, even in postings that began on entirely different footing :D
I don't think you will benefit any more from this thread Mr Terrorizer and i would sugest you do a 'search' on this and other forums and webpages, because that is a much more efficient way to get the information you require ;) |
|
|
| terrorizer |
Can't argue with that. The only conclusion that I can draw from this thread is something like: "Those NS chips suck. But... Wait a minute... As a second thoght... I think they rule"
Don't mean to offend anyone, this thread my lead to something constructive after all. |
|
|
| till |
| quote: | | The only conclusion that I can draw from this thread is something like: "Those NS chips suck. But... Wait a minute... As a second thoght... I think they rule" |
there would be much more to see in case you accept between black and white there lay a lot of different colours. |
|
|
| terrorizer |
No need to take everything literally.
What about the Toshiba chips? |
|
|
| Morse |
Hi Till;
>>>...yes, they may have tonal differences, but different than differences from different tube devices or topologies...<<<
Thinking about it in the broad light of day reminds me of some experiments with "dampers" (rubber rings around the glass bottle of the valve). Some valves that were rather sibillant were tamed nicely, while others that didn't have that problem lost some of the "lifelike quality" that I cherish (not nearly as bad as trying to drive a highly reactive load with a 2A3 though - that's a really ugly exercise!). In other words, microphonics may well play a part of the 'valve magic' (This isn't a condemnation - I'm first and foremost a valve guy! If a mechanically introduced 'distortion' makes the music more pleasurable and lifelike, I'm actually HAPPY). With chips, I can't see microphonics being present to anything like the same extent, or being of the same nature, due to the construction of the device (won't claim there are NO microphonics without testing - that's always a risky claim).
I'm sure that there are other mechanisms involved in different valve tonalities, but it would also be interesting to take valves with the same characteristic curves (like the 6V6 and the 6AQ5), but different envelopes, and listen for any audible effects that might be due to the different physical construction.
>>>...I suspect the chips has a average lower spectrum of harmonics than the common valve (ot ZEN - like) amp, but amount of harmonics over output voltage is different. The ZEN f.e. has very low THD at low levels, and THD goes up with power...<<<
Yep, that's "valvelike" indeed - I've got most of the parts for a Zen (sans PS trafo and caps), so maybe I should build one of those too - if nothing else, for comparison.
>>>...The chip has relative much more THD at low levels, going down to to a min. at maybe 50 or so% of maximal power output, and climbing again. The distortion at low power levels is heavy crossover distortion...<<<
Hmmm, my understanding (perhaps VERY flawed!) is that chips usually use laser trimmed parts with very close values to each other. There's also the issue of thermal tracking of devices on a common die vs discrete devices - in other words, parts matching and thermal tracking is to such a high degree that the crossover from one device to the other is minimised in a way that discretes can't manage.
All I can say for sure is that this is the first solid state hifi amplification I've ever heard that didn't sound 'gritty' or 'unpleasant' to my ear; rather an exciting thing to discover, actually. The "reason why" is still very much open to debate. Maybe if I do build that Zen and find a second type of sand amp I can tolerate, I'll have some answers.
>>>...So with very efficient speakers (you will need to reproduce dynamic range) you always listen a chip amp at power levels with a lot of crossover distortion. With a SET or ZEN at 0,01 to 0,1 W average output power you alway listen with much less distortion than with the chip....<<<
Hmmm, I don't know that I need hyper-efficient speakers with the 16 wrms my TDA2030 puts out, but I've hooked it to my 96dB efficient Fostex' and it sounds pretty good there (the 2A3 still has the goods on it though - thank God, I'd be despondent if a $50 chipamp were to beat out my 2A3, given the amount of $$, research, and effort I put into building it....). At the moment it's hooked to my 89dB SPL RS40-1354/RS40-1270D's (you can see pics of 'em at http://home.earthlink.net/~shidenkai/index.html Since they cost about $45USD to build, it seemed appropriate! ;)
>>>...I use a chipamp for my computer soundcard, as whats coming out there is allready ****, and a class a amp burns a little too much expensive energie for this purpose...<<<
Similar concept here - my TDA2030A mostly sees duty with TV sound (why put hours on a $100USD set of valves for TV?!). Also, I use it a lot when I'm going to be using the hifi for less than several hours - I don't like short cycling my valves. When the 2A3 goes on, I'm going to use it all day (built the PS to CCS standards, so it'll pull 24 hour duty shifts without getting too warm).
>>>...hey, iīm not sure about any thing, and iīm only a mechanics student. Not EE...<<<
No problem! I'm just a physicist who's retooling to be a chemist (more $$ to be had there in industry). EE is "just" a hobby (okay, obsessional devotion is more like it), but I'm still just learning too....and making plenty of mistakes along the way.
Thanks for some really interesting posts, Till. You've made me think and that's always risky! ;) This is proving to be an interesting thread....
All the best,
Morse |
|
|
| h_andree |
I'm experimenting with different version of
the NI LM3875 design. I find that the sound is
very good. I've not finished the project but I'm going
to perform an AB comparison with my SET amps.
I find that the sound varies with the components.
I tried 6 different resistors from - input to ground
and they all sounded different. Some had better
dynamics, some sounded a little nasal. When
you do a comparison you should first optimize the
GC.
I think that a chip-amp with the right components
attached to it can perform similarly to an excellent tube
amp.
The main reason why I'm interested is because I want
to have one amp per driver. It's very unpractical to
use 1 class A monoblock tube amp for tweeter,
midrange and bass driver. Chip amps give me the
capability of using active cross-overs and a nice amount
of power per driver.
Harry Andree |
|
|
| jean-paul |
| quote: | Originally posted by Ropie
It never ceases to amaze me how quickly the same old arguments rear their heads again and again, even in postings that began on entirely different footing :D
I don't think you will benefit any more from this thread Mr Terrorizer and i would suggest you do a 'search' on this and other forums and webpages, because that is a much more efficient way to get the information you require ;) |
Well Ropie, what stops you from giving the required information except for pointing at Decibel Dungeon ? You're right about doing a Search though. Everything has been said and all there is left is repeating it.
The National LM3875 and its brothers are not the most popular just because of the sample program.
Terrorizer ( friendly name ), you can try the LM3875 without any problem on the Peerless speakers. The price could be a reason too. In BrianGT's group buy they'll cost you 4 $ per piece. A complete kit with the insulated version of the chip and PCB's will set you back 70 US $. Just have a look here:
http://brian.darg.net/order/ |
|
|
|
