Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 21, 2012 15:57:45 GMT
reading reviews of 'C'Moy' amplifiers all over the place, mostly with rail splitters and claims of being very powerfull I decided to build me a testbed and roll some op-amps in it and test for real world specs... Wanna know? download this pdf: www.mediafire.com/view/?p2hnepv9u8dlbjeperhaps a fun thread to discuss various of these amplifiers in general.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 21, 2012 18:16:02 GMT
Looks like the LM4562 is the best performer in that application (no observations and slightly higher power tha LT1361), am I correct?
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 22, 2012 5:59:31 GMT
LT1361, LM4562, OPA2132 and OPA2134 are all good choices for these types of amps. a few % more power isn't really noticeable, you need a good 3dB to become aware, that's a factor 2 in power. To have it twice as loud you need to increase the power by a factor 10 (10dB) Some of the opamps exhibit 'soft clipping' others hard clipping. The soft clipping ones may sound nicer when the clipping border is reached. Didn't make notes about soft clipping.
|
|
|
Post by PinkFloyd on Sept 22, 2012 8:16:58 GMT
Hi Frans, OPA2227, 2604 and 2107 are also very good.... my personal favourites for stability would be the 2132, 2134, 2227 and 2604 with the 2227 being my preferred sonic signature.... you can also try OPA2228 but it needs to be optimised and is not really a direct "dump it in" chip roll.... for dumping in without further ado the 2132 / 2134 really are your friends
|
|
mrarroyo
Been here a while!
Our man in Miami!
Posts: 1,003
|
Post by mrarroyo on Sept 22, 2012 12:28:23 GMT
Awesome report, which I understood half of it much less be able to build one. Is there an offering out there that is comparable to your findings? I mean something under $75 USD that could be bought and meets your conclusions? Thanks.
|
|
|
Post by PinkFloyd on Sept 22, 2012 13:51:01 GMT
Awesome report, which I understood half of it much less be able to build one. Is there an offering out there that is comparable to your findings? I mean something under $75 USD that could be bought and meets your conclusions? Thanks. X2...... that is an AWESOME report Frans, we really are fortunate to have you contribute to the forum with these gems Not only are they interesting, they are THE de facto reference points which can be trusted. All the best, Mike.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 22, 2012 19:31:42 GMT
It's (almost) a copy of the Zigis. In fact about 90% of all those 'altoids' C'Moys have a similar design. Marginal difference is in the input circuit, decoupling of the opamp and added 3.9 Ohm output resistors (for more stable operation). Zigis also uses OPA2134 and LM4562 b.t.w.
What these amps teach us is that you don't need a powerful amp in most cases. Most headphones will give around 90dB SPL with 1mW already (not all of them). 90dB is already pretty loud. Twice as loud (100dB) only requires 10mW which these amps can deliver. To reach yet double the perceived loudness yet again (110dB) you need 100mW already (which these amps can on 60 to 120 Ohm headphones) With very dynamic classical recordings this means you can listen to 85-90dB average SPL and still hear110dB peaks without distortion. The higher wattage amps are needed for 'power hungry' headphones or if you play really loud.
Mind you... these amps won't do that well on 16, 32 and 300 Ohm headphones and is where they act in a less desirable way (especially the 9V C'Moy's). headphones like the K601 (120 Ohm) can play really loud on these types of amps and are a perfect match.
I don't have other opamps to test (the ones Mike mentioned) if someone wants to know if certain (dual opamps DIP-8) qualify and how they test... stick them in anenvelope (well packaged) I'll test and ship 'm back. Shipping won't be more than 1 Euro in 'letterbox' mail.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 22, 2012 19:48:53 GMT
Not half. It always amazes me how well the Zigis drives the K601. You just wonder where all of the volume can possibly come from.
Mind you... these amps won't do that well on 16, 32 and 300 Ohm headphones and is where they act in a less desirable way (especially the 9V C'Moy's).
Why is that, Frans? I don't get why those particular impedances don't work so well, yet the 120 K601 is amazing!!
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 22, 2012 22:25:07 GMT
That's because op-amps cannot supply much current (nor were they intended to). For low impedance headphone you need relatively high currents and low voltages. For high impedance headphones you need relatively high voltages and low currents. For medium impedance headphones you need average voltages (9V is already decent, 18V is better) and the current drawn at those voltages isn't as high as when low impedance headphones are used and thus the opamp can handle this load without too much problems.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 23, 2012 7:18:10 GMT
Oh, of course. I get it now. 120 ohms is the optimum between too much current or too much voltage. Thanks Frans.
The Zigis sounds so powerful that you tend to forget that it only has a 9v battery.
Maybe there's something about the simplicity of the design which seems to make CMoys sound good?
It does seem quite funny that a good headphone can sound so good from such a cheap amplifier!
Sent from my iPod touch using ProBoards
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 23, 2012 8:50:12 GMT
That's because op-amps cannot supply much current (nor were they intended to). Frans, does this mean that low impedance cans are fine at low volumes, but turn the wick up and most op amps start to struggle to supply enough current? What I'm getting at, what exactly is better then for a low impedance can which requires so much more current? Obviously high impedance requires a big voltage so we go back to power supply but what about supplying oodles of current via op amps? Valves are the same as well in that they seem to prefer higher impedances.
|
|
|
Post by gommer on Sept 23, 2012 9:35:06 GMT
Frans, just wondering. Did you ever experiment with opamps + output transistor driver stage within the feedback loop? That should resolve the current limit for low ohmic 'phones, but of course it introduces zero crossing distortions. The question is, how well can different types of opamps mask this. It would remain a simple enough C'Moy type amp.
Marc
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 23, 2012 13:18:12 GMT
It would be interesting if we could take something like the the Zigis and make it even more capable, hopefully staying within the confines of the existing enclosure Especially with 32 0hm in mind as that is the rating of my fave portable use cans (senny HD238)
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 23, 2012 14:31:18 GMT
One thing you can do with the Zigis Chris, is plug it in. Up to 30v as well, I think it is.
Sent from my iPod touch using ProBoards
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 23, 2012 15:03:16 GMT
Frans, does this mean that low impedance cans are fine at low volumes, but turn the wick up and most op amps start to struggle to supply enough current? Indeed. What I'm getting at, what exactly is better then for a low impedance can which requires so much more current? special opamps as the TPA6102 or output stages as those from the Sansa series. C'Moys with NJM4556 e.t.c. Obviously high impedance requires a big voltage so we go back to power supply but what about supplying oodles of current via op amps? It's what the CHAmp does... it's essentially a C'Moy with 4 opamps in parallel to form a 'power' output stage (with some extra configurable features). Valves are the same as well in that they seem to prefer higher impedances. Indeed especially the OTL (output transformer less). These are so well liked because of the high output impedance, similar to adding 120 or sometimes even more series resistance. Amps with transformers (especially those in a feedback loop) do not exhibit this trait.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 23, 2012 15:09:39 GMT
Frans, just wondering. Did you ever experiment with opamps + output transistor driver stage within the feedback loop? That should resolve the current limit for low ohmic 'phones, but of course it introduces zero crossing distortions. The question is, how well can different types of opamps mask this. It would remain a simple enough C'Moy type amp. Marc The opamp version of the SC amp, the output stage of the Aune T1, The Larks Guanzo and some soundcards e.t.c. all use this topolgy. I never made such a circuit nor specifically tested it as it simply works. With a class-B added output stage in the feedback loop you would be asking for crossover type distortion. How much would totally depend on the openloop gain and frequency range (slewrate) of the opamp. Most of these transistor output stages have a small AB class setting with the aid of 2 diodes. a 'C'Moy' by definition is a design where the opamp is the gain+output stage.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 23, 2012 15:14:08 GMT
It would be interesting if we could take something like the the Zigis and make it even more capable, hopefully staying within the confines of the existing enclosure Especially with 32 0hm in mind as that is the rating of my fave portable use cans (senny HD238) Put in an NJM4556 and add 2 output resistors, (between 3 and 5 Ohms) and increase the battery voltage to 18V. It would essentially become a single opamp O2 (the O2 didn't have those output resistors added that's blasphemy in his court).
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 23, 2012 21:25:38 GMT
Both very useful.
Frans,
When using the NJM4556 chip is it hand in hand with the extra battery (external battery pack)? Or will it be happy on a single.
Cheers,
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 23, 2012 21:57:16 GMT
The NJM4556 should work down to 4.5V so it should work O.K. on 9V. The opamp will be happier on 18V and would increase output power considerably (a factor 4 or so). More headroom doesn't hurt (like Ian already stated)
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 23, 2012 22:04:15 GMT
Thanks Frans.
|
|
|
Post by scappers on Oct 1, 2012 13:06:40 GMT
|
|
jonclancy
Been here a while!
Mr. Ripple Eater
Amateur EAGLEist
Posts: 1,131
|
Post by jonclancy on Nov 18, 2012 11:48:04 GMT
I don't have other opamps to test (the ones Mike mentioned) if someone wants to know if certain (dual opamps DIP-8) qualify and how they test... stick them in anenvelope (well packaged) I'll test and ship 'm back. Shipping won't be more than 1 Euro in 'letterbox' mail. Ah, a bit late to the party. If you're reading this, Frans, I have a selection of OPA that might be interesting to test. Including a load on SOIC devices on adapters. CHeers Jon
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Nov 18, 2012 13:47:25 GMT
I don't have other opamps to test (the ones Mike mentioned) if someone wants to know if certain (dual opamps DIP-8) qualify and how they test... stick them in anenvelope (well packaged) I'll test and ship 'm back. Shipping won't be more than 1 Euro in 'letterbox' mail. Ah, a bit late to the party. If you're reading this, Frans, I have a selection of OPA that might be interesting to test. Including a load on SOIC devices on adapters. CHeers Jon Same goes for me Frans - when I was first messing around with my Beresford 7520 DAC a friend sent me quite a few to try. If you're still interested I could ship you the lot for testing. Cheers, Dave.
|
|