|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
In Reply to: Re: OPA2134 vs OPA2604 vs OPA637 posted by eagle1 on January 30, 2002 at 13:17:56:
Original opamp was OP275 from AD. It's no match to any of the three opamps mentioned above. I'm not sure if OPA627 is better or different than 637 (sonic wise) in any way. Otherwise I also have top grade 627's that I can try.Yeap I'll report back...
Follow Ups:
doubt that top grade, i.e. lowest offset voltage and/or military tempareature will do any good in audio applicationsany FET input op amp will have problems when used in non-inverting mode due to common-mode-input-voltage-dependend input transistor capacity causing distortion
the 6x7 may be less prone due to alleviating action of cascode circuit
the 627 and 637 are the same except for minimum noise gain - in applications above noise gain of 5, the 637 will offer higher bandwidth which helps to reduce hf and intermodulation distortion
look at D. Self web site - he tested op amp performance with an audio precision one unit - not the most meaningful test, but still, it is interesting to observe discrepancies to manufacuturers' data sheets! the OP-275 turns out to be pretty poor btw!
i have had good experience unsing a 627 in inverting mode and with a diamond transistor output buffer that keeps thermal influence from output load outside of the chip - minimizing thermal coupling and hence LF distortion
for non-inverting mode I recomment a fully complementary discrete op-amp
greetings,
Eric
Hi,> any FET input op amp will have problems when used in non-inverting
> mode due to common-mode-input-voltage-dependend input transistor
> capacity causing distortionI would re-phrase this tro include ALL Op-Amps with "long tailed pair" type inputs, be they discrete, monolithic, Fet or BJT input. All suffer from various common mode induced problems, though FET OPA's do less so then BJT ones.
Only OPA's build on the current feedback complementray technology (AD811, LM6181/82/71/72 and other similar ones) which operates without the long tailed pair input is free from those effects.
Later T
Hi Thorsten,
> any FET input op amp will have problems when used in non-inverting
> mode due to common-mode-input-voltage-dependend input transistor
> capacity causing distortionYou seem to be so sure to emphasise the above line.
Maybe I am totally wrong but with a cascode connected input stage the above effect will be nill. Or are you not referring to the Miller effect of the input stage?.
ELSO
Hi,> You seem to be so sure to emphasise the above line.
Actually, I was commenting on it.
> Maybe I am totally wrong but with a cascode connected input
> stage the above effect will be nill.Not so.
> Or are you not referring to the Miller effect of the input
> stage?.I was refering to common mode effects in general. The original poster referred to common mode modulation of the Input Fets by the signal, unless the cascode elements directly reference the common point of the long tailed pair cascoding is not effective in this case, though the overall magnitude of the effect would be larger if miller amplified.
I'm much more concerend about common mode induced distortion in circuits with a long tailed pair input (BJT, MOSFET, J-Fet - all are affected) if the circuit is operated non inverting.
Ciao T
./
ELSO
Hi Eric,Thanks for some interesting notes. I do have some questions and hope you don't mind answering...
> any FET input op amp will have problems when used in non-inverting mode due to common-mode-input-voltage-dependend input transistor capacity causing distortion
By "non-inverting" do you mean signle-ended signal sending into "+" input? Or do you actually mean "non-differential"?
Also in above sentence you mean "capacitance" instead of "capacity", right? Can you elaborate a bit more on the mechanisms (I already understand Cgd, or Cbc for bipolar does depends on bias conditions) and problems associated with the effect you mentioned above? And why such effect doesn't happen when the opamp is not operated under "non-inverting" mode?
I visited D. Self's site before, pretty interesting. Will visit again to correlate his results with what I heard.
-Ehien
Hi,please do not forget that bandwidth is heavily depending on the closed loop gain. So the ranking of the op amps concerning their sonic characteristics may vary depending with gain settings. This holds especially when comparing OPA627 and OPA 637. At high gain settings the OPA627 bandwith may become insufficient
KlausB
Hello old friend!I just called to say
I AM HERE
groman
Ha! Ha!That's why the Black Cube sounded so bad compared to my own phono stage. Given the price difference, 800$, I was suprised, to say the least.
Thanks Ehien,
Daniel Trudeau
Newer version of Black Cube (mine should be the original version) uses OPA2604, which I believe is the main reason they sound better.By the way, the link below will bring you to a photo of dual-wired 637 sitting in the BC. It's ugly but it sounds good... :-)
Thanks Ehien. Well, it's not that bad, altought, I will never sell something that look like that! But as DIYer...Regards,
Daniel Trudeau
for the moment, anyway- OPA637.After 24hr of burn in (power on but no music playing) I compared the dual-wired OPA637 pair to OPA2134. The differences were much more subtle than the ones between say the groups of OPA2604/2134/637 and OP275/5532/AD746/AD823. I had to listen longer to gradually realize the differences.
OPA2134 sounded more impactful but a bit leaner, especially when compared to OPA637 (not so compared to most other opamps). With the right music 2134 can bring a right amount of excitments.
On the other hand, OPA637 had the quality that draw you in. It had more body compared to 2134 and sounded warmer as mentioned earlier. Sometimes it seemed to lack a little bit of high frequency engery/extension, but it could also be that the treble is now smoother. I plan to leave the dual 637 on my phono amp and do a longer term audition.
By the way throughout this test Classic Record's Kind of Blue reissue was used as the source. The setup was Grado Platinum cart-Rega RB300 arm-Rega Planar 3 table-Lehman Black Cube phono stage-C-J CAV-50 integrated-B&W Matrix 802s3 speakers.
A real winner might not be confirmed until more LPs and types of music are played.
-Ehien
Hi Ehien,It seem's that I'm not the only one playing with op-amp in phono stage these days. I first try NE5532 as input and OPA2604 as output. Then change the 5532 for an OPA2134. Then split the input for two OPA134. Change the 2604 to 2134. Put back the 2604. Change the 2604 for an AD826. The AD826 is clearly above all those. I still have some LT1364 to try as output amp.
This takes me more than an year. Right now, I'm comparing OPA134 to OPA227 to LT1115 for the input stage. After I settle on one of them, I'll do a match between it and the OPA627. I can't use OPA637 because at HF, the gain go under 5, so instability can happend.
Regards,
Daniel Trudeau
Hi all,From replies, I see that's a important topic! A great way to learn BTW.
Ok, as Ehien state, the 627-637 is a warm op-amp, even warmer than the OPA134, wich I dislike. So the OPA627 as input op-amp for my RIAA is out. 40$ worth of useless op-amp, at least for that application.
The 604 being to veiled for my taste, I tried OPA227 and LT1115. The 227 remind me of the 604 but with less HF and midrange edge. It stills manage to give me that "kick ass" bass and midrange speed I like and that I miss in warm sounding op-amp. The proplem is that subjective speed is detrimental to midrange transparency and grain-free highs.
Came in the LT1115. I didn't read enough the datasheet at first so I was wondering why I did get some weird noise in HF, like someone searching to lock on a channel on short-wave radio. After reading back, I discovered that this op-amp don't like to see unity-gain i.e. like the OPA637. But I think: I got these for an year without listening to them, so I tried them anywhay. Wouha, the bass and the midrange became really good. The focus was great and I experienced the right balance of speed and transparency. The LT1115 is a bit warmer than the OPA227 but achieve what the BB can't: absolutely no edge in voice, er...for an op-amp, that's it. The Linear offer my enough speed to enjoy rock, be-bop and massive classics. In the same time, the more delicate music is not trashed as with the 227. I get lost in the sound and finally listen music without A.N. The only problem is that f... trouble with treble. As it can't handled them right, cymbals lost some sharpness and bite.
That experience trouble me. I was thinking that the erratic behaviour at HF should destroy completely the sound, it doesn't append. Now, I think that the best I can do is to swap to passive filtering with a LT1115 as input and any high-speed op-amp at output.
Regards,
Daniel Trudeau
Daniel,Thanks for the inside scoop. I think your preference on LT1115 over other opamps has a lot to do with application. In the preamp part of a phono amp, LT1115 seems to be an ideal device due to its exceptionaly low noise, low distortion, and moderate slew rate. But it might not be a good choice (just speculation as I've not used LT1115) for output stage of a phono stage in driving cables.
I've not touched the preamps (AD SSM2017) used in my Black Cube as they are in the RIAA circuitry. Maybe there are even more rooms to improve, if I can identify a proper replacement (such as LT1115?).
Cheers.
Ehien
Dan:Have you had a chance to read about or actually try the new AD8610?
The datasheet has some interesting charts and info, especially the shootouts with the OPA627.Michael
This opamp and its comparison to OPA627 are very interesting. But this is another damn surface mount component. Look like I'll one day have to design and build a header to accept these ever-popular SMT opamps.
Well, I did worry a little bit about stability before installing 637's. However, as it will take me longer to study the feedback arrangement of Black Cube than soldering the 637's, I simply went ahead and tried them out. Obviously, so far so good. I'll try 627's when I get the time.Look to me AD825/826 have the potential to be very good. However I have no chance to try them. The SMT single 825 is available from LC Audio as dual configured opamps. Maybe you can report back after comparing 627/637 with 825/826.
By the way, I seem to prefer BB (now part of TI) over AD equivalents. For instance AD746 was listed as OPA637 dual equivalent; and AD823 was listed as OPA2604 equ. on AD's cross reference list. However, in my system these two AD's were no match to their counterparts.
Just my thoughts...
-Ehien
Hello again Ehien,I'll try AD826 against 627, but it will take some times. I'll built a newer phono stage soon to incorporate the 50KHz pole and will go full monoblock, batteries PSU too. And no, I wouldn't buy the LC Audio soon, I have enough of op-amp here, there's no way I will pay someone to do a job I can do. But for begginer, the LC Audio solution is easy.
Me too, I prefer BB over AD equivalent. And I've only seen one BB chip going bad so far. AD is not bad, but doesn't give the same MTBF as BB.
Stability have more to do with PS lines than feedback. In my basic design, the OPA627 oscillate. I will fix that today and include it in the match.
Regards,
Daniel Trudeau
no decent op-amp should fail in any decent application! MTBF is a parameter that makes sense in the digital world!unless you are talking about how often a circuit that is on the fringe of oscillation will really oscillate. but this is more a question of design errors rather than op-amp quality
Hi Capslock,Like any digital device, it's a matter of alot of things. You should know that some manufacturers clearly annonce that there products isn't appropriate for life support. Of course, brand new, they should work OK. But because they know that they can't specify enough reliability in rough conditions and the way IC fails, they protect themself. BB products can be use in critical application, even if not in life support. That's what I mean by MTBF for op-amp.
Regards,
Daniel Trudeau
> Stability have more to do with PS lines than feedback. In my basic design, the OPA627 oscillate. I will fix that today and include it in the match.You meant only for OPA627, right? Otherwise for OPA637 since it's not unity-gain stable one has to take care of both feedback (thus gain) and PS to ensure stability.
nt
Hi,> I can't use OPA637 because at HF, the gain go under 5, so
> instability can happend.Well, first insert the missing 50KHz breakpoint in the active RIAA and you will find in most cases, the gain is sufficient for guaranteeing stability.
Ciao T
Hello T,Yes, I'll have to check that. I have a new circuit that take care of this breakpoint, I just need the time to try it. I still not fully understand why a 50KHz pole change so much things. Maybe it's most a phase behaviour than a relative voltage attenuation. Getting the phase right makes the soundstage real.
Regards,
Daniel Trudeau
take a look at the AD797 data sheets for ways to degrade op-amp bandwidththen, there used to be a pretty instructive BB app note called "op-amp stabilility and gettings things right for a change"
Enjoy!
This post is made possible by the generous support of people like you and our sponsors: