Report back after, say, 7 days...
Too much of a good thing?
I'm using 4 of these babies: 2 in CD player, 2 in preamp (2x gain stages).
I think it's perhaps too much and it maybe just makes the sound a little clinical and uninvolving, good though it is. I might now try a pair of OPA627s to cover one gain stage in the preamp.
Would I be better changing the first or 2nd gain stage? The design is the Rod Elliott project 88. I think I'm only using gain on one of the two op-amps, not sure which though! The pre was designed to work with OPA2134 but others seem to offer more (certainly LM6172 gives great excitement, at the expense of smoothness.)
Simon
I'm using 4 of these babies: 2 in CD player, 2 in preamp (2x gain stages).
I think it's perhaps too much and it maybe just makes the sound a little clinical and uninvolving, good though it is. I might now try a pair of OPA627s to cover one gain stage in the preamp.
Would I be better changing the first or 2nd gain stage? The design is the Rod Elliott project 88. I think I'm only using gain on one of the two op-amps, not sure which though! The pre was designed to work with OPA2134 but others seem to offer more (certainly LM6172 gives great excitement, at the expense of smoothness.)
Simon
earl of sodbury said:Follow-up to my earlier query
Obtained some LME49710NA, popped them in the cradles Stello so thoughtfully provided for their OPA627 predecessors in the CDA320's output stage...
And
Bass improved in every way, better sense of acoustic space, and a more natural instrumental timbre are the immediately obvious benefits, and that's straight out of the tube.
If you "believe" in run-in then there should be more to come...
Downside is the extra insight is revealing of DAB (routed thru the Stello's DAC) radio's HF oddities.
Otherwise the best single-component upgrade I've yet tried. Thanks all for the advice.
churz, eofs
So what is the circuit your using it in? I/V, inverting, non-inverting, what gain, what values of feedback.....
That op amp family is unity gain stable, into 100pF. Some care needs to be taken though to prevent boosting or ringing.
If your running a gain of 10, but have a feedback cap, guess what, that's unity gain as for as the op amp is concerned.
You may know all this, and sorry if I'm stating the obvious. The 4562 is also much higher bandwidth, so a layout that's suffecient for a 20Mhz op amp may be marginal for a LM4562, although in my experience its not a squirly op amp by any means, it could be your "radio HF oddities".
The other common mistake is that people sometimes think bypassing should be vrom +v to -v. That might be fine for typical audio bandwidths and tame an op amp, but the the bypassing really should be +V to gnd, and -V to gnd, because the op amp is delivering its current to ground. If not, and the op amp is moving HF signal to the load, then it has to find a path +V to gnd for example, and that can have lots of inductance and cause issues.
A lower bandwidth op amp just won't do that.
Its becomming a novel, so i'll stop writing until I get some feedback.
PortlandMike
Ryssen said:
You report that it sounds great, so I suspect you have no issues.
That said though, the reason for using a feedback cap is to roll of the high frequencies.
If you need to roll off the highs, then do it.
If not, then running it with some gain at high frequencies will make the gain and phase margin huge, and make the transient response better. It also improves the capacitive load driving significantly and reduces ringing if there is any.
Try it and let us know what you find.
If you don't drive much cap, then all you need is a gain of 2 to get almost all the benifit. If you wanted some roll off from a cap you could add a resistor in series with the cap to stop the roll off at very high frequencies. Keeping the HF gain at 2 or so gives you almost all the benifit in gain and phase margin. I'm sure I'm confusing things though.
Another point is that if your running large values of feedback resistors the cap might be required to minimize peaking in the freq response. The comment here is "don't run high values of feedback resistors." This is audio man, and with this op amp I think you should run the feedback R's in the 1k to 2k range. Its got the snoot.
A good friend tells me a CRD to +V is good too, I think that's been mentioned already though.
Bottom line is, try it, listen and see!
Best Regards,
PortlandMike
Tried without the feedback caps,sems the highs opened up a little more,my dumbass why did I use that cap?Thanks for telling!🙂
Will listen more..
Will listen more..
Hi Portlandmike and Ryssen,
Following the quick opamp lesson for "non-technical melo-maniacs " ,
IF we have a nice Voltage output DAC wich comes with a drastic digital filter maybe we can save a little cap from the feedback loop of the opamp output amplifier section...unless we have very noisy ambience that could inject some EMI/RF to the signal path...
Is it that simple?
Good musical weekend 🙂 for you
M
Following the quick opamp lesson for "non-technical melo-maniacs " ,
IF we have a nice Voltage output DAC wich comes with a drastic digital filter maybe we can save a little cap from the feedback loop of the opamp output amplifier section...unless we have very noisy ambience that could inject some EMI/RF to the signal path...
Is it that simple?
Good musical weekend 🙂 for you
M
That´s how I see it,dont´t know why else one should use a cap.Maybee it´s better to use scope to be sure.🙂
maxlorenz said:
Maricio,
Not sure I'm following you. Are you talking about a I/V op amp, or an op amp used in the output of a DAC system?
There are other ways to filter out HF rather than a feedback cap.
Like a RC. That method also doesn't depend on the op amp bandwith to work.
I once had a very early CD player that actually had a LC filter. I think it was like 6th order. Sounded better than anything else I could find for some years.
Mike
Voltage output DAC (I/V inside DAC chip)-->buffer amplifier opamp section.
That's how my M-audio superDAC is built 🙂
More things to study...
I asked a scope for Christmas but nothing hapenned 🙁
Thanx
M
I spent some times this afternoon to try different op-amps at the DAC output of an Tag AV32R. Two friends where there. One is a pro pianist teacher, the second one is a violin student. They are not audiophile at all and don't know what is an op-amps...
We tried AD8620, OPA627, LM4562, LM6172 and OPA827 (prototype from TI) and the original OPA2134. The listening session was done one a pair of Dynaudio BM5.
The winner is the LM4562 and by far.
The second is the OPA827, and after AD8620, OPA627, OPA2134 and the last LM6172.
.
We tried AD8620, OPA627, LM4562, LM6172 and OPA827 (prototype from TI) and the original OPA2134. The listening session was done one a pair of Dynaudio BM5.
The winner is the LM4562 and by far.
The second is the OPA827, and after AD8620, OPA627, OPA2134 and the last LM6172.
.
Portlandmike said:
Hi, I'm very much a noob to this and so can't provide a satisfactory answer I'm afraid. I'm not the circuit designer - it's a commercial unit for which no schematics are currently obtainable, TTBOMK. I do know that the analogue ouptput stage is class-A biased and runs commensurately hot, and that gain is higher than average.
In-situ the chips are running thermally stable at ~35 centigrade, and on full gain there are no audible noise issues (at least, nothing that wasn't already there 😉 ) Similarly over a wide range of music and test tracks there are no obvious acoustic anomalies.
DAB digital radio in the UK is transmitted in a basically obsolete, highly compressed, low bitrate MP2 format
Users with hi-fi systems widely anecdotally report audible HF anomalies with a broad range of receivers and accompanying systems - my comment reflected the fact that these known anomalies were more audible with this chip in place. It's my belief that this is due to the very-obviously wider bandwidth this chip has.I've now put around 120 hours worth of signal through the new chips, and as mentioned, with a broad range of musical styles - there is nothing but a very significant and clear improvement to my ear / subjective preference.
Midrange reproduction has now "caught-up" with the early and obvious gains to LF and HF reproduction, the sense of naturalness and sonic authenticity in particular is substantially greater, and there is no accompanying "sterility" to the sound.
For my tastes there are no downsides, and careful listening has failed to reveal any distortions or unwanted signal gains or losses. I'm very, very pleased with the results of this admittedly haphazard trial - it's by some margin the largest single gain in performance that a single component change has ever wrought anywhere in my system 😀
I entirely accept this implementation (and this answer!) are probably sub-optimal, and hope that as my knowledge slowly grows, both can be improved!
cheers, EofS
Having some trouble getting the LME49740 Quad and LME49740 single.Can´t get any sample,and going to order from Digikey soon,but they don´t have them.
Where do I get them,or any one have some to sell to me?
Maybee a group buy?
Where do I get them,or any one have some to sell to me?Maybee a group buy?
Portlandmike said:
Actually, the feeback cap forms a RC filter in the feeback of the op-amp. It's just that this only works on a NI topology (due to the R path to ground). It can be done on inverted topology, but is far more complex. The confusion is that NI has a minimum gain of 1, so unlike a conventional RC, you can't totally eliminate the HF portion, only reduce it in comparison to the rest of the signal (assuming G > 1).
Fenris said:
Fenris,
I may be confused in what you saying, but here is a stab at it.
On NI configuration, that is non-inverting:
T=Gain=Vout/Vin= (Zf/Zi+1) Thus is Zf is a cap at high frequencies, then you still can't get gain less than one.
I think that's what your saying above, but then that would be an inverting amp.
For an inverting amp:
Gain=Vout/Vin=Zf/Zi, so you do actually get the filtering action desired, but with the potential issues I mentioned for the op amp gain and phase margin.
I don't see what's so complex about filtering in the inverting configuration. Its the easy one!
I think this is all extra words because I suspect that was what you were saying and I got confused.
Cheers
Mike
Hi,
But You could easily add a RC-snubber-circuit at the output to GND.
This is common practise and can be found e.g. in many Phono stages.
jauu
Calvin
But You could easily add a RC-snubber-circuit at the output to GND.
This is common practise and can be found e.g. in many Phono stages.
jauu
Calvin
Portlandmike said:
Mike, the NI one filters because it's a series circuit to ground. ( http://en.wikipedia.org/wiki/RC_circuit ). The Inverted is not referenced to ground and would be classified as a modified integrator. There are more potential issues with phase and interaction with the inv topology, and usually there's no reason you need gain of less than 1.
Thanks guys for the comments 🙂
I have this excellent book downloaded:
http://focus.ti.com/docs/analog/cat...ctName=slod006b
But no time to study it yet 🙁
Cheers,
M
I have this excellent book downloaded:
http://focus.ti.com/docs/analog/cat...ctName=slod006b
But no time to study it yet 🙁
Cheers,
M
- Status
- Not open for further replies.