[ Home | Staff & Contacts | HiFi Playground | Listening tests | DIY & Tweakings | Music & Books ]

Author: Paul Hunting - TNT UK
Published: March, 2023

Introduction

This piece is a listening comparison between five possible replacement op amps for the NJM4560 fitted in the Philips CD753. The NJM4560 op amp is also fitted in the Marantz CD5000 and numerous CD players.

The five op amp options were selected from a sift through of about a hundred or so op amps. My criteria for finding a replacement op amp is that the replacement should be low noise, have a highish common mode rejection ratio (CMRR) value. In addition the replacing op amp should have as high as possible slew rate. I was strict in my criteria for a low noise op amp, I was semi strict about CMRR rate at the top octave as this metric gives a good idea of top octave instrument performance. I was less dogmatic about slew values. Other users will have their own criteria. I also sought op amps likely to be compatible with the CD753 circuit. I define an op amp as compatible where the replacing op amp had specifications close in values to the specifications of the target op amp (NJM4560) with regard to Bias Input Current, Gain, Current draw and Voltage Range specifications. The CD753 circuit is optimised for the NJM4560 and selecting an op amp with specifications close to this op amp should result in a stable performance of the replacing op amp in the CD753.

The five op amps listened to were: OPA1612, MUSES8820, OPA1692, LME49723 and the NJM8801. All five op amps have more than half the noise compared to the NJM4560. The LME49723, MUSES8820 and NJM8801 are fairly traditional op amps. They don't have internal capacitors, there is a capacitor in the NJM8801 to protect a diode, further these op amps need an external bias and gain resistor circuit. These three traditional op amps had a poor low slew rate and a poor comparatively low top octave CMRR value, both which limit reproduction tone. The other two op amps the OPA1612 and OPA1692 are what I'd term super op amps and appear to combine low noise, high slew and a high CMRR value at the top octave. Op amp performance metrics have been previously reviewed.

Philips CD753

The CD753 host player is sound challenged by its DAC in that the fitted TDA1549T only has a second order noise shaping filter and lets through too much shaped quantized noise. This means that instruments played in the top octave can sound exposed and sometimes hard (using the OP275 and system fitted with a damped metal dome tweeters). For op amp selection ones with a higher value Common Mode Rejection Ratio in the top octave should limit this high frequency exposure.

Op Amps

I previously used the OP275 op amp in the CD753. The OP275 is a replacement for the NJM4560. But I wanted a more detailed op amp than the CD753. I also suspect the OP275 is an off fit in this CD player with regard to common mode voltage range giving rise, playing some CDs, to mild OP275 transistor saturation and to a, not unpleasant, bloated sound. It took about a year to fully realise this.

LME49723 vs Muses8820

On paper the LME49723 is a better specified op amp, it has lower noise and a higher CMRR value compared to the Muses8820. Listening to the Morton Feldman opera (Neither) there was more detail with the LME49723 fitted. The Muses reproduction of the same piece was rounded, less detailed and not so many instruments could be heard. With Black Angels by George Crumb (Stanislas Quartet) the LME49723 played this with too much clarity and there was not enough body to the high frequency violin notes. The Muses8820 tonally gave this piece a more interesting rendition, and violins had more body here. This pattern was repeated for other CDs. If I had to make a choice now I would opt for the detail of the LME49723 and try adding playback warmth elsewhere. But there does appear to be an issue with a lack of top octave warmth and treble instrument weight with the LME49723 fitted in the CD753.

LME49723, Muses8820 & NJM8801

Allegrasco is the third track on North Shore (CD) by Gavin Bryars (Harmonia Ensemble), it is a duet between clarinet and piano. The MUSES8820 played Allegrasco with too much warmth and there was distortion to the clarinet at the beginning of this piece. The NJM8801 played this same track with what could be described as manic overtones which ruined the presentation. The LME49723 captured the instrument separation, it also reproduced the change of moods and tempo of this track and better presented its progression. With Track two Intermezzo a play of clarinet, piano and cello, with the LME49723 fitted, it was easy to distinguish between the instruments, even when they were playing a similar note progression. Both the MUSES8820 and the NJM8801 lost the clarinet detail with this track. Instrument separation was poorest on the NJM8801. North Shore the first track is a duet of cello and piano. With the MUSES8820 there is detailed insight into the piano playing and this added to the performance; at least for me. The tone, with the Muse, came across as more sonorous rather than giving a rendition of individual instrument tone. LME49723 played this track with better tone and separation. At times the cello bow strokes came across and could be heard. The NJM8801 played the cello, and at times the piano with distortion. There was no similar treble distortion, playing the same track, with the LME49723. In all the LME49723 in the CD753 played this Gavin Bryars CD the best while the MUSES8820 gave a passable, and less detailed, performance.

The MUSES8820 is a warm op amp, and you may need to be careful with partnering equipment as the sound might tip over into lushness and bloating if your system is already warm.

LME49723 vs OPA1692

The LME49723 has less noise than the OPA1692. This is noticeable with ensemble pieces. Listening to The Beating of Wings (Andrew Poppy) instrument separation was better on the LME49723. The piano drifted in and out of focus with the OPA1692 but was better separated and more focused with the LME49723 in place. The OPA1692 seemed to give this piece a faster playback and notes are exactly where they should be. The performance of the OPA1692 was good, top octave instruments were well presented, with good tone and the whole presentation was listenable. On some material though the OPA1692/CD753 seems to lose dynamics. Is this a result of a mismatch in the op amps higher typical required input bias current (300nA) compared to the target NJM4560 op amp figure (40nA) and as delivered by the CD753. If the host CD player is providing too low an input bias current to the OPA1692 then this may affect how the op amp internal capacitors are charged, affecting op amp dynamics and bandwidth. But the OPA1692 sounds fast, dynamic and had good timing on most CDs. The timing is better on the OPA1692 than the LME49723. In the end, though, I found the OPA1692 too sophisticated and on the reproduction of some music this turned the presentation into lush romanticism which is fine if you like this; lushness may be a characteristic of the loss of dynamics. The OPA1692 is the better op amp but does not suit my taste so well, I also found it inconsistent between CDs and I question its fit in the CD753. In the end I just preferred listening with, the too much clarity in the top octave, LME49723.

LME49723 vs OPA1612

I was not expecting much from the OPA1612 as the gain specifications are not compatible but from fitting it produced a listenable sound. However it was apparent that the OPA1612 noise performance was not met in the CD753, probably in part because of the CD753s external to the op amp feedback resistor and this adds noise (see OPA1612 datasheet). I'm wondering if this resistor can be removed or replaced by a wire link for this op amp. But the OPA1612 starts at such a low noise that the addition of this resistor noise still means that this op amp should be low noise. However, noise of the OPA1612 also rises with source resistance (see datasheet). The TDA1549T DAC fitted in the CD753 outputs a load resistance of 3k ohm, so this too will add to this op amps noise.

There is a gain loss with the OPA1612 but the higher performance of this op amp (high slew and high CMRR) was compensation enough and the gain loss was not so noticeable on most material. There also seems to be some added, but not unpleasant, distortion but looking at the datasheet I can't really see where this is coming from; there could be a voltage anomaly somewhere or perhaps the Philips CD753 is just not delivering enough current to the op amp. The distortion is worse with soprano vocals. For the OPA1612/CD753 combination slower uncomplicated music sounds better and here the instrument tone and performance reproduction is spot on. Listening to a multi instrument or a faster pace piece and the OPA1612/CD753 produced a veiled uninteresting performance. There was more detail, interest and clarity playing the same pieces with the LME49723 fitted. The LME49723 has less top octave instrument tone though. Civil Wars by Philip Glass was painful with the OPA1612 in place with not enough detail and too much vocal distortion, but this CD played fine with the LME49723. The OPA1612 is not such a good fit in the CD753.

LME49723

The LME49723 is a fine match in the CD753 despite having a high typical input bias current compared to the NJM4560 and a higher current draw than this CD player can deliver. The gain against frequency curve of the LME49723 (datasheet Fig 44 page 11) is near identical to that of the target op amp, the NJM4560 (datasheet page 3), and implies that at least the surrounding op amp bias and feedback resistors of the Philips CD753 are compatible with the LME49723. Further the common mode voltage range (+- 14v) is also compatible between the CD player and this rail to rail op amp. Played dynamics with the LME49723 are good. There is though a barely noticeable perceived muddle to some complex material (Yo Shakespeare by Michael Gordon, Iceberg Ensemble) and here the music comes across as a bit slow, lumpy and disjointed, perhaps the 10% increase in 1/f noise of the LME49723 is the culprit here. Even given this, I preferred the LME49723 performance. Read more on op amp compatibility issues here on TNT-Audio.

Conclusion

I am leaving the LME49723 in place. I like the detail, soundstage and clarity of this op amp. There is a question of top octave exposure and instrument weight but the top is rarely harsh or uncomfortably exposed. And top octave tone loss may be minimised by tweaking the playback system elsewhere. Others might prefer the Muses8820, the OPA1962 or indeed the OP275. These op amps offer a less detailed performance than the LME49723 but have a warmer tone.

I rejected op amps like the NJM4562 because the CMRR (70db) at high frequencies is too low. But this op amp is a low noise fit for the CD753 and a replacement for the NJM4560 but don't expect much in the way of top octave tone. I thought the OP27 would work but the input offset voltage is out of range. Also the AD8599 looks as if it would fit but this is expensive, further input offset voltage is also out of range compared to the NJM4560 and the op amp also draws a high current. And the LME49990 looks OK but it is pricey, draws a highish current and gain looks high though the datasheet does not give enough information about gain. The number of replacement options for the NJM4560 increases if you are willing to accept a higher noise value.

In general, performance did match op amp performance metrics as stated in the datasheet. Specified low noise op amps produced a more detailed presentation and a better sound stage, op amps with a high CMRR rate at the top octave had better top octave tone and high slew op amps gave better tone. The Muses8820 largely debunked this relationship however and produced a warm sound that did not reflect its stated metrics. I would argue though that this is a sonorous blanket warmth and does not reflect instrument tones. Saying that, the warmth it produces is very listenable.

So called super op amps like the OPA1612 and OPA1692 that seem to offer everything: low noise, high slew and high CMRR values, appear sensitive to their surrounding circuitry and their performance can decline when installed in the wrong circuitry. However they have a good top octave performance and tone.

The most difficult thing in rolling op amps is finding op amps that are both a performance upgrade and that are a compatible fit. Op amps are not designed to be upgrades for 20+ year old CD players and perhaps it is too much to expect that they would work well here. I suppose I just struck lucky with the LME49723.

Postscript

I added top octave warmth to the LME49723/CD753 combination by changing the speaker tweeter series crossover capacitor.

© 2023 Paul Hunting - paul@tnt-audio.com - www.tnt-audio.com

[ Home | Staff & Contacts | HiFi Playground | Listening tests | DIY & Tweakings | Music & Books ]