Burr Brown OPA 2143 vs NJM2068

I'm not sure about repair cost but he is offering very old op-amp not even half as good as OPA2134. I suspect he might have this amp handy. My advice is to replace it with LM4562 - much better amp than both. Be sure to replace it in both channels - this is dual amp but might have both amps in one channel. LM4562 can be purchased from Digi-Key for about $5. Be sure to order right suffix (DIP, SOT8 ...). OPA627 is a single amp (won't fit).

If you still want to go with NJM2068 ask this guy about supply voltage since OPA2134 starts at 5 volt minimum while NMJ2068 starts at 8V minimum.

NMJ2068 datasheet looks pretty old. They claim better performance than 4558 type amps. Nobody remembers Rayteon 4558 since it was designed/used more than 20 years ago.

LM4562 is the newest audio amp from National Semiconductor (they got awards for it) intended for hi-end audio. It has excellent AC (THD=0.00003% IMD=0.00005%) and DC (0.2uV/degC) specifications. Unity Gain Bandwidth is 55MHz and slew rate is 20V/us. It work from 5V to 36V. Power supply rejection is excellent. It has pretty strong drive (600 Ohm specified). As for the known product - Benchmark DAC1 USB uses it in the output stages replacing previously (DAC1) used NE5532. Replacing all chips in Benchmark is probably too expensive for Benchmark ($5 for LM4562, $0.55 for NE5532) since there is a lot of them (about dozen packages) and NE5532 is very decent in the AC (pretty bad at DC).

It is difficult to talk about sonic character since I have never heard it (new product) and Benchmarks design objectives are very specific (promoting natural sound and avoiding warm sound). My audio experience is also very limited.

How it compares to OPA627 - I don't know. According to Benchmark's technical director John Siau they tested OPA627 and got worse results than NE5532. Mr Siau even stated in details which harmonics got worse. On the other hand Steve from Empirical Audio who moded a lot of Benchmarks says that OPA627 is better since it performs better/cleaner at low levels (microdynamics). Steve is probably the person to answer question about sonic characteristic of LM4562.

OPA627 is, as far as I know, a single package amp while LM4562 is dual amp available in DIP, Metal Can and SOIC (SMT) packages from Digi-Key for about $5. LM4562 is very good at AC and DC and having wide banwidth and wide supply range can replace any audio op-amp in the dual package. I think that National created also single packages but with different part number - I can find out.

Word about NE5532 - it was initialy designed by Signetics. Signetics became member of Philips Company and early Benchmarks used thin sounding Philips NE5532. Around year 2001 Philips factory burned down and they stopped making NE5532 selling license to Texas Instruments. TI redesigned them with larger die and the newer Benchmarks that use them have much fuller sound. Benchmark has output cap since NE5532 has pretty bad DC offset and drift.

In general audio OP-Amps are a very narrow class of all OP-amps. If you're not sure if amp belongs to this class download data sheet in pdf and check for THD vs Frequency and THD vs Output Voltage charts. Don't even bother if you don't see it.

TL... is probably TL082 or 72 or 62 or something from this family (pretty bad for audio). If you can go with OPA2134 (dual) then go with LM4562 - it is exceptional amp. It is probably 8 pin DIP thru hole plastic package (2 rows 0.3" apart of 4 pins each 0.1" apart. Pinout of packages is standard - always the same. Both OPA2134 and LM4562 are dual and pinout is the same. Be sure that package is plastic DIP as I described a not a metal can. Metal can package (not very common) can be soldered into DIP (same order of pins) but not the other way around. LM4562 is available in all versions. When you change amps go for the best and most expensive. Few dolars more won't make any difference to you. It makes often difference to manufacturer but sometimes reason is different. They just simply cannot constantly change amps when new and better once are released (every year). It would require prototype testing over and over again and create nightmare in service. I suspect that they never change to better amps without a big reason. Design is treated as closed. Be sure to test your amp with cheap speakers. In spite of op-amps being stable there is always a chance as with any repair of bad solder joint or short - very small chance but still..

Joaco - TL081 is a single amp. It is very old and very cheap amp designed as a improvement over 741 type (one of the first op-amps). It is not an audio amp. LM4562 comes in single version with completely different number starting with LME.... - I have to check again.

What's wrong with using the original part? - without going into technical stuff let just say that Texas Instr. introduced this part in October 1976 and LM4562 was introduced around 2006. Not only 30 years difference but also TL081 was design as universal amp while LM4562 was designed specifically as an audio amp (with incredible DC spects).

I was refering to Joaco post about power amp that has TL081. As for 5532, it depends - if it's thin sounding Philips amp I would replace it, but if it is TI amp I would leave it. But if your amp has 5532 it probably has output capacitor (like Benchmark DAC1). Change to LM4562 and you might be able to remove this cap.

Joaco - AD818 is a video amp. It could be used in audio applications (Rowland uses video amps in Capri preamp) but care must be taken. Pinout of the NULL pins is different so verify if NULL feature is used. In addition bias current is much higer. 100MHz bandwidth in not always a blessing. Wide bandwidth amps can often oscillate since small circuit capacitance at 100MHz can "eat out" phase margin.

Typical bias current for LM4562 is 10nA. Taking into account possible unbalance of 10k we get 0.1mV unbalance.

LM4562 has excellent AC characteristics but the sound is what matters. Benchmark uses it as a driver for XLR outputs because of the ability to drive 600 ohm (not necessarily because of sound). Getting rid of output cap would be a good thing but my Benchmark is still under warranty. NE5532 has bad rap on DIY forum - perhaps because it is so cheap. Somebody tried already replace NJM2068 with OP2132 and complained about metallic sound.

I do not have experience with LM4562 so I would not push for replacement of OPA627 or OPA2134 but original question was about replacement for NJM2068 - pretty old dual amp. Bias current runs here 150nA typ. and 1000nA max (it doesn't get much worse than that) - LM4562 is good candidate here.

Input offset current is 11nA typ and offset voltage 0.1mV. Total of 0.3mV giving generous 10k unbalance. Gain after DAC is practically none. Sub 1mV offset might be a problem for some power amps but not for mine (5Hz-65kHz).