Need help to upgrade this opamp

[Dyonight]

Hi!

I have a SSL Alphalink Mx madi converter and I'd like to upgrade the opamps in the A/D section.

Those inside are 4565 opamps (there's some other number that I don't understand... please see picture).

What could be  a drop-in replacement for those that would improve transient, response and lower distortion and noise a bit?

I have read somewhere the "NJM4580" could be a better option for a NJM 4565.... but mine doesn't mention NJM anywhere in the name, is it the same part?  Would the NJM4580 be the best performing upgrade or is there something better out there?

Thanks!

 

[Andy Peters]

Hi!

I have a SSL Alphalink Mx madi converter and I'd like to upgrade the opamps in the A/D section.

Those inside are 4565 opamps (there's some other number that I don't understand... please see picture).

Likely a date code.

What could be  a drop-in replacement for those that would improve transient, response and lower distortion and noise a bit?

I have read somewhere the "NJM4580" could be a better option for a NJM 4565.... but mine doesn't mention NJM anywhere in the name, is it the same part?  Would the NJM4580 be the best performing upgrade or is there something better out there?

‘“NJM” is the prefix the New Japan Radio Corp puts on their op-amps, in the same manner as TI putting “TL” and “OPA’“ on their parts, Analog Devices using “OP” and the former National Semiconductor using “LM.”

We could recommend a dozen devices that spec “Better,” but the question is would you notice, and are they optimal for how the device is used in the circuit?

The first question I’ll ask you is how good are your SMT rework skills? Are you willing to risk destroying the PCB on a $3000 product?

 

[Monte McGuire]

A/D converters are brickwall lowpass filtered, and the passives around the amplifier determine the circuit's transient response, not the op amp, as long as the amp is fast enough. This is a 10MHz amp, so it's time and frequency response are controlled by the passives.

If you remove too much noise, it could "un-dither" the converter and make it sound worse. You'll have to try it out and see, barring any schematics or knowing the ADC chip in question.

Lower distortion is possible. The NJM 4580 has higher gain bandwidth and thus lower distortion. If you spend a few cents more, the TI LM4562 or LME49720 will outperform any NJM chip re. distortion. You run the risk of instability using a faster op amp, but a modern surface mount PCB is likely to work well, and the LM4562 has a generous phase margin, and it'll tolerate load capacitance pretty well. Of course, this depends on the ADC chip and how the circuit works. Any chance of finding a schematic?

 

[Dyonight]

Likely a date code.

‘“NJM” is the prefix the New Japan Radio Corp puts on their op-amps, in the same manner as TI putting “TL” and “OPA’“ on their parts, Analog Devices using “OP” and the former National Semiconductor using “LM.”

Thank you for the clarification!

We could recommend a dozen devices that spec “Better,” but the question is would you notice, and are they optimal for how the device is used in the circuit?

You're very right, the problem is it looks like something went wrong and one opamp died and as causing pop and white noise sound, so I had no choice but to replace it. I had some opa2134 around  (I read "somewhere" that it is a drop in replacement) so I replaced it and now the noise is completely gone.

The first question I’ll ask you is how good are your SMT rework skills? Are you willing to risk destroying the PCB on a $3000 product?

I take the necessary precautions so yeah I can make it work!

 

[Dyonight]

A/D converters are brickwall lowpass filtered, and the passives around the amplifier determine the circuit's transient response, not the op amp, as long as the amp is fast enough. This is a 10MHz amp, so it's time and frequency response are controlled by the passives.

Is the OPA2134 be a suitable replacement? All I think I understand so far is that it's a fet input op and the original 4565 is a bipolar input op. Not sure I understand the what it mean though...

You run the risk of instability using a faster op amp, but a modern surface mount PCB is likely to work well, and the LM4562 has a generous phase margin, and it'll tolerate load capacitance pretty well.

Would it be better suited to replace the 4565 or the OPA2134 is good enough? So far it does an excellent job!

Any chance of finding a schematic?

If it's mandatory I'll ask SSL if they can provide one for maintenance. The Mx serie is discontinued anyway so maybe they won't care too much giving that away.

 

[Monte McGuire]

Is the OPA2134 be a suitable replacement? All I think I understand so far is that it's a fet input op and the original 4565 is a bipolar input op. Not sure I understand the what it mean though...

Would it be better suited to replace the 4565 or the OPA2134 is good enough? So far it does an excellent job!

The OPA2134 is a good sounding amplifier, probably nicer than the 4565, but there are cleaner amps today. If it's working well, then maybe there's no point in replacing it. Its errors will be to sound a little brighter than reality, but in a nice way. If you wanted better "transient response", a cleaner amp might be disappointing.
[/quote]

 

[Dyonight]

The OPA2134 is a good sounding amplifier, probably nicer than the 4565, but there are cleaner amps today. If it's working well, then maybe there's no point in replacing it. Its errors will be to sound a little brighter than reality, but in a nice way. If you wanted better "transient response", a cleaner amp might be disappointing.

Thanks for confirming compatibility!

As the sound goes, there's nothing to complain, it perform admirably, sounds kinda "alive" (preferable to "real" IMO) and it's very stable.

The previous 4565 were very likely defective but the noise floor was all over the place in TotalMix, some channels were around -104db while other were around -85 and some channels had hickups and noise burst. The OPA2134 stay around -105, -103db stable on all channels, which is a huge improvement!

Thanks a lot for all your inputs! It's been very helpful!