Well, those LME49860s should sound EXACTLY like an LM4562, since it's really the same chip that can withstand higher rail voltages. Do the bigger rails provide better sound? I dunno.......maybe? Of course, what I'D like you to do is put sockets in and swap 'em between them and those OPA285s that you like 
sos we can get the results!! My take on the LM4562, and why I have 'em almost everywhere in my audio gear, is that it doesn't "sound" like anything---completely transparent to my ears---you get out exactly what you put in. Some refer to them as "sterile" but for REPRODUCING music, that seems to me to be the best. Not so for my guitar amps, however---the LAST thing I want is flat and clean!! Hooray for Leo Fender!!
sos we can get the results!! My take on the LM4562, and why I have 'em almost everywhere in my audio gear, is that it doesn't "sound" like anything---completely transparent to my ears---you get out exactly what you put in. Some refer to them as "sterile" but for REPRODUCING music, that seems to me to be the best. Not so for my guitar amps, however---the LAST thing I want is flat and clean!! Hooray for Leo Fender!!
Sounds like fun, I will probably join you and give them a shot as well since they are 44V capable and I have +/-20V rails in the two MX-50 and +/-18V in the two 5050..
They may be too noisy for the replay amps, but would suit everywhere else.
YMMV of - course and my recollection (primitive FFT hardware at the time) was that there was no discernible difference in performance. But, and here's the big but, I preferred the original OP285s to the replacement 4562s.
In theory I might still have some of those PMI OP285s around here - will have to look.
They may be too noisy for the replay amps, but would suit everywhere else.
YMMV of - course and my recollection (primitive FFT hardware at the time) was that there was no discernible difference in performance. But, and here's the big but, I preferred the original OP285s to the replacement 4562s.
In theory I might still have some of those PMI OP285s around here - will have to look.
Well, the comparison between the OP285s and the LM4562s will have to rest on YOUR shoulders 
because they're too rich for my blood, at ~ $6 EACH!! Hmmmmm..... I was just looking at the electronics in the KING (ATR-100) and Alastair used LM318s---a super-fast (70nV/µsec) opamp. They are sorta noisy, though (15nV/√Hz), so he used a discrete transistor (CD701) front end on the repro head.
It all just goes to prove my premise---measuring and hearing are different animals. There are things we can definitely hear but can't put a measurement number on (how do you measure why a tube sounds so good?) and things we can measure but can't hear (can you definitely say you can hear the difference between 0.00003% THD and 0.00001% ?).
because they're too rich for my blood, at ~ $6 EACH!! Hmmmmm..... I was just looking at the electronics in the KING (ATR-100) and Alastair used LM318s---a super-fast (70nV/µsec) opamp. They are sorta noisy, though (15nV/√Hz), so he used a discrete transistor (CD701) front end on the repro head.It all just goes to prove my premise---measuring and hearing are different animals. There are things we can definitely hear but can't put a measurement number on (how do you measure why a tube sounds so good?) and things we can measure but can't hear (can you definitely say you can hear the difference between 0.00003% THD and 0.00001% ?).
Many good points raised, dinner awaits so I have to be quick.. TBH I doubt I could ever hear the difference between 0.5% and 0.05% distortion let alone anything better. I suspect in many cases 1% THD unless mixed with supply ripple is not going to be audible either..
I am really debating how much I want to try OP285, they are expensive, 10 would run $56 + shipping + adapters + sockets. So I'll think about that a bit/lot more.. I may have some dips lurking around, but where did I put them? Hmmm...
I believe you can use the same trick on the 5534.. Think I have seen it or maybe even done it and have forgotten about it. (Mists of time thing)
I am really debating how much I want to try OP285, they are expensive, 10 would run $56 + shipping + adapters + sockets. So I'll think about that a bit/lot more.. I may have some dips lurking around, but where did I put them? Hmmm...
I believe you can use the same trick on the 5534.. Think I have seen it or maybe even done it and have forgotten about it. (Mists of time thing)
Hi dotnect335,
He is referring to the use of transistors as the front end of the 5534. Look a the schematic of a Cyrus 3 phono section for an example.
Also, because we are now using super quiet audio spectrum analyzers these days, we can now see what people talk about hearing. That's quite different than when we only had numbers to go by. We are also looking well below anyone's threshold of hearing. In my case (and Kevin's), we can see down to -130dB or so, and up tp just short of 96 KHz. If I fall back on my older equipment (HP 339A + HO 3585A), I can see the audio spectrum to 1 MHz, depending on where the 339A rolls off. The 3585A is good until 40 MHz on its own. We are seeing lower than our own interal noise floor, and well beyond the high frequency response of our hearing. The response we are looking at is to a 24 bit depth (so ~ 123 dB) without averaging, and the math allows us to see much lower into the noise floor.
So these days, if you can really hear something, we most definitely can see it and quantify it.
-Chris
He is referring to the use of transistors as the front end of the 5534. Look a the schematic of a Cyrus 3 phono section for an example.
Also, because we are now using super quiet audio spectrum analyzers these days, we can now see what people talk about hearing. That's quite different than when we only had numbers to go by. We are also looking well below anyone's threshold of hearing. In my case (and Kevin's), we can see down to -130dB or so, and up tp just short of 96 KHz. If I fall back on my older equipment (HP 339A + HO 3585A), I can see the audio spectrum to 1 MHz, depending on where the 339A rolls off. The 3585A is good until 40 MHz on its own. We are seeing lower than our own interal noise floor, and well beyond the high frequency response of our hearing. The response we are looking at is to a 24 bit depth (so ~ 123 dB) without averaging, and the math allows us to see much lower into the noise floor.
So these days, if you can really hear something, we most definitely can see it and quantify it.
-Chris
Doh! Now I see! But I can't find a Cyrus 3 schematic---all the page links have expired.
Is there an affordable audio spectrum analyzer available these days? I used to use an H-P 3585 in the factory back in the day, but those were very expensive!
That is encouraging! Can you "see" this difference between an OPA285 and an LM4562 that kevinkr is speaking of?
Hi dotneck,
The software for the RTX would be Virtins M.I. (Multi-Instrument) which has native support for the RTX, meaning it is calibrated and controls the instrument. That software can be had with options, but the entire package with everything runs about $500 US. Well worth it, steep learning curve - just like Audio Precision or the Keysight unit.
The HP 3585A/B only needs a quiet front end to notch out the fundamental. I used my HP 339A for that, but it wasn't quiet enough. I was doing work beyond where I could see it. The RTX brought instant clarity and understanding since I could now see what I was doing.
This question you might want to put to Kevin. I'm not certain how comfortable he is with the RTX and it's display or not. There are two components to this. Making the measurement, and interpreting the results.
-Chris
Sort of. The RTX - 6001 is the least expensive high performing instrument around. I think it runs about $2,500 US landed. The next closest runner is more like $10K US. The ones I would love to have run about $35K. That would be the Keysight U8903B optioned for stereo & digital, up to 1.5 MHz operation.
The software for the RTX would be Virtins M.I. (Multi-Instrument) which has native support for the RTX, meaning it is calibrated and controls the instrument. That software can be had with options, but the entire package with everything runs about $500 US. Well worth it, steep learning curve - just like Audio Precision or the Keysight unit.
The HP 3585A/B only needs a quiet front end to notch out the fundamental. I used my HP 339A for that, but it wasn't quiet enough. I was doing work beyond where I could see it. The RTX brought instant clarity and understanding since I could now see what I was doing.
I imagine you can see the difference while in circuit. All I can speak of is what I've been doing. So far I've been able to predict how something would sound, or explain why something sounds a certain way. But this doesn't cover things that people might think they hear, but don't really. That's the thing that you have to be so careful of, the stuff that goes on between the ears that manifest as hearing something that really isn't there. At this point, I'm pretty convinced that if there is something there to be seen, we can in fact see it. Keep in mind that this involves interpreting what is seen on the screen as well. Years using the 339A and 3585A seemed to have prepared me for this fairly well.
This question you might want to put to Kevin. I'm not certain how comfortable he is with the RTX and it's display or not. There are two components to this. Making the measurement, and interpreting the results.
-Chris
Ouch!! Way to rich for my blood! But it sounds like an absolutely fascinating device!
I totally agree that interpreting the data is at least as important as measuring it. Could you post examples of what you mean by "predict how something would sound"? With this device, can you "see" the difference in sound between, say, a 6V6 tube and a 6L6 tube? Or the difference between an NE5532 and an LM4562?
Nowadays one can do pretty comprehensive measurements with a good sound card. All you need is a low distortion swept sine generator and a PC. I also use a pink noise generator for quick frequency response visualization and playback EQ adjustment, as I described in the other thread:
Reel to reel tape recorder adjustment
Reel to reel tape recorder adjustment
SpectraPLUS-RT can do 1/3 octave full functional until 30 days. TrueRTA also looks good, but not free if you want to go for 1/3 or 1/6 octave display. Both can generate pink noise, too. I use a CM6206-based external sound card, but I think most would do. You can do a loop test for getting a baseline.
Hi dotneck335,
You can't tell the difference between chips unless they are installed in the intended circuit first. Comparing really good op amps would probably show results well below our limit of hearing. But if there really is a difference, you would see it.
You can use a sound card, but you absolutely do need a good front end for it. Go for a 24 bit card with a high sample rate. I've used products from Soundblaster (Creative) and Emu (also Creative). Their high end stereo music card would be the one you want.
As for the RTX, it sounds fantastic! It's cost is even cheap compared to the high end D/A converter boxes. You can afford it, you just have to make it happen. My recent purchase of a Keysight 34465A meter was possible only because I sold some stuff that I didn't want to sell. You just have to make what you want happen and sometimes make a few decisions that you wouldn't normally. So now my bench has an excellent DVM and also an excellent audio processor. Money well spent. I bought the RTX in the group buy, so that was for 1/2 of the list price.
-Chris
You can't tell the difference between chips unless they are installed in the intended circuit first. Comparing really good op amps would probably show results well below our limit of hearing. But if there really is a difference, you would see it.
You can use a sound card, but you absolutely do need a good front end for it. Go for a 24 bit card with a high sample rate. I've used products from Soundblaster (Creative) and Emu (also Creative). Their high end stereo music card would be the one you want.
As for the RTX, it sounds fantastic! It's cost is even cheap compared to the high end D/A converter boxes. You can afford it, you just have to make it happen. My recent purchase of a Keysight 34465A meter was possible only because I sold some stuff that I didn't want to sell. You just have to make what you want happen and sometimes make a few decisions that you wouldn't normally. So now my bench has an excellent DVM and also an excellent audio processor. Money well spent. I bought the RTX in the group buy, so that was for 1/2 of the list price.
-Chris
Tentative OpAmp Report
I'm lazy, I won't lie.. So tonight I soldered an LME49860 to an adapter and plugged it into the tape pre-amp op-amp socket in my MX-5050MKIII-2. I did verify very briefly with my scope that it was not oscillating.
No measurements of any kind because in addition to be lazy I got home from work at 8PM tonight and I am rather tired, but... curious.
I guess it's that 30dB or so of extra OLG over the NJM2043 or perhaps a slew rate that is about 5X faster, but despite my initial pronounced skepticism I do notice the highs are a bit a cleaner sounding so perhaps upgrading at least the op-amps in the high gain positions might be worthwhile.
The sole one installed in the MKIII does run quite hot on the +/-18V supplies so I suspect in the MX-50 they will run hotter still with the +/-20V supplies. I am planning on little heat sinks for this devices. I would estimate case temperature at > 60°C so the die is probably toasty.
I am still not convinced that I should change them all, it strikes me that the most critical ones would be those that are running at significant closed loop gains.
I'm lazy, I won't lie..
So tonight I soldered an LME49860 to an adapter and plugged it into the tape pre-amp op-amp socket in my MX-5050MKIII-2. I did verify very briefly with my scope that it was not oscillating.No measurements of any kind because in addition to be lazy I got home from work at 8PM tonight and I am rather tired, but... curious.
I guess it's that 30dB or so of extra OLG over the NJM2043 or perhaps a slew rate that is about 5X faster, but despite my initial pronounced skepticism I do notice the highs are a bit a cleaner sounding so perhaps upgrading at least the op-amps in the high gain positions might be worthwhile.
The sole one installed in the MKIII does run quite hot on the +/-18V supplies so I suspect in the MX-50 they will run hotter still with the +/-20V supplies. I am planning on little heat sinks for this devices. I would estimate case temperature at > 60°C so the die is probably toasty.
I am still not convinced that I should change them all, it strikes me that the most critical ones would be those that are running at significant closed loop gains.
Attachments
Hmmmm.....interesting....thanks for reporting your findings! From the data sheets I have, it appears to me that the 49860 is ~3.5 x faster than the 2043, and ~5 x quieter. Distortion specs are impossible to compare, because New Japan Radio does NOT provide THD stats of any kind on this device that I can find. The 49860 is ~-130db and I would guess that the 2043 is quite a bit (20-30db) WORSE than this, based on other NJM products that I have found specs on. Whether or not that THD difference is audible or not, I can't say. The noise difference, if detectable at all, would ONLY be relevant in the first repro head preamp. I would think that the speed and distortion differences would be cumulative in the remainder of the circuitry; once the signal is slew-limited and/or distorted by one stage it remains so; therefore one 2043 in the path would taint the rest. But again, this is tape and these differences may not be audible at all. I am surprised at the 49860s running hot, as they cannot be drawing much current at all!I'm lazy, I won't lie..
I am still not convinced that I should change them all, it strikes me that the most critical ones would be those that are running at significant closed loop gains.
It's dissipating 140mW at idle due to the supply voltage and that package is really, really small.