I was in a hurry and had to post-and-run, but yes, they were discontinuing that and many other of the LME series a few years ago, but then they decided NOT to discontinue that one. The whole story should be in this thread:
TI to discontinue many ICs!
TI to discontinue many ICs!
I just received "Small Signal Audio Design," by Douglas Self, as a Christmas present. The author dedicates several chapters to op-amps and phono preamps, both moving magnet and moving coil. Chapter 11, on noise and distortion, shows the 5534A as the clear winner, even above the LM4562.
Of course, the best results depend as much on the circuit topology as on the op-amp selection. Since the 5532 is a dual op-amp, your circuit isn't going to provide the same performance as a circuit designed for the single op-amp 5534. Self says that it's possible to get within 2 dB of theoretical perfection by using a 5534A as the main op-amp with a 5532 implementing load-synthesis to remove the Johnson noise of the 47 kΩ load required by standard cartridges.
Note that he says Fairchild, JRC, and ON-Semi versions have 6 dB to over 9 dB lower THD at 20 kHz than other manufacturers. So, rather than replace your 5532 with a different op-amp, maybe stick with the 5532 but choose a different manufacturer of the 5532.
Personally, I'd recommend trying one of Self's schematics based on the 5534A.
Of course, the best results depend as much on the circuit topology as on the op-amp selection. Since the 5532 is a dual op-amp, your circuit isn't going to provide the same performance as a circuit designed for the single op-amp 5534. Self says that it's possible to get within 2 dB of theoretical perfection by using a 5534A as the main op-amp with a 5532 implementing load-synthesis to remove the Johnson noise of the 47 kΩ load required by standard cartridges.
Note that he says Fairchild, JRC, and ON-Semi versions have 6 dB to over 9 dB lower THD at 20 kHz than other manufacturers. So, rather than replace your 5532 with a different op-amp, maybe stick with the 5532 but choose a different manufacturer of the 5532.
Personally, I'd recommend trying one of Self's schematics based on the 5534A.
I had the opportunity to check the thd of TI ne5542 and ne5534 on AP1 and AP2 and they do show very low thd compared to most of the ic's on the market, but i actually built all the circuits based on ne5534 that he shows in his book to trick the common mode distortions and for the line levels, lm4562 subjectively was simply BETTER .I didn't try to build it's one stage mm phono amplifier based on ne5532(34) though, because i simply don't trust one stage op-amp with voltage feedback designs no matter the op-amp involved...The book is indeed a very good book, no matter what others might say, but the level of subjectivity in audio is a bit more important than it is in other sciences.
Subjectivity in audio? Can of worms there. Sample of one sighted listening is at best a statement of preference, no science at all.
Audio engineers are often in need to trick the clients self esteem and ears if they don't have "overkill designs" for cheap at hand!
IMHO any mistrust in these classic VFB designs is primarily rooted in the combination of parts that are poor output drivers of modest GBW (e.g. NJM4558) with RIAA feedback networks of very low high-frequency impedance (e.g. Rg = 180 ohm or below). A 5532 class part and a "Neumann pole" should generally eliminate such concerns.
Is there any adapter board out there that would allow replacing one dual opamp with two singles like 5534A (compensation cap included)? Sure would be handy in a case like this, assuming the feedback network is low enough in impedance for the noise advantage to be relevant. A 5534 has both lower voltage and current noise than a 5532, the A part even more so than the regular one. (Apparently the 5534A effectively used to have BC549 dies as inputs, at least back in the day. Probably has higher input capacitance than some other bipolars as a result, but I mean, it'll be used with low external impedances, so who really cares.)
I think the 5534 is very well suited for an RIAA application, and would not be one bit surprised if it had been designed specifically with this very use in mind. Its input noise on MM can be beat only by exotic low-noise FET input parts like the AD743 (or going discrete low-noise FET input stage, as a number of manufacturers eventually did in the '80s), and it has very robust output driving for the day. Input and common mode nonlinearity are of little concern in a phonopre.
Another part that I might consider is an NJM2068, but while it seems to have voltage noise in the 3 nV/sqrt(Hz) vicinity, I am not sure about current noise levels, and it's not as robust an output driver so an external buffer may be advisable, which brings up the parts count and power consumption.
Is there any adapter board out there that would allow replacing one dual opamp with two singles like 5534A (compensation cap included)? Sure would be handy in a case like this, assuming the feedback network is low enough in impedance for the noise advantage to be relevant. A 5534 has both lower voltage and current noise than a 5532, the A part even more so than the regular one. (Apparently the 5534A effectively used to have BC549 dies as inputs, at least back in the day. Probably has higher input capacitance than some other bipolars as a result, but I mean, it'll be used with low external impedances, so who really cares.)
I think the 5534 is very well suited for an RIAA application, and would not be one bit surprised if it had been designed specifically with this very use in mind. Its input noise on MM can be beat only by exotic low-noise FET input parts like the AD743 (or going discrete low-noise FET input stage, as a number of manufacturers eventually did in the '80s), and it has very robust output driving for the day. Input and common mode nonlinearity are of little concern in a phonopre.
Another part that I might consider is an NJM2068, but while it seems to have voltage noise in the 3 nV/sqrt(Hz) vicinity, I am not sure about current noise levels, and it's not as robust an output driver so an external buffer may be advisable, which brings up the parts count and power consumption.
Last edited:
The NJM2068 performs exceedingly well about 1dB quieter than the NJM5532. So not worth swapping out for and yes the 5532 has better output drive.
I modified a marantz pm62(technics su v6 modified clone) into the real technics preamp.It has a differential fet front end so noise is not a problem, i used every single op-amp i could use to drive that riaa network(lm4562, njm4562, njm2068, ne5532 , opa 2134, opa2132...) no matter what you do, the dust on 33rpm simply killed the music.It sounds really fine with clean records and also with any record on 45rpm.That is really interesting that it doesn't matter how much dust you have on 45 records , it still sounds fine, but the dust on 33 kills the music.I never had problem with dust when i used valve preamps.Probably the stylus have more kinetic energy to remove the dust and the SNR is higher anyway so any noise or transients might be masked very well at 45rpm.
I think that i built almost all the well known topology, but i didn't build the D.Self phono preamp as it is.
I think that i built almost all the well known topology, but i didn't build the D.Self phono preamp as it is.
That's a really cool idea. "Someone" should create a quick circuit in EAGLE, submit the design to a place like OSHPark, and share the layout so that folks could order the boards at $5 per square inch.
Then again, such a thing may already exist.
Could?? just back shelf your suspect./current phono amp.
Then.. Build... Salas' simple one.
More than a fair chance it will sound/perform a helluvalot better too.
Then.. Build... Salas' simple one.
More than a fair chance it will sound/perform a helluvalot better too.
That looks almost perfect. The pins appear to be a high quality option.
Unfortunately, this board has no solder pads for compensation capacitors or offset voltage nulling. I don't know how often this aspect of single op-amps is what makes them superior in performance to dual op-amps, but not having the option to try it out is rather disappointing.
If you don't mind three-dimensional circuits, you could solder a small leaded 22 pF NP0 capacitor straight to the pins of the NE5534A. (Assuming that the feedback network attenuates less than three times at high frequencies, otherwise such a capacitor will only degrade performance.)
Last edited:
According to Wikipedia the Neumann pole is complete nonsense:
RIAA equalization - Wikipedia
"In 1995 one non-academic source erroneously suggested that Neumann cutting lathes applied a single high-frequency pole at 3.18 µs (50 kHz)[8] and that a complementary zero should therefore be included upon playback. However, no such pole exists.[9]
For example, the RIAA pre-emphasis in the popular Neumann SAB 74B equaliser reaches a maximum at 100 kHz, and in addition to this, the circuit also applies a second-order roll off at 49.9 kHz, implemented by a Butterworth (maximally flat) active filter, plus an additional pole at 482 kHz.[2] This cannot be compensated for by a simple zero even if it were necessary, and in any case, other lathes will differ. Correction upon playback is not, in fact, required, as it is taken into account at the cutting stage when manual equalisation is applied while monitoring initial cuts on a standard RIAA playback system. Nevertheless, the use of the erroneous zero remains a subject of some debate among amateur enthusiasts."
I'm sure someone does adaptor boards with a location for compensation capacitors, that was just the first that came up on a search.
People simply forgot to use coupling capacitors... but anyways offset nulling is usually done with another op-amp.