Is there any audible difference between the ON Semi version of the NE5532 (in SOIC-8) and the Texas Instruments version? I seem to have read somewhere that the T.I. version had higher distortion at high frequencies.
Indeed, the best were the discontinued Signetics parts, but the difference is beyond audibility I believe.
JRC were best, in comparison to ON and TI.
Now Nissinbo or something, company got merged.
For really low noise, AD797 is a candidate.
Now Nissinbo or something, company got merged.
For really low noise, AD797 is a candidate.
I believe that New Jap Radio's datasheets are misleading. They don't quote specs in the same manner as most other companies; thus I don't trust 'em.
Doug Self measured slightly higher noise in the TI parts when writing his book - Small Signal Audio Design. TI has been recently switching over to newer fabs so that information may or may not be valid.
It's many years since I did the test, but one of the first things I wanted to do when I got my UPV was compare the Ti against Signetics 5532. I was cynical and suspected it was an Internet rumour. I was wrong. The Signetics part had better THD+N combined (it was a very small difference). There is also the NJM5532DD. I have some Philips-branded 5532s somewhere (same as Signetics?).
What's a "UPV"? I believe that Signetics acquired Philips at some point. I did find some Signetics NE5532s a few years back, but they're seemingly impossible to find now; also they never made them in SOIC-8 format. which is what I'm needing now.
Doug Self did an amp for Elektor with a zillion 5532s as the output stage. As part of that project, he tested many samples of all the 5532s available and found small but significant differences in THD & noise. He reported his results in another Elektor article IIRC.
The best are, without doubt, the Mullard/Philips/ (Signetics?) ones made by Southampton Virgins in the great Mullard factory. They are like valves from Mullard, Blackburn; at one time, the biggest valve factory in the World. Alas, both factories are long gone.
Circa 1980, there was a world shortage of NE5534ANs; the Low Noise version of 5534. At Calrec, we set up a jig to select 5534s for Mike amps & Virtual Earth mixers; the most noise critical applications. EVERY SINGLE MULLARD/PHILIPS 5534 met the higher 5534AN standard. I must have tested more than a 100 myself. I modded the jig to test 5532s and found the same but we didn't test as many as 5534s.
A decade or so later, I came back to commercial electronic design and was very disappointed to find this happy state was no longer extant
I think it was mainly TI bits from Thailand that I was looking at but no maker had the reliable, consistently good performance of the Mullard/Philips devices. If your 5532s are marked Mullard, they will be the stuff hand carved by Southampton virgins
IIRC, there are only 4 masks for 5532/4 in existence. You can clearly see the 4 different types in Self's measurements, including some NOS Philips devices. If anyone has a scan of Self's article, please post it here.
I'm not sure NJM5532DD are 'real' 5532s. NJM/JRC make a huge range of OPAs, some very good, but based on a completely different topology. The big difference is the usual NJM OPA has PNP inputs while a 'real' 5532 is NPN input. But NJM datasheets usually show the topology.
The best are, without doubt, the Mullard/Philips/ (Signetics?) ones made by Southampton Virgins in the great Mullard factory. They are like valves from Mullard, Blackburn; at one time, the biggest valve factory in the World. Alas, both factories are long gone.
Circa 1980, there was a world shortage of NE5534ANs; the Low Noise version of 5534. At Calrec, we set up a jig to select 5534s for Mike amps & Virtual Earth mixers; the most noise critical applications. EVERY SINGLE MULLARD/PHILIPS 5534 met the higher 5534AN standard. I must have tested more than a 100 myself. I modded the jig to test 5532s and found the same but we didn't test as many as 5534s.
A decade or so later, I came back to commercial electronic design and was very disappointed to find this happy state was no longer extant
I think it was mainly TI bits from Thailand that I was looking at but no maker had the reliable, consistently good performance of the Mullard/Philips devices. If your 5532s are marked Mullard, they will be the stuff hand carved by Southampton virginsIIRC, there are only 4 masks for 5532/4 in existence. You can clearly see the 4 different types in Self's measurements, including some NOS Philips devices. If anyone has a scan of Self's article, please post it here.
I'm not sure NJM5532DD are 'real' 5532s. NJM/JRC make a huge range of OPAs, some very good, but based on a completely different topology. The big difference is the usual NJM OPA has PNP inputs while a 'real' 5532 is NPN input. But NJM datasheets usually show the topology.
dotneck: what gives you the ability to say NJM is good or bad?
You have experience, access to test reports and so on?
In general, the deviation from spec (6 Sigma), a statistical measurement, is very much lower for all Japanese electronic parts, their transistors and capacitors are famous for that.
NJM / JRC was well known for their quality.
Also, how many do you need?
They are about 50 cents here, if in stock.
The TI made ones go for half that.
It is amusing to me that people looking for a single tube make assertions about businesses making millions of devices daily.
They did not get big without good quality, and keeping the customers happy in many ways.
You have experience, access to test reports and so on?
In general, the deviation from spec (6 Sigma), a statistical measurement, is very much lower for all Japanese electronic parts, their transistors and capacitors are famous for that.
NJM / JRC was well known for their quality.
Also, how many do you need?
They are about 50 cents here, if in stock.
The TI made ones go for half that.
It is amusing to me that people looking for a single tube make assertions about businesses making millions of devices daily.
They did not get big without good quality, and keeping the customers happy in many ways.
Really intriguing point re: NJM5532DDs. Thank you for that. I have become a relatively recent convert to them, having seen one or two Ti ones fail without obvious cause. I didn't think of investigating the topology and went on the basis that - in the same way Ti's modern opamps appear to have wafer contamination, creating burst / popcorn noise - that they were less likely to have the same issue. I use them regularly in a VF configuration, where previously the LM4562 was used. The NJM part gives the same or better SNR, but I don't have to worry about popcorn (and it also runs at higher V, so I don't have to drop as much heat in the PSU). As a non-inverting VF, the NJM part performs as well as any IC I've tried (but it should do as a VF!).
A UPV is a recently-discontinued analyser. My best one gets within a dB or so of the AP 2722 (used by Mr Self) in terms of THD+N. It is a workhorse that I use every day, unlike the 2722 that can live for years without seeing action.
A UPV is a recently-discontinued analyser. My best one gets within a dB or so of the AP 2722 (used by Mr Self) in terms of THD+N. It is a workhorse that I use every day, unlike the 2722 that can live for years without seeing action.