Users have complained about popcorn noise in the LME49720 and LM4562 many times, which, as far as I know, are essentially the same part. TI has said repeatedly that they will look into it but have given no official response. It is a shame, these chips have great specs and were supposed to be a replacement for the 5532 in terms of similar price and superior performance. It seems that the 5532 from the 1970’s is a formidable opponent.
Ah, the reply I was hoping for. With this many noisy parts, I figured this had to be a known issue. My parts have been in my stash for several years. I put the lot numbers from the chips in the description but have not deciphered them yet.
@@JohnAudioTech that explains a lot. You probably have parts from before they corrected the issue. I heard about popcorn noise in these parts ages ago. So you probably have really old parts. If your wallet allows for it, you could buy a few of the new ones and make a follow up video. Great work!
@@analog_joeSearching around, I've found lots of complaints about the popcorn noise issue (Texas Instruments E2E forums), but nothing definitive on TI actually correcting the issue. Can you provide info on when TI actually took action to correct the issue?
@@JohnAudioTech I remember reading it on the TI forum as well. Although I might’ve been looking for the LME49720 which is basically a rebranded LM4562. However, you are right, I made a couple of searches and found nothing definitive from TI. I might’ve imagined it, but I am almost positive that they did mention that they changed their semiconductor process for that part. All the posts I’ve seen are also very old, 6 years or more. But be it as it may, many users complain about several National part numbers having this issue, not just the 4562 or the 49720. Maybe it is best to avoid them and stick to Burr-Brown parts
Indeed TI did change its fabrication process, however, I am unaware if it made things better or worse. Be it as it may, popcorn noise is not the only issue of this part. It appears that there is also an EMI problem Best to avoid it. I edited my original post.
About 8 years ago, I designed and built my phono amp using LM4562s. I firstly bought from well known auction sites (found them to be useless fakes) then from Farnell - the best. LM4562 is a highly strung opamp, they need an essential cap from +ve to -ve, I soldered directly to the IC's pins, this stopped spurious HF oscillations. IMHO, with the design you have C2 and C3 being electrolytics, I would never use - they can be very noisy, I would use film, but at 22uF is the reason why I would not use this design. All the best, Beamer
Freezer spray is something I found essential for my workshop. It's especially useful in complicated circuits with many chips. Any change in the noise as you freeze a chip instantly tells you unambiguously where the problem is. I've had quite a few chips that work like this, often with intermittent problems which are hard to trace.
Supposed to be ultra low noise, but they are either faulty from manufacture, , remarked or fakes. As they are now considered obsolete (not recommended for new use) you likely are getting old wafers that have been in storage for a while being packaged, and thus they might have been contaminated and not washed properly at some point, leading to this higher noise from surface contamination. Unless you get those that are JAN-TX rated, you will probably have this. Time to look for an alternate, of which there are plenty around, though often not with the high power voltage ratings though. these though should be free of popcorning, I would complain to the supplier if you still have all the order details.
If you're using any thick film resistors they can contribute significant noise. Electrolytic caps can be noisy also, but generally it isn't noticed if they are on the power rails for the op amp.
Pop and hisses can indicate bad caps and choke coils. Could also indicate bad resistance per power inputs across a chip. LM384 if you don't add a resistor between GND and VCC You'll experience pops and hisses.
maybe you can answer my question no one has been able to answer me? I have different preamplifier boards (pcb) from ebay that are designed differently and not the same. Some have coupling capacitors, some don't. I have several different ones op-amp ic. models, brands. I try the ones on (naim nap250) clone card (pcb) from ebay. Treble, bright frequencies are always destroyed, less well, precisely as the frequency rolls off too early. If I connect this amplifier directly to my DAC. then the problem is gone. I have never been able to figure out why. Some preamplifier cards, pcb. have different potentiometers and some have input relay control and some only a manual switch. have also tried different rca cables. Do you have any idea what the problem is
I once got a bad LM3886TF with bad hiss and popping noise. It was sample from TI(National semiconductor)! They gave 5 of them but 4 were fine. Only one was noisy. At first I thought I had bad capacitor somewhere.
Did you test with open input? What about testing with a real pick up input (MM or MC?) - do you see the same ? None the less, LM4562 is prone to have popcorn noise at very hig input resistance (manufacturing issue) - but is it relevant in your application?
I can hear it with the MM cartridge connected or open. I can't hear it when music is playing, but can hear it above surface noise between tracks. Very relevant to a part marketed as a high quality audio component.
@@JohnAudioTech Then it's bad. I use to buy a lot of LM4562 and then screen them for pop-noise. I supposed you did but then from Mouser, Digikey or similar (not from eBay seller in China)
Does anyone know if TI manufactures National Semiconductor branded LME49720's or are they selling off the old stock manufactured by Nat Semi? I know the 4562's are TI because there is no NS logo on them. They supposedly contain the same circuit but it's a patented circuit that's why we don't get the "Equivalent Circuit" diagram on either data sheet. It's a pretty inventive input bias cancellation circuit but it requires about 4x more silicon than your average audio op amp; an extra npn diff pair in series with the collectors of the usual pnp input pair, a current mirror consisting of at least 3 transistors, a bias duplicator transistor that feeds the current mirror, and then at least 5 diodes and/or Vbe drops to bias everything properly to achieve negligible input bias current which of course should result in very low noise and exceptional linearity. Ideal matching of that silicon would result in a noise figure of 0dB. Of course there is no ideal silicon but in this case there are more opportunities for small inconsistencies to become big noise- exactly what we heard here.
I don't understand the comments suggesting adding various passive components to eliminate the noise. If adding stuff is often your first inclination, do yourself a favor and read this comment. First, it's usually wise to assume that a published schematic is complete until otherwise proven. With that in mind, we should rate our possible solutions in order of relevance of course, but also in order of ease and simplicity. Since adding components betrays the schematic and requires disassembly, de-soldering and soldering, it rates very low in both categories. In fact, that would probably be the very last thing we should try right? Instead, consider starting with the existing components that are in series with the noise like the op amp was in this case. Relevance of op amp to noise: It was the series component preceding the noise so relevance = very high Ease/simplicity of replacing op amp: Socketed so ease = very high Solution #1: Replace op amp. Noise eliminated. Repair complete. The creator solved the problem in one simple, solderless step. Imagine if he had added more components which could have possibly created more problems. I hope this comment helps those who have been solving problems backwards.
@@AnalogDude_ They're not necessary but they are helpful. Especially used in high power audio amplifiers where heat and age will begin to increase the drift of the opamp circuit. Check any service manuals for old school stereo systems, and you will find these capacitors used in every single one of them. These capacitors are absolutely required in BTL applications where floating grounds are used.
You could try the OPA2189 precision ultra low noise op-amps. But I would still replace the caps and make sure all the resistors are 1% metal film (normal standard these days, so cheap).
Or even corrosion within the package? Intel has had issus with corrosion intrusion in some CPUs recently, I'm sure a common op-amp gets a lot less TLC.
Knockoffs!!! I drive by Texas Instruments daily install sockets but Discrete for $20 production op-amps I've had to swap in all types of equipment since the 80s
Not surprising John. Possibly even Digi-key and other sources are receiving knock off chips from other countries. QA is not superb, as reflected by their low cost.
It would be nice to use an oscilloscope to look at what is going on. I would also return the bad chips to DigiKey or Mouser, or at least get a future credit. I know it's not a lot of money, but you don't want to reward companies that are sending you junk...
BTW I would replace all the caps and make sure NONE of the resistors are not carbon (unlikely), but you might want to replace them with known metal film types. I suggest this becasue the noise is still FAR TOO HIGH, the hiss should be nearly non existent (with the op-amps specs), not the large crap you have.
@@stevenbliss989 Resistors all look good from what I can see - even some pricey Dale parts on there. Electrolytic caps look like Panasonic from the markings and the film caps are mostly Wima from the red package.
Users have complained about popcorn noise in the LME49720 and LM4562 many times, which, as far as I know, are essentially the same part. TI has said repeatedly that they will look into it but have given no official response. It is a shame, these chips have great specs and were supposed to be a replacement for the 5532 in terms of similar price and superior performance. It seems that the 5532 from the 1970’s is a formidable opponent.
Ah, the reply I was hoping for. With this many noisy parts, I figured this had to be a known issue. My parts have been in my stash for several years. I put the lot numbers from the chips in the description but have not deciphered them yet.
@@JohnAudioTech that explains a lot. You probably have parts from before they corrected the issue. I heard about popcorn noise in these parts ages ago. So you probably have really old parts. If your wallet allows for it, you could buy a few of the new ones and make a follow up video. Great work!
@@analog_joeSearching around, I've found lots of complaints about the popcorn noise issue (Texas Instruments E2E forums), but nothing definitive on TI actually correcting the issue. Can you provide info on when TI actually took action to correct the issue?
@@JohnAudioTech I remember reading it on the TI forum as well. Although I might’ve been looking for the LME49720 which is basically a rebranded LM4562. However, you are right, I made a couple of searches and found nothing definitive from TI. I might’ve imagined it, but I am almost positive that they did mention that they changed their semiconductor process for that part. All the posts I’ve seen are also very old, 6 years or more. But be it as it may, many users complain about several National part numbers having this issue, not just the 4562 or the 49720. Maybe it is best to avoid them and stick to Burr-Brown parts
Indeed TI did change its fabrication process, however, I am unaware if it made things better or worse. Be it as it may, popcorn noise is not the only issue of this part. It appears that there is also an EMI problem
Best to avoid it. I edited my original post.
About 8 years ago, I designed and built my phono amp using LM4562s. I firstly bought from well known auction sites (found them to be useless fakes) then from Farnell - the best. LM4562 is a highly strung opamp, they need an essential cap from +ve to -ve, I soldered directly to the IC's pins, this stopped spurious HF oscillations. IMHO, with the design you have C2 and C3 being electrolytics, I would never use - they can be very noisy, I would use film, but at 22uF is the reason why I would not use this design. All the best, Beamer
Freezer spray is something I found essential for my workshop. It's especially useful in complicated circuits with many chips. Any change in the noise as you freeze a chip instantly tells you unambiguously where the problem is. I've had quite a few chips that work like this, often with intermittent problems which are hard to trace.
thank you this video a motivation to finally finish this project :D just needs an enclosure for like a year
Supposed to be ultra low noise, but they are either faulty from manufacture, , remarked or fakes. As they are now considered obsolete (not recommended for new use) you likely are getting old wafers that have been in storage for a while being packaged, and thus they might have been contaminated and not washed properly at some point, leading to this higher noise from surface contamination. Unless you get those that are JAN-TX rated, you will probably have this. Time to look for an alternate, of which there are plenty around, though often not with the high power voltage ratings though. these though should be free of popcorning, I would complain to the supplier if you still have all the order details.
LM4562 is still an active product and it will be for many decades to come just like the 5532
If you're using any thick film resistors they can contribute significant noise. Electrolytic caps can be noisy also, but generally it isn't noticed if they are on the power rails for the op amp.
Super LM. ❤
Pop and hisses can indicate bad caps and choke coils. Could also indicate bad resistance per power inputs across a chip. LM384 if you don't add a resistor between GND and VCC You'll experience pops and hisses.
maybe you can answer my question no one has been able to answer me?
I have different preamplifier boards (pcb) from ebay that are designed differently and not the same. Some have coupling capacitors, some don't. I have several different ones
op-amp ic. models, brands. I try the ones on (naim nap250) clone card (pcb) from ebay. Treble, bright frequencies are always destroyed, less well, precisely as the frequency rolls off too early.
If I connect this amplifier directly to my DAC. then the problem is gone. I have never been able to figure out why.
Some preamplifier cards, pcb. have different potentiometers and some have input relay control and some only a manual switch. have also tried different rca cables.
Do you have any idea what the problem is
I once got a bad LM3886TF with bad hiss and popping noise. It was sample from TI(National semiconductor)! They gave 5 of them but 4 were fine. Only one was noisy. At first I thought I had bad capacitor somewhere.
Did you test with open input?
What about testing with a real pick up input (MM or MC?) - do you see the same ?
None the less, LM4562 is prone to have popcorn noise at very hig input resistance (manufacturing issue) - but is it relevant in your application?
I can hear it with the MM cartridge connected or open. I can't hear it when music is playing, but can hear it above surface noise between tracks. Very relevant to a part marketed as a high quality audio component.
@@JohnAudioTech Then it's bad. I use to buy a lot of LM4562 and then screen them for pop-noise.
I supposed you did but then from Mouser, Digikey or similar (not from eBay seller in China)
How you tried turning it off then on again?
Try them a 2 layer PCB with dedicated ground just to be sure.
Does anyone know if TI manufactures National Semiconductor branded LME49720's or are they selling off the old stock manufactured by Nat Semi? I know the 4562's are TI because there is no NS logo on them. They supposedly contain the same circuit but it's a patented circuit that's why we don't get the "Equivalent Circuit" diagram on either data sheet. It's a pretty inventive input bias cancellation circuit but it requires about 4x more silicon than your average audio op amp; an extra npn diff pair in series with the collectors of the usual pnp input pair, a current mirror consisting of at least 3 transistors, a bias duplicator transistor that feeds the current mirror, and then at least 5 diodes and/or Vbe drops to bias everything properly to achieve negligible input bias current which of course should result in very low noise and exceptional linearity. Ideal matching of that silicon would result in a noise figure of 0dB. Of course there is no ideal silicon but in this case there are more opportunities for small inconsistencies to become big noise- exactly what we heard here.
I don't understand the comments suggesting adding various passive components to eliminate the noise. If adding stuff is often your first inclination, do yourself a favor and read this comment. First, it's usually wise to assume that a published schematic is complete until otherwise proven. With that in mind, we should rate our possible solutions in order of relevance of course, but also in order of ease and simplicity. Since adding components betrays the schematic and requires disassembly, de-soldering and soldering, it rates very low in both categories. In fact, that would probably be the very last thing we should try right? Instead, consider starting with the existing components that are in series with the noise like the op amp was in this case.
Relevance of op amp to noise: It was the series component preceding the noise so relevance = very high
Ease/simplicity of replacing op amp: Socketed so ease = very high
Solution #1: Replace op amp.
Noise eliminated.
Repair complete.
The creator solved the problem in one simple, solderless step. Imagine if he had added more components which could have possibly created more problems. I hope this comment helps those who have been solving problems backwards.
2:59 why are the C2 and C3 capacitors?
I saw another circuit also doing a similar thing.
They remove output DC offset.
@@yellowest_ball to ground?
@@AnalogDude_
To the reference voltage (the non-inverting input of the opamp), which in this case is indeed ground via R1L.
@@yellowest_ball Ya, i have used non inverting opamps, but not used caps to ground it, i saw no schematics using it.
@@AnalogDude_
They're not necessary but they are helpful. Especially used in high power audio amplifiers where heat and age will begin to increase the drift of the opamp circuit. Check any service manuals for old school stereo systems, and you will find these capacitors used in every single one of them.
These capacitors are absolutely required in BTL applications where floating grounds are used.
That's a Rod Elliot pre amp. Seems to be excessively noisy for some reason?
You could try the OPA2189 precision ultra low noise op-amps. But I would still replace the caps and make sure all the resistors are 1% metal film (normal standard these days, so cheap).
Is this due to doping corrosion on the actual silicon, due to defects in manufacture? I have heard of this being an eventual issue for some op amps.
Or even corrosion within the package? Intel has had issus with corrosion intrusion in some CPUs recently, I'm sure a common op-amp gets a lot less TLC.
Knockoffs!!! I drive by Texas Instruments daily install sockets but Discrete for $20 production op-amps I've had to swap in all types of equipment since the 80s
Should have used transistors!
thanks
Not surprising John. Possibly even Digi-key and other sources are receiving knock off chips from other countries. QA is not superb, as reflected by their low cost.
It would be nice to use an oscilloscope to look at what is going on. I would also return the bad chips to DigiKey or Mouser, or at least get a future credit. I know it's not a lot of money, but you don't want to reward companies that are sending you junk...
from our tests defective ratio is about 25% horrible
Horrible , and from supposedly official sourced retail. How come these haven't been recalled?
JRC\NJM or TI, anything else is junk.
Quality downfall. We can't escape it.
IF AMERICANS STOP BUYING CHINESE WE CAN...
That is still a LOT of noise for op-amps that are supposed to be VERY LOW noise, ....I am suspicious!!!!!!!!!!!!!!!!!
BTW I would replace all the caps and make sure NONE of the resistors are not carbon (unlikely), but you might want to replace them with known metal film types.
I suggest this becasue the noise is still FAR TOO HIGH, the hiss should be nearly non existent (with the op-amps specs), not the large crap you have.
@@stevenbliss989 Resistors all look good from what I can see - even some pricey Dale parts on there. Electrolytic caps look like Panasonic from the markings and the film caps are mostly Wima from the red package.