The LM4562's EMI seems to be weak against Wi-Fi radio waves (2.4GHz).
I made a headphone amplifier with the configuration of LM4562-LME49600 more than 10 years ago. When I connected to the i-pod and accessed the Internet via wi-fi, I heard a burst noise called “BoBoBo…”.
Of course LME49720, 49860, 49710 are the same.
I made a headphone amplifier with the configuration of LM4562-LME49600 more than 10 years ago. When I connected to the i-pod and accessed the Internet via wi-fi, I heard a burst noise called “BoBoBo…”.
Of course LME49720, 49860, 49710 are the same.
Although there is a possibility, it will not occur if OP-AMP is replaced with another type (NE5532, OPA1612 etc) with the same board.
If is the design from LME49600 datasheet I made 2 for two colleagues around 10 years ago and were no problem with noise from mobile phone. You can see my PCB on a different forum.
Last edited:
The OPA2189 is a precision opamp that is probably chopper stabilized. Supply goes up to 36V btw.
I used the OPA189 in a circuit once and the circuit had some problems with noise in the upper part of the audio band. I didn't proof that the chopper frequency of the OPA189 was the source of the noise but it is what I assumed after some time.
I made some measurements with the OPA1641 together with the BUF634 in an headphone amplifier and the circuit produced a garden fence like distortion whîle K2 and K3 with the LT1028, OPA1655 and the OPA1611 were 'dominant' and higher harmonics were barely visible. Maybe the OPA1641 was just too slow.
The ADA4898-2 has a thermal pad and therefore is not really compatible to the LM4562.
The OPA1622 has a totally different package.
Don't know the other opamps yet.
ADA4627-1 is the best op amp I've heard thus far and what I'm using in my Echo Gina 24/96 and TC Impact Twin. However, note that it is only rated for +/-15V supplies, would require two SOICs mounted on an adapter to replace an LM4562, and is somewhat pricey. Have tried the ADA4627-1 in a variety of circuits and would describe it as having nice detail in the highs, deeper in the lows, and overall more natural than the LM4562 family, which I find a tad thin and sterile. Local supply bypassing across the rails with something like 0.1µF may be advisable.
https://www.analog.com/media/en/technical-documentation/data-sheets/ADA4627-1_4637-1.pdf
https://www.analog.com/media/en/technical-documentation/data-sheets/ADA4627-1_4637-1.pdf
I've replaced the OP275 and OPA2227 on one of my Lynx L22 sound cards by bunch of LM4562.
That was a stupid move: Lots of popcorn noise. Next time I will try the OPA1612.
Read more.
Cheers,
E.
That was a stupid move: Lots of popcorn noise. Next time I will try the OPA1612.
Read more.
Cheers,
E.
In addition to the fact that the LM4562 is vulnerable to EMI, there are also many reports that popcorn noise and burst noise occur with a relatively high probability.
All the LM4562s I've dealt with have confirmed that they are sensitive to noise from RF demodulation, but should I say fortunately, I have yet to experience Popcorn noise or burst noise that occurs independently without being demodulated by RF.
The EMI environment of the sound cards in the PC should be pretty bad, but will that your LM4562 individual with a lot of popcorn noise produce popcorn noise at the same in frequency and level even in a clean EMI environment?
I just want to know if it's from RF demodulation or just noise on its own.
All the LM4562s I've dealt with have confirmed that they are sensitive to noise from RF demodulation, but should I say fortunately, I have yet to experience Popcorn noise or burst noise that occurs independently without being demodulated by RF.
The EMI environment of the sound cards in the PC should be pretty bad, but will that your LM4562 individual with a lot of popcorn noise produce popcorn noise at the same in frequency and level even in a clean EMI environment?
I just want to know if it's from RF demodulation or just noise on its own.
I've seen LM4562's exhibit ultra low frequency popcorn noise myself. Amusingly it manifested as unexpectedly wild voltage swings at the output of a DC servo. Further investigation found that the servo itself was just fine, and the problem child was the "forward path" which was an LM4562. Swapping in a half dozen brand new, virgin fresh, LM4562 chips revealed that the problem appeared and disappeared when I installed and replaced the original, bad, 4562.
Edmond Stuart, Mark Johnson
Thank you both for your reply.
After all, it means that there is also a problem of noise alone.
I still have dozens of unopened LM4562, LME49860, etc., so I would like to continue to carefully observe this noise.
Thank you both for your reply.
After all, it means that there is also a problem of noise alone.
I still have dozens of unopened LM4562, LME49860, etc., so I would like to continue to carefully observe this noise.
Last edited:
After a couple of weeks I'm very happy with the LM6172. Lovely sound, the very realistic piano tone is particularly pleasing, but it's generally very neutral. I like it.
For some reason I tried a whole bunch of Dil 8 opamps in a test cirquit. After I had tested a good handfull, some started to show strange results, which really confused me. I repeated the test with some of the prior tested ones and got partial strange results too.
I had done only maybe 50-70 insertions, but the socket was already so worn out that it's contacts got unreliable. A socket with these round "precission" contacts.
A nice way to ruin an opamp test that was meant to be objective.
Now I use the cheap version for testing.
I had done only maybe 50-70 insertions, but the socket was already so worn out that it's contacts got unreliable. A socket with these round "precission" contacts.
A nice way to ruin an opamp test that was meant to be objective.
Now I use the cheap version for testing.
LM6172 is great as I/V, its only significant drawback is higher noise which I take to be responsible for the lack of air. AD811 would likely be lower noise but probably needs circuit modification as its CFB.
If you leave a bunch of 4562/49720s to soak overnight, even after you've vetted them for noise, you invariably find that one or two will have developed the familiar 'trampoline' effect on the FFT when you get into the workshop the next day. Most of the time, the burst / popcorn issue arises within a couple of minutes of being powered for the first time, but it can start hours down the line, which is immensely frustrating.
There is a posting here somewhere (from a TI employee?) that claims that they are now making them on a new wafer fab line, so the burst noise issue ought to have gone...
One thing is for certain: if you can weed out the noisy OAs (anywhere between 30-50% - it's a total PITA), the 4562 / 49720 is a fantastic opamp. It really needs to see a source Z that's low, and even becomes unstable when fed with medium Z. As to why Natsemi's team put a phono preamp in the apps data is beyond me... It's just about the worst OA for a MM input that you could find and Self showed it to be noiser than a TL072 in this app...
My feeling regarding the EMI / RFI reports is that either the layout / implementation is creating ultrasonic instability, or the PCB isn't sufficiently shielded. I've left 4562 / 49720 circuits on long-term soak and watched the FFT. There are no problems if your topology, PCB and enclosure are properly designed.
I don't know where you get this fear of popcorn with this operation noise, but I have used it in several applications, including my DAC where it is used in the filter and it does not have a bit of noise. In the DAC, I've been using it daily for over 10 years now and the noise is so low that in combination with an M60 you can't hear anything at full volume, even with your ear close to the speakers. And regarding EMC, I did a test in which I put a mobile phone on the DAC box and called and absolutely nothing was heard in the speakers even though the volume was at maximum.
I think the biggest issue is implementation, which leaves a lot to be desired in most DIY projects.
Just a few examples: there is a lot of noise about jitter and all kinds of schemes are designed and used to reduce jitter, but they are made on separate boards connected with simple cables, some even over 10 cm long, and I don't have seen in any image posted from such assemblies (DACs in general) some measure of impedance matching is taken for the very fast signals used.
Very fast signal does not necessarily mean high frequency, the slew rate value determines the speed of the signal and all established jitter reduction schemes have a very high slew rate to make the rectangular signal look good, but to preserve its quality, all fast signal lines must work on adapted impedances and the connection cables between the boards only adapted signal lines are not. Even when they are made on the same board, the connection lines between the integrated ones are not adapted.
And even if you happen to find one with problems, all you have to do is change it and use the one with problems in non-audio modest schemes.
I think the biggest issue is implementation, which leaves a lot to be desired in most DIY projects.
Just a few examples: there is a lot of noise about jitter and all kinds of schemes are designed and used to reduce jitter, but they are made on separate boards connected with simple cables, some even over 10 cm long, and I don't have seen in any image posted from such assemblies (DACs in general) some measure of impedance matching is taken for the very fast signals used.
Very fast signal does not necessarily mean high frequency, the slew rate value determines the speed of the signal and all established jitter reduction schemes have a very high slew rate to make the rectangular signal look good, but to preserve its quality, all fast signal lines must work on adapted impedances and the connection cables between the boards only adapted signal lines are not. Even when they are made on the same board, the connection lines between the integrated ones are not adapted.
And even if you happen to find one with problems, all you have to do is change it and use the one with problems in non-audio modest schemes.