• State
  • Locked Locked
  • Replies 29 replies
  • Subscribers 35 subscribers
  • Views 20913 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

LM4562 / LME49720 low frequency noise

Victor Mickevich
Prodigy 225 points
Other Parts Discussed in Thread: LM4562, LME49720, MC33078, LME49860, NE5532, OPA1642, OPA2134, OPA1612, OPA2604

I was ordered about 400pcs of these chips in SO8 package from Farnell and from Digikey within some years period. I am using these chips for ultra low distortion audio range sine oscillators manufacturing. But, I need to make noise spectrum measurement control before trying these chips, because situation is non understanding for me in frequency region lower than 100Hz. Only approximately half of all chips has noise density in this region similar as in higher frequency region, but approximately 20% of all chips has very high noise and flicker (more than ten times higher) in this low frequency region. In other ways these chips are very nice, and allow to build oscillators with distortions level about -150dB at 1kHz.

Thank  you,

Victor.

 

  • 0
    TI__Intellectual 2145 points
    Hi CC,
     
    Sosun 会统一把上次培训的文件发出来。谢谢！

    Thank you and Best Regards!
    Roman Wang/王良藩
    Signal Chain Analog FAE,DongGuan
    Mobile: 139-1665-8076
     
     
     
  • 0
    TI__Guru 52115 points
    Hi Victor,

    Which kind of help do you need from us?

    Andy
  • 0 in reply to Andy Liu (BMS)
    Prodigy 225 points

    Hi, Andy.

    Please advise, if you can, that observed situation is normal, or maybe something is wrong.

    Thank you.

  • 0 in reply to Victor Mickevich
    TI__Guru 52115 points
    Hi Victor,

    Here is what I know about these two parts:
    Noise test is implemented for these two parts but measurement time is pretty short. In the noise test, it is very hard to screen 100% noisy parts.

    Andy
  • 0 in reply to Andy Liu (BMS)
    Prodigy 225 points

    Hi,

    Here are some pictures of spectrum measured in same conditions.

    1) very good chips (around 50% of all quantity):

    2) good chips (around 30% of all quantity):

    3) Noisy chips (around 10% of all quantity). Two pictures from different chips)

    4) Very noisy chips (around 10% of all quantity). Three pictures from different chips:

    Last chip measured in wider frequency range:

  • 0 in reply to Victor Mickevich
    TI__Guru 52115 points
    Hi Victor,

    Thanks for these pictures you posted.

    The wafer process of LM4562/LME49720 is Bip process and sometimes the Bip process causes the pop-corn noise or burst noise problem. Any Bip process op-amp has the possibility of pop-corn or burst noise sometimes. It happens in wafer processing and it is very difficult to screen the devices with pop-corn noise within spec in testing process.

    Andy
  • 0 in reply to Andy Liu (BMS)
    Prodigy 225 points
    Hi,
    Thank you Andy for the advice. If I properly understand, then it is no way for to use these chips without user's input control, when excellent performance is needed in all audio range.
  • 0 in reply to Victor Mickevich
    TI__Guru 52115 points

    Hi Victor,

    In my opinion, if you really care about the "excellent performance in all audio range", I suggest you do the screening by yourself. Sorry, it is not a very good answer, but I cannot offer you a better one. 

    Andy

  • 0
    Intellectual 415 points

    I am also seeing burst noise in various production lots of the DIP LME49720 and LME49860. This was observed at the output of a 40 dB gain instrumentation amp input. The NJM5532, NJM2114, NJM2068 and MC33078 do not exhibit this burst noise in the same circuit and same PC board.


    It sometimes takes several seconds to occur so if the test time is short, parts with this defect are not screened. This LME49720 sample was a brand-new from the tube lot JR18ABE3.


  • 0 in reply to Wayne Kirkwood
    TI__Guru**** 189399 points
    Hi, Wayne, Victor,

    As Andy mentioned, it might be difficult for us to 100% fully screen out all these parts.

    However, if you would like to return these to where you purchased them, they can send them back to use for Device Analysis so perhaps we can improve the test program to better try and screen this effect out for you.

    We are always looking to continually improve.

    -d2
  • 0 in reply to Don Dapkus
    Prodigy 225 points
    Hi Don,
    Thank you for your offer.
    Maybe I can send noisy chips direct to some address of your company? Otherwise seems relatively hard and long way...

    Regards.
  • 0 in reply to Don Dapkus
    Prodigy 230 points

    Has anything been done to remedy the situation of different noise 4562s?

    D

  • 0 in reply to David D
    Prodigy 225 points
    Seems no. Last 50 pieces ordered from Farnell in July showed the similar situation - around 20% of the chips was not usable cause of the big noise.
    Victor.
  • 0 in reply to Victor Mickevich
    Intellectual 415 points

    Thanks David and Victor for bringing this back onto radar. Since the LME49710, LME49720 aka LM4562 etc. have been brought back to life on an 8" fab I have to wonder if the stock that's exhibiting the 1/f and burst noise is new or old.


    Any indication Victor when and where your 50 were made?

    I've got a circuit right now on the bench where I'm comparing the NE5532, NJM2114, OPA2134, OPA2604, OPA1642, OPA1612 and the LME49720. The circuit is not particularly critical as to whether it needs to be FET, bipolar or low Ib. Bandwidth, drive current and THD are the relevant parameters. Until I reluctantly tried the LME49720 (due to the LF noise and sourcing issues) the OPA1612 performed best. Glad I revisited the LME49720 because it outperforms all of them on distortion. The OPA1612 is a very close second and if it were available in through hole it would be my first choice because its very, very quiet with a smooth-sounding, not rocky, noise floor.

  • 0 in reply to Wayne Kirkwood
    Prodigy 225 points
    Unfortunately I do not know where and when last ordered chips were made. Chips in the SO8 package have the manufacturer's code "63L3" in the first line.
    Regards.
  • 0 in reply to Victor Mickevich
    Prodigy 230 points

    Have you by chance tried this with the OPA1612?

  • 0 in reply to David D
    Intellectual 415 points

    I've recently used the OPA1612 in both a Moving Coil preamp and a Mic Preamp (both using external transistors on the front end) and they were very, very quiet. (The MC preamp used ZTX851 transistors the Mic preamp THAT300s.)

    Also compared the OPA1612 to an NJM2068 and OPA2134 in a moving magnet preamp and it was, as expected due to higher current noise, noisier than the BJT '2068 and FET '2134. The outcome was no great surprise but the noise character of the OPA1612, which was exaggerated due to the current noise and inductive source impedance, was still clean-sounding noise. In the same circuit the LME49720 (and the LME49860 RIP) were unusable due to burst noise.

    The OPA1612 does not seem to have the process contamination of the LM497X0 that produces rocky-sounding noise.

    I have not tested the OPA1612 with high value bias resistors to see if it has the latch-up/bias current correction start-up issues of the LME49720.

    In terms of noise performance in the real world the OPA1612 beats the LME49720.

  • 0 in reply to David D
    Prodigy 225 points

    As I know, OPA1612 isn't better for my application. Also, according the data sheet, OPA1612 may has higher input bias current, and has no data for to drive 600 ohm load. And finally, OPA1612 is more expensive.

  • 0 in reply to Wayne Kirkwood
    Prodigy 225 points
    If maybe you are interested, I have some spectrum measurement and pulse response data for popular opamps in the one chip MM phono preamp implementation. You can see links in this thread (post #209) : www.diyaudio.com/.../293756-suggestion-opamps-21.html
  • 0 in reply to Victor Mickevich
    Intellectual 415 points

    Thank you for the link Victor.

    "[OPA1612] has no data for to drive 600 ohm load."

    The 600 Ohm THD load data is featured on page 1 of the datasheet and also characterized on page 8.

    The OPA1612 does have a higher maximum Ib. The advantage of the OPA1612 in my experience is that you don't have to throw 20% or more of them away due to noise. So they may not be as expensive as they appear to be.


    100% of my limited supply of LME49860 and LME49720 over several different dates codes have burst noise in high gain circuits. I would not enjoy screening SMT parts.

  • 0 in reply to Wayne Kirkwood
    Prodigy 225 points
    Thank you. Missed these pictures in OPA1612 datasheet, saw only tables.
    OPA1612 price is for 2...2,5 EUR higher, but the final result for me probably would not be better.
  • 0 in reply to Don Dapkus
    Prodigy 225 points
    TI, what's going on? The previous order in February this year - 50% usable chips. The latest fresh order in May - one usable chip from the five!!! The absolutely catastrophic situation!
  • 0 in reply to Victor Mickevich
    TI__Expert 6985 points
    Hello Victor! I am sorry for your frustration. Can you please share more information on what you are experiencing as well as which device you are referencing in this thread of messages? Thanks, Jeff
  • 0 in reply to Jeff Kohnle
    Prodigy 225 points
    The problem is still with the LM4562/LME49720 noise, or I do not know how it can be called - last LM4562 chips from the Mouser have chaotically output jumps from one level to another with the period of some tenth parts of second and more. The picture is similar as in Wayne Kirkwood's post in this thread. When I started this thread, there was the 20% of faulty chips. Now I have 80% in trash.
  • 0 in reply to Jeff Kohnle
    Prodigy 225 points

    Here is the fresh picture of LM4562 output. TI, what I can do with that chips?

  • 0 in reply to Victor Mickevich
    Intellectual 415 points

    That is exactly the type of burst noise I saw in earlier (6") production runs of the LME49720.

    I bought a few from Mouser some time ago that were clean but I don't have any recent lots to test.

    TI really needs to get some of your parts back to do failure analysis. I would return them to the distributor for credit, replacement or reimbursement. Those parts will go back to TI and they can start looking at this.

  • 0 in reply to Wayne Kirkwood
    Prodigy 225 points
    Today I checked 20 pcs else from my latest order. Overall situation is little better - 13 from the 25 chips have the "burst" noise and are unusable. Similar situation was observed with the previously ordered chips on February this year from the Farnell - around 50% bads.
  • 0 in reply to Wayne Kirkwood
    Prodigy 225 points
    I'll be glad to send the noisy parts back to TI, but it is not so easy for me. Our local re-seller says that the Mouser don't want to get parts back. Usually the Mouser says: "... we do not accept returns from resellers..."
    Easiest way would be, if I can send the noisy chips direct to TI.
  • 0 in reply to Victor Mickevich
    Prodigy 225 points

    Finally the Mouser accepted the return. Thanks Mouser for this.

    Last week received 100pcs of the LME49720 ordered from TI store. Country of origin marked as Philippines. After the test I have 23 noisy parts. Asked TI store for the return. Waiting for the answer.