• State Resolved
  • Locked Locked
  • Replies 9 replies
  • Answers 1 answer
  • Subscribers 48 subscribers
  • Views 564 views
  • Users 0 members are here
Support feedback
Options
Options
Related

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.

Original question:

TL074: Voltage follower running very hot

TL074: TL074 running hot as a voltage follower

Murray Hodge
Prodigy 20 points

Part Number: TL074
Other Parts Discussed in Thread: LM339, TL072, LF347

Tool/software:

I have TL074 op amps running as sample and hold voltage followers with the +ve input going to the output of a 4051 switch with a 10nF cap to hold the voltage between refreshes.

The TL074 is running very hot and it is not a load problem. If I remove the LM339 which is the only load the TL074 voltage followers this does not alter the heat of the 074.

The TL074 that is being used as the level shifter for the saw wave is only getting slightly warm.

Regards

  • 0
    TI__Mastermind 27868 points

    Murray,

    1. I don't see the specific package designation for the TL074 that you are using.  Please double check the package designation and pinouts to make sure that the device is not miswired.  For example, are the positive and negative supplies reversed. I doubt this is the issue, but it does happen so please double check.  I can check if you give me the full part number.
    2. You mention that you removed the LM339 and you still get the self heating.  Can you please remove Q1/Q2 as well?  This is to check that absolutely nothing is connected to the output.
    3. Do you know if the device is functional at all?  Maybe an initial transient damages the device and it then draws a lot of current.  If the device is damaged without anything connected to the output, the damage is likely from the power supply.  The input has a 47k ohm resistor in series with it and normally this would protect the input.  So, if you place a new TL074 on the board and the output is not connected to anything, the only place where electrical overstress can happen is on the power supply.
    4. Did you purchase the TL074 from a  TI’s authorized source?  It does happen that sometimes TI device are counterfeited.  Here is the TI anti-counterfeit policy:   https://www.ti.com/support-quality/quality-policies-procedures/anti-counterfeit.html . In the case of counterfeiting, sometimes the internal device really does not function as expected. I will say that normally counterfeiting happens on more expensive devices so I don't think that is likely here.
    5. Does the device always get hot?  
    6. If the device has a package like an SOIC, it is generally possible to desolder pins and lift them off the PCB.  You can trouble shoot by doing this.  For example, you can lift the output pin.  You could also lift the power supply pins and add a series ammeter to measure IQ and see if IQ is excessive.
    7. As an experiment, do you have a different op amp with the same pinout?  Does that device also get hot?
    8. Please double check your PCB layout.  Maybe there is a routing error.
    9. Also, with the PCB, maybe you have a PCB defect?  For example an unexpected short.

    I hope one of the above suggestions help. This is an unusual issue as you don't expect an op amp with nothing on it's output to self heat.

    Best regards, Art

  • 0 in reply to Art Kay
    Prodigy 20 points

    Thank you for your help Art. To answer your questions

    1) This is SOIC 3.5mm SMD package. Yes the part is put on correctly and is working well apart from getting very hot.

    2) The Q1/Q2 are actually on a different TL072 chip and this chip is not getting hot. The chip that is getting hot has only 4 s/h amps on it.

    3) Yes the device is working

    4) These were purchased from Digikey in the US

    5) Yes. In a previous question thread it was suggested that the common mode of the LM339 might be the cause. Removing the 339 does not make any difference.

    6) By removing the 339 there is nothing connected to the output and there is no rouge resistance or capacitance on the circuit that I can find. There are actually 12 circuits using 3 TL074 chips for all the S&Hs. All three of the 074s doing the S&Hs are behaving the same and getting very hot. There are also 3 074s doing the voltage level adjust and these are only getting slightly warm. 

    7) I have ordered an ST version of the TL072 and Texas LF347 chips to try in the board to see if there is a difference.

    8) This board has been on sale for a while now and is working correctly otherwise. I worry about the reliability of the product with this happening.

    9) No. This is happening on all the boards but interestingly some run hotter than others for no reason that I can see. My supply lines are +-14v.

    Regards

    Murray

  • 0 in reply to Murray Hodge
    TI__Mastermind 27868 points

    Murray,

    1. Great answer, thanks this helps a lot. 
    2. If I understand this correctly, the application works correctly even with the TL072 getting hot.  The product is released and you are concerned about a reliability issue (but the system works correctly otherwise).  Please correct me if I am misinterpreting your comments.  If the produce is working with the hot device, I share your concern.
    3. The device draws about 1mA of quicken current (P = 2*14V*1mA = 20mW).  This should not cause noticeable self heating.
    4. For item 2 above you mention that Q1 and Q2 are not on the circuit we are discussing.  Correct?  If not, does the schematic you posted have the circuit?
    5. I guess that the op amp is driving a capacitor through a switch (sample-and-hold).  A LM339 comparator is also connected.  I am unsure specifically which devices and configuration is being used.  If the op amp has a capacitive load and the signal is switching fast there will be a transient load current.  I think that you are likely dealing with something like that. 
    6. It would be useful to measure the input and output signals.  If there is a switched capacitive load, the average supply current will increase.  You could measure this with most DMM by lifting the supply pins and placing the ammeter in series with the supply.
    7. I think it is likely that the issue is not specifically related to the op amp but rather there is an unexpected load.  If you substitute the TL072 with a different op amp I think you will see the same self-heating.  Alternatively, if you test TL072 devices to a simple buffer circuit such as U8A and U8B with R77 and R78 removed, you should not see self-heating.  Note: I do not see decoupling capacitors on U8A and U8B.  These may be present but somewhere else on the schematic.  If not, that can be an issue.  Decoupling is required for good operation.  Some op amps will become unstable without proper decoupling.  Sometimes a circuit may work without decoupling but not always and proper decoupling is really a requirement for reliable operation.

    I hope I understood your comments correctly.  This is an unusual issue and I do agree that it needs to be resolved even if your product appears to work correctly.

    Best regards, Art 

  • 0 in reply to Art Kay
    Prodigy 20 points

    Hi Art

    Sorry if my schematic is unclear. There are 3 quad 074s that only have 12 S&H voltage followers (U7,8,9) and 3 other 074s (U2,4,6) that do the voltage conversion and have the FET attached as a waveform switch. The 3 S&H 074s are the only ones that are getting hot even when the output is disconnected. The S&H output load is only the input pin of the LM339 and the -ve input pin of the same 074 opamp and removing the 339 makes no difference to the heating. 

    So in summary, The 074s that are getting hot only have the output connected directly to the -ve input and the +ve input is only connected to a 10nF cap and the output of the 4051. The only load on the output is the 074 -ve input pin and the input pin of the LM339 and removing the 339 makes no difference.

    In trying to isolate what might be happening I have tried setting the 4051 output (and 074 input) to just 0v and also just +5v but this makes no difference.

    I will provide an update to see if another brand or the different type of JFET op amp makes any difference.

    Thanks again for your help.

    Kind regards

    Murray

  • 0 in reply to Murray Hodge
    TI__Genius 14795 points

    Hi Murray,

    Art is out for the next week so I'll try to cover here. I am now editing my original post because I realize that I had some errors/misunderstanding in my previous comments. You are saying the circuit with the JFET has no problems, correct?  All of the quads that use the JFET circuit are not hot, correct?

    Are these warm amplifiers still working normally?  You described this case: 

    In trying to isolate what might be happening I have tried setting the 4051 output (and 074 input) to just 0v and also just +5v but this makes no difference

    This was a good test, but I'm not sure if you are saying that the amplifiers worked fine for this case?  Or you were only checking if they were still hot?  Is it the case where these hot amplifiers are no longer functional?

    The last of my multiple questions here is:  is this a new circuit, or is this something that has worked previously, only just started having problems?

    Regards,
    Mike

  • 0 in reply to Michael Hartshorne
    Prodigy 20 points

    Thanks Mike for your input and I'm sorry for the confusion. There are 12 S&H circuits which use all of 3 074s (3x4=12). There are 3 other 074s that are doing the 12 wave shaping and level shifting. Only the 074s that are running as S&Hs are running hot. The only input to the S&H is the output of the 4051s and the connection from the output of the S&H back to the -ve input.

    The board and circuitry is running as it should and outputting as it should. The only problem is the S&H chips running too hot to touch and I am worried about reliability.

    Regards

    Murray

  • 0 in reply to Murray Hodge
    TI__Genius 14795 points

    Hi Murray,

    Ok, got it. I don't think you sent the schematic for the sample and hold, correct?  In the schematic above, I only see the JFET circuit and the buffer. 

    We have a few different types of sample-and-hold devices that use a closed-loop architecture, see below:

    www.ti.com/.../snoa223.pdf

    Are either of these configurations what you are using?

    Regards,
    Mike

  • 0 in reply to Michael Hartshorne
    Prodigy 20 points

    Yes, the 4051 switch, the 10nF cap and the JFET opamp is the S&H.

    I think I have solved the problem. I removed all ICs from the board except the 3 S&H 074s and they were still getting very hot.

    I purchased some TI LF347DR JFET opamps with the same pinout and these do not heat up. I have yet to try some ST brand 074s but I think I have the answer.

    So my conclusion is it is the Quiescent or supply current that the chip draws (TL074 10mA max LF347 11mA max). While the specs are the same the TL074 runs very much hotter. This also has something to do with my board layout as it is a pretty small board and these parts are packed closely and not allowing the PCB to act as a heatsink. Something in the way the 074 has been designed means it heats up more.

  • +1 in reply to Murray Hodge
    TI__Genius 14795 points

    Hi Murray,

    Ok, yes the ~9 mA*30 V is dissipating 1/4 W; for a Theta-JA of 160 °C/W, that equates to roughly a 40 degree rise.  The TL074 has moved to a smaller die so it may be that the small die size gives worse thermal performance than other devices on old processes.  So I will consider this one closed unless I hear otherwise, you can ping this post again if anything else comes up.

    Regards,
    Mike