• State Resolved
  • Locked Locked
  • Replies 16 replies
  • Answers 5 answers
  • Subscribers 47 subscribers
  • Views 744 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.

TL074: Voltage follower running very hot

Murray Hodge
Prodigy 20 points
Part Number: TL074
Other Parts Discussed in Thread: LM339, TL331, LM393, LM339LV

Tool/software:

Hi

I have a TL074 acting as a voltage follower on the output of the 4051 analog switch to decode a mux signal. This is followed by another TL074 which inverts and ranges the output.

My problem is the voltage follower TL074s runs too hot to touch. The converter TL074 runs as I would expect.

Voltage supply is +-14v.

I can find no rouge oscillation and the chip running hot has me baffled.

Thanks in advance

Murray 

  • 0
    TI__Guru* 77541 points

    Hi Murray, 

    Murray Hodge said:
    My problem is the voltage follower TL074s runs too hot to touch. The converter TL074 runs as I would expect.

    TL074s are driving 10nF capacitive load without compensation. If you are measuring with DMM, you may not see the oscillation. But you will see the oscillation if it is measured with scope. 

    The circuit looks like the following, and you need to insert  Riso between 100Ω to 1kΩ in order to drive the 10nF capacitive load. You have 4 channels and each channel is driving 10nF loads, and they are getting hot because of the high frequency oscillation at output. 

    TL074 Driving 10nF 07182025.TSC

    If you have other questions, please let me know. 

    Best,

    Raymond

  • 0 in reply to Raymond Zhang1
    Prodigy 20 points

    Thanks Raymond for your reply. I am a little old school and I am not familiar with the term Riso. I am not driving the 10nF as it is on the input to the follower. The output goes directly to the input of a LM339 comparator with no other connections and I cannot see any oscillations. The 339 should only need nA current on it's input so maybe I need a resistor between the 074 and the 339. I'm finding it odd as this type of opamp use has been around since the 1970's. Regards

    Murray

  • 0 in reply to Murray Hodge
    TI__Guru* 77541 points

    Hi Murray,

    Murray Hodge said:
    I am not driving the 10nF as it is on the input to the follower. The output goes directly to the input of a LM339 comparator with no other connections and I cannot see any oscillations.

    Ok, I see. LM339's input should have 10s pF parasitic capacitance at the input + parasitic PCB etc. You still need to place 50Ω or 100Ω resistor after the TL074's output. It will improve that phase margin of the TL074 op amp. 

    There was a PCN issued in TL07x family in recent years, see the link below. So the post-PCN part has FAB change. The TL074's BW is increased and open loop output impedance is slightly altered due to new FAB process. This will affect the loop gain and op amp stability of the previous existing circuit. 

    If you are not sure, please check on the TL07x's date code. There are two digit number followed by a letter, and it should say 37 or larger, which is a date code. Please send me a TL074 IC's image and I can tell me if it is pre or post PCN products. The pre-PCN TL074 op amps should not have the overheating issues. 

    https://mm.digikey.com/Volume0/opasdata/d220001/medias/docus/5488/PCN20230627002.1.pdf

    If nothing has changed and there are no unused TL074 op amps (unterminated or floating), then it has to be a poor phase margin related to the overheating issues. Please place 50-100Ω resistor after each op amp's output and the overheating issues should resolve itself. If there are other unrelated changes in the existing circuit, please let me know.

    If you have other questions, please let me know. 

    Best,

    Raymond

  • 0 in reply to Raymond Zhang1
    Prodigy 20 points

    Thanks Raymond. Adding a resistor will require a design change and costs but if it will fix it that sounds good. Here is a picture of the chip in question. They are running too hot to touch and I am not sure if I can sell the product as it is.. They were purchased from Digikey.

    Regards Murray

  • 0 in reply to Murray Hodge
    TI__Guru* 77541 points

    Hi Murray, 

    The above image indicated that you are using the pre-PCN TL074C, and the op amps were manufactured in May, 2022. The overheating is not due to the op amp's instability. 

    Can you tell us which TL074C op amps are hot? Currently, I suspected that TL074C's output voltage may be too high that exceeds the LM339's input common mode range. Per the LM339's datasheet, the input should be kept between 0 and 5V-2V=3V (if Vs = 5Vdc is used). It is likely that your TL074C's output voltage exceeds the 3V requirements and source too much current..

    Can you tell me the TL074's output voltage ranges and LM339s' input voltage range - make sure that the LM339's input are not violating the input Vcm requirements. 

     

    Here is the Application Design Guidelines for LM339, LM393, TL331 Family Comparators Including the New B-versions, see the link below. 

    https://www.ti.com/lit/an/snoaa35f/snoaa35f.pdf?ts=1752859346056&ref_url=https%253A%252F%252Fe2e.ti.com%252F

    Could you snap a LM339 IC image onboard? We would like to check the comparator's date code. Please map out the voltage readings at TL074's output and LM339's input (should be measured at similar voltage levels), and tell us which TL074s are hot to a touch; we will provide you with suggestions next week. 

    Did you mention that the existing circuit has been working for years? We want to know what has changed on the BOM. So far, we ruled out TL074C op amp, which is the pre-PCN op amps. But we also want to confirm the onboard's LM339's date code. If it is necessary, I will forward your query to the comparator team next week (LVHAC team). 

    If you have other questions, please let us know. 

    Best,

    Raymond

  • +1 in reply to Raymond Zhang1
    Prodigy 20 points

    Common Mode? I never thought or checked that. One input is 0-5v in a Sawtooth wave and the other is a voltage 1-4v to get a variable pulse width from the 339 output. I will look at lowering the Saw level to 0-3v and the other input I have control over via software. The 339 will still give 0-5v PW. Interestingly the other 074 in the circuit which is supplying the Sawtooth wave also gets above room temp. Would current reducing resistors on both 339 inputs help?

  • 0 in reply to Murray Hodge
    Prodigy 20 points

    Hi Raymond, I want to thank you for your help with this. The CM is the problem and when I lower my control signal below 3v the 074s run much cooler. I was parking the PW control voltage at +5v when it was not selected which was most of the problem. I will change the 339 voltage to +15 in the next hardware release which should solve all CM issues. Thanks again, Murray

  • 0 in reply to Murray Hodge
    TI__Guru* 77541 points

    Hi Murray, 

    Murray Hodge said:
    I will change the 339 voltage to +15 in the next hardware release which should solve all CM issues.

    You seem to figure out the TL074's overheating issues. When IC's temperature over 60C, it is hot to the touch (likely you can place a finger over the IC without burning sensation). Between 60-70C, you may be able to touch for no more than 2-3 seconds. Over 70C, you are going to sense sharp "burning" or "hot" sensation to the touch. 

    Based on the LM339's date code, the comparator was manufactured around Aug. 2021. It is likely that this may be pre-PCN device as well. (PS. I am not very familiar with comparator's FAB change history, so this is my guess). There is also a PCN associated with LM339 product due to FAB change. 

    LM339's input stage defined the operating region. Exceeding the recommended operating region, the differential input pair is no longer behaves predictably, and the internal junction may conduct unintended current (outside of datasheet's specification).

    If you are going to modify the next hardware revision, I would recommend to place 10kΩ resistor between each output of TL074 and input of LM339 comparator - to limit the comparator's input current.

    If you have other questions about the LM339 comparator's behaviors, I can forward the query to the LVAC team. 

    Best,

    Raymond  

     

  • 0 in reply to Murray Hodge
    TI__Guru 63416 points

    Hi Murray,

    The LM339 input will remain high-impedance even when above the supply…up to 36V. So even if the op-amp drives to +15V, it would not draw any current (bias current would be <1pA when above ~4V on a 5V supply). So I do not think that is the problem.

    However, if the op-amp outputs are trying to go negative below ground, then the input of the LM339 is clamping the op-amp output to GND. That will cause things to heat up as the op-amp output tries to sink 10's of mA's of current.

    The danger is that the LM339 *WILL* fail at cold – especially with when both inputs are greater than 3.5V input. Input range shrinks at cold. It may seem to “work” at room temp right now, but it will eventually fail (see section 6.7 about the “feature”).

    Please see the LM339 family appnote, sections 6.4 and 6.7, this is all explained there:

    https://www.ti.com/lit/an/snoaa35f/snoaa35f.pdf

    The LM339 you show is the original Ji1 die from 2021 – so like the TL074, nothing has changed.

    I think the problem is elsewhere…

    What are the other two comparators U1:C and U1:D doing?

    Have you measured the voltages at all the pins? Are they what you expect? If there are any negative voltages anywhere on the LM339, then that is your problem...

    Once the heating problem is found, I would recommend using the LM339LV, which has a R-R input.

  • 0 in reply to Paul Grohe
    Prodigy 20 points

    Thanks Paul

    No 339 inputs are going negative as far as I can see. The too hot to touch 074 was actually a faulty 4051 that was outputting a neg voltage. One input on the 339 is a 0-5v saw wave which should never go negative and the other input is under software control and has a range of 0-3v which comes from the 4051 sample and hold 074. The other 339 comparators are working exactly the same for other channels. I think this is solved as the board is running about room temp now. I will look into the LV version.

  • +1 in reply to Murray Hodge
    TI__Guru 63416 points

    Hi Murray,

    We really recommend using the LM339LV device. You are hanging on the edge of operation with the LM339 and it will eventually fail.

    We still recommend a resistor (~1k-10k) between the op-amp output and the LM339 input - to limit the current during a transient or failure.

  • 0 in reply to Paul Grohe
    Prodigy 20 points

    Thanks Paul. I have ordered the LV version and will use it in all future products. In the common mode section of the data sheet it has (GND) and (V+) in brackets for the Min and Max. What do the brackets mean?

  • 0 in reply to Murray Hodge
    TI__Guru 63416 points

    Hi Murray,

    The parenthesis really don't need to be there. It's a convention to state that they are supply names.

    Sometimes there is a value in reference to the supply, and it would be placed in parenthesis, such as (VCC-2) or (V+) + 0.2.

    Just remember that the LV is 5V supply max. You could also power the LM339 from the +15V supply and eliminate the input range issue.

    Yes...Those 4051 CMOS switches can be damaged easily - usually by shorting one of the outputs (to much current) or inputs exceeding the supply. I have replaced may of them over the years...

    Glad to help!

  • 0 in reply to Paul Grohe
    TI__Guru 63416 points

    Hi Murray,

    I see you have a bunch of 10,000pF caps directly on the lines of the mux. Those are large enough to store some serious charge.

    I could see a scenario where the caps are charged up to several volts and the power is turned off - then you would have multiple caps (off channels) dumping current back into the switch (ESD diodes clamped to the dead supplies). It may eventually get tired of that...

    If you keep having 4051 failures, try removing the caps.  I would recommend a series resistor between the switch output and cap to limit any discharge current - like say 470 ohms to 1k. But beware of the low-pass filter that creates. Also make sure the mux input is not being over-driven.

    As part of the final design check, you may want to scope the various nodes at power-on and power-off. You have multiple supplies and there could be a power sequencing problem..

  • 0 in reply to Paul Grohe
    Prodigy 20 points

    I have been repairing vintage synths for 50+ years and almost without exception they put a 0.01uf cap on the output of the 4051s with no resistors.

    This is an example from a Roland synth from early 1980s. This could be why they fail after 40+ years and I keep a draw full of 4051s.

    The cap is needed to hold the sample and hold output until the next refresh and some have 20+ outputs and the refresh time can be long due to component speeds and cap charge time.

    Thanks for your help.

  • 0 in reply to Murray Hodge
    TI__Guru 63416 points

    Hi Murray,

    I also have been fixing audio and test equipment for 40+ years...and the 4051's are popular input selectors and range-changers. So they get beat up in all sorts of nasty ways..

    The caps on the outputs are not an issue during operation. The charge current will be limited by the source - but the uncontrolled discharge during power-down could cause bruising to the ESD cells.

    As a heads-up, the classic TL074 JFET input did not have a ESD clamp to V+ (could take up to 36V, even with no power). The PCN'ed TL074 do have a ESD clamp to V+ - so a similar issue could occur with the op-amps.