• State Resolved
  • Locked Locked
  • Replies 12 replies
  • Answers 2 answers
  • Subscribers 50 subscribers
  • Views 650 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.

TL081H: Output stage difference from legacy TL081

Alex Pogossov
Prodigy 140 points
Part Number: TL081H
Other Parts Discussed in Thread: TL081, TL084H, TL084

Hello,

From datasheet SLOS081M is appears that output swing, especially positive for the new TL081H is close to the rail (500mV).

Old TL081 can go only 1.5V from the rails.

 However, functional diagram (s.8.2, p.29 of the document) shows a rather conventional ols school output power stage which would not go closer than probably 0.9V to the rails, at best.

Does the new TL081H has some unique (undisclosed) power stage circuit, or is there an error in the datasheet?

Regards, Alex

  • 0
    Guru 267410 points

    The functional diagram shows the TL081.

    The TL081H is a CMOS device with a completely different architecture. Its internals are not documented, but probably look like most other CMOS opamps:

  • 0 in reply to Clemens Ladisch
    Prodigy 140 points

    You might be right, Clemens. Higher noise density particularly at low frequencies gives away CMOS nature. I guess it does not use the folded cascode, but a more conventional current mirror at V+ rail, but it is only my speculation. Or only one folded cascode as it is not fully rail-to-rail input part. 

    However, the datasheet, for some reason, now combines both old TL08x and new TL08xH into one document which is confusing given they are quite different beasts.

    Maybe an engineer from TI can further elaborate on the differences between TL08x and TL08xH.

  • 0 in reply to Alex Pogossov
    TI__Guru*** 144736 points

    Hello Alex,

    The shared data sheet is conducive to encouraging adoption of TL07XH. Many decades ago JFET was the only way to get low input bias current. TL07XH is a CMOS device that is similar to TL07X (in bandwidth and slew rate) but better in most ways. The input is just NMOS transistors and the output is MOSFET drain output like the picture.

  • 0 in reply to Alex Pogossov
    Guru 267410 points

    The TL08xH actually has rail-to-rail inputs to prevent phase reserval, but the characteristics of the P-channel pair are not guaranteed.

  • 0 in reply to Ron Michallick
    Prodigy 140 points

    Glad to hear that it is actually R2R output. Also, output impedance without feedback is low (125 Ohm), so there must be some local feedback around the drain output stage.

  • 0 in reply to Clemens Ladisch
    Prodigy 140 points

    Interesting Clemens... and surprising to say the least.

    By looking at the humps/steps of Voffset versus Vcommon-mode graph, it looks like TI somehow "managed" to get crossover between n-MOS and p-MOS input differential stages right in the middle of the common mode range. How did they do it, I am wondering? Usually crossover happens somewhere about 2V off the rails (when one of the diff pairs goes into cut-off) and in the most of common mode range an amp works linearly (with both diff pairs together). Obviously not in TL081H. It looks like if the amp is used with bipolar supply, THD will be increasing with audio signal diminishing. 

    For the specs to look better TI specified THD at 6Vrms. Had they specified it at 500mVrms, THD would have looked awful. So this op-amp is no good for audio.

  • 0 in reply to Alex Pogossov
    TI__Guru*** 144736 points

    Clemens, there is no PMOS backup pair. Once both NMOS cutoff, the output may be low or high based on the balance of the rest of the amplifier input stage.

    Alex,

    I believe the "cross over" in the data sheet chart is an artifact of the test method used to get the data.I suspect that there is some load in the setup that changes current flow direction. That contributes to the little bump. 

    I did a quick search though my own data. I didn't find a 40V Vos vs Vcm chart , but I did find a 5V (+5V, 0V) chart (two samples).

     

  • 0 in reply to Ron Michallick
    Prodigy 140 points

    Thanks Ron,

    It looks like you know more about TL081H that the guys who wrote the junk datasheet.

    I wish (Christmas wish) that TI amends the datasheet, remove all [bulls**t]/junk and publish true information and correct data. It would have cleared confusion, unnecessary Q&As and contribute to TI's reputation.

  • 0 in reply to Alex Pogossov
    Guru 140200 points

    Hi,

    the TL084H shows a noise of about 80nV/SQRT(Hz) at 100Hz while the legacy TL084 only shows about 26nV/SQRT(Hz).

    Kai

  • 0 in reply to kai klaas69
    Prodigy 140 points

    Yes, high noise gives away CMOS front end. I guess it is impossible to make a JFET rail-to-rail op-amp, as close to a rail one of the JFET differential pair would saturate and the JFETs will draw gate current. It is totally unacceptable. So we have to take a penalty of high noise for rail-to-rail input feature.

    Even if a JFET op-amp common mode range includes one of the rails, still it will not be able to reach an opposite rail as close (1.5V) as TL081H.

    I believe TL081H is a useful device, but the datasheet needs corrections.

  • 0 in reply to Alex Pogossov
    Guru 140200 points

    Hi Alex,

    what are your exact needs?

    Kai

  • 0 in reply to kai klaas69
    Prodigy 140 points

    Hello Kai,

    My application is single supply - 35V, but common "ground" is positive. So TL081H looks perfectly suited, as its common mode range includes positive rail rather than a negative rail. It will work as a relatively slow integrator, so no issues with bandwidth and slew rate.

    The main reason I started this thread is because the datasheet for TL081H was soooooo confusing and in fact did not do TL081H justice. Below are errors in the datasheet:

    1) Nowhere it says that TL081H is CMOS op-amp;

    2) Schematic diagram depicts an old school TL08x, which is not capable of rail-to-rail output operation;

    3) Voffset versus Vcommonmode plot shows some bumps in the middle which look like crossover distortion in the front end and raises suspicion about suitability for audio.

    Thankfully the "gurus" on the forum clarified all this junk in the datasheet, and now I designed this op-amp in hoping it will work as expected.

    === If TI makes a separate "true" datasheet for CMOS TL08xH, and keep it distinguished from the old JFET TL08x, that would be so much better, IMO. ===