• State Resolved
  • Locked Locked
  • Replies 10 replies
  • Answers 1 answer
  • Subscribers 49 subscribers
  • Views 1057 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.

LMH5401: High gain, High bandwidth, low noise

Matth. B.
Expert 1145 points
Part Number: LMH5401
Other Parts Discussed in Thread: THS4509, THS7530, LMH6629, TINA-TI, OPA855, THS4551, THS4541

Hello,

I need to design an amplifier with differential inputs (50 Ohms impendace or could be high impedance if it's an issue), single ended output, ac coupling.

I need a gain of 40 dB but I can reduce it to 30 dB and a bandwidth of around 100 MHz and it must be low noise.

I'm not sure to achieve all these requirements with only one opamp but I will try.

I have chosen 3 opamp, LMH5401, THS4509 or THS7530.

I also asking myself if it is best to build an instrumentation amplifier based on LMH6629.

Could you help me to make the right choice.

  • 0
    TI__Mastermind 26665 points

    Hello Matthieu,

       It might be difficult to achieve 40dB at this bandwidth with just one stage, but yes we can try once we narrow down the amplifier you need.

      Would you be able to share the application because using a fully differential is different compared to using an instrumentation amplifier. Both do have differential inputs, but there are different use cases for each product type. 

       Also, what is the loading requirements, i.e. what will this amplifier be driving?

    Thank you,

    Sima

  • 0
    Expert 1145 points

    Hello Sima,

    The application is to build an LNA with differential inputs (in order to use with differential signals). I don't know the system to measure so I assume the LNA would have 50 Ohms inputs but as I wrote could be High if it's a problem.

    About the load, it's a spectrum analyzer so a load of 50 Ohms (perhaps AC coupling).

    Thanks

  • 0 in reply to Matth. B.
    TI__Mastermind 26665 points

    Hello Matthieu,

      Thanks for the additional information. The three amplifiers you have chosen are good choices. Here are additional choices that could work, but would need a second stage: list. 

      I set up a Tina-TI simulation using the LMH5401, a very high 8GHz bandwidth part. I had to use a higher feedback resistor than recommended, but you can get 130-160MHz of bandwidth at 40dB using this part. 

      If you want to play around with the resistor values a bit more, I have attached an excel sheet below which helped me pick out the values for RG and RT. 

      For a part like this, I would recommend first testing this using our LMH5401EVM; since, you have to be extra cautious with the board layout. 

    LMH5401_HighGain40dB.tsc

    6558.LMH5401-SP_Single-ended to Differential Caluclator.xlsx

    Thank you,
    Sima

  • 0 in reply to Sima Jalaleddine
    Expert 1145 points

    Hello Sima,

    Thanks a lot for your support and the files provided. Just one remark, I'll use only one output of teh amplifier so my gain will be 6dB less than the one in your simulation but will always be ok for my application.

    Again, I need amplifier with the less noise. By checking the datasheet, I see that THS7530 could have less noise than LMH5401 but the THS7530 seems not to be a "standard" opamp (variable gain amplifier) and I don't know this kind of amplifier.

    Can I use it like a standard amplifier for my case or the best choice for you is the LMH5401?

  • 0 in reply to Matth. B.
    Guru 48395 points

    Morninng Matthieu

    Couple of ideas

    1. Since you say AC coupled, baluns can be very useful here

    2. Using an input step up 1:2 turns can get you diff to single at the input to us a simple op amp (OPA855 is pretty good)

    3. If you want to stay with FDA's, there are some very low noise decomp ones out there, but not from TI. There you would use an output balun like I showed in our char circuits for the THS4541 and THS4551 - on the THS4551, there are TINA files including those in the product page listing. 

    Here is an example using a very good Macom 1:2 balun that is way over 200Mhz BW in 50ohm I/O. The input to the gain R is relatively flat, but the final load (at 40dB gain) is too rolled off. I had left a 0.18pF parasitic on the feedback, 

    If i remove that parasitic, the BW goes out nicely, only 1.5dB down at 100Mhz now - lots of interesting things happening with that input balun - mostly it reduces the apparently noise gain for the op amp if  the source is considered, also a good noise reduction using that. 

    And the file, 

    OPA855 balun input high gain.TSC

  • +1 in reply to Michael Steffes
    Guru 48395 points

    I still had this file open in TINA so I ran an output spot noise at the OPA855 output pin, this 150nV would be input referred to the transformer by 200X to show 0.75nV equivalent input spot noise. The op amp itself is 1nV which  is increased by the resistor terms and current noise terms. That 200ohm input R required to get the 50 ohm match at the input of the 1:2 step up is 1.8nV by itself, but the transformer does act to reduce the apparent noise at its input to a very low 0.75nV. The 40kohm bias setups are intended to cancel bias currents (and the output does in fact sit right at 2.5V bias) and their noise is well filtered by the 10nF cap. 

    I think the op amp itself is pretty overcompensated at this higher gain (noise gain is 1+20k/400ohm = 51) In theory, I can reduce the phase margin with a feedback and inverting to ground cap to extend this by up to 1.6X  if need be. With 8GHz GBP, that NG=51 should be producing 157MHz which might be enough, but easy to get more with a couple of caps. 

  • 0 in reply to Michael Steffes
    Expert 1145 points

    Hello Michael,

    Thanks a lot for all your ideas, I think I have all to build my LNA.

    Thanks again

  • 0 in reply to Michael Steffes
    Expert 1145 points

    Another think I forgot. This design is nice for high bandwidth but I have to design also for low frequencies (around 10 Hz) so I can't use a balun.

    So, I think an FDA (or an instrumentation opamp) will be the right choice (perhaps not TI like you said).

  • 0 in reply to Matth. B.
    Guru 48395 points

    Table 4 in this article might be useful, you might have to bug the editors to the figure to expand again - link appears to be broken, 

    https://www.planetanalog.com/input-and-output-voltage-range-issues-for-high-speed-cfas-and-fdas-insight-2/

  • 0 in reply to Matth. B.
    TI__Mastermind 26665 points

    Hello Matthieu,

    Matth. B. said:

    Again, I need amplifier with the less noise. By checking the datasheet, I see that THS7530 could have less noise than LMH5401 but the THS7530 seems not to be a "standard" opamp (variable gain amplifier) and I don't know this kind of amplifier.

    Can I use it like a standard amplifier for my case or the best choice for you is the LMH5401?

      Yes that is correct, you can use it as a standard amplifier with a fixed gain or, depending on your input signal, even vary the gain between 30 to 40dB using the gain setting pins of the amplifier The LMH5401 would be the next choice if you are looking for the simplest setup for a FDA.

       But, both THS7530 and LMH5401 will provide a space constrained setup that will pass DC frequencies, as well as providing a high gain. For higher input frequencies, Michael's balun + OPA855 would be the next lowest noise option.

      More VGA/PGA FDA options that can provide 30-40dB of gain at your bandwidth can be found here. 

    Thank you,
    Sima