• State Resolved
  • Replies 15 replies
  • Answers 5 answers
  • Subscribers 16 subscribers
  • Views 275 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

TMUX136: acPSRR is not specified, can you provide any kind of information ?

Peter Soerensen
Prodigy 60 points
Part Number: TMUX136
Other Parts Discussed in Thread: TMUX1072, TMUX4053

Tool/software:

I have looked through all yours multiplexers and competions, the best and almost only choice to fullfill our requirement in function, very low capacity, supply voltage, size and price is the TMUX136.
Its primary use case is digital, but I will use it for analog signals. Multiplexers intented for analog use normally have an ACPSRR spec of say 40dB, TMUX136 datasheet do not.
Our frequency range for PSRR is only from 10K to about 10MHz, anythink lower or higher is don't care. Can you provide any information of PSRR we can expect?
Is the internal construction of the TMUX136 different from analog muxes. is there a mux with similar internal construct that has a PSRR spec?
We also want to know if there can be coupling from the SEL  pin to signal?
The noise signals on power and SEL will be below 10uV. We want the coupled noise on the A/B/COMMON to be significant lower like <100nV!




  • 0
    TI__Genius 10055 points

    Hello Peter,

    Our multiplexers are analog, which means they can pass both digital and analog signals. 

    Regarding the PSRR concern, we usually recommend a 0.1uF and 1uF decoupling capacitors on the supply pin for best noise performance and eliminate any power supply noise coupling concerns. We do not have the PSRR characterized or specified for this device.

    I can't share too much information about the internal structure of the device, but it has a charge pump topology. Please see FAQ for more information. 
    The TMUX4827 also has that structure, please see its ACPSRR spec below:

    Yes, there will be some coupling from the SEL pin to the signal, but implementing decoupling caps on the line should take care of the noise. 

    Thanks,

    Nir 

  • 0 in reply to Nir Gilgur
    Prodigy 60 points

    Thanks for your information, great help.
    Even though we have extensive supply filtering we can see our supply switch frequency on the amplifier output.
    We use TMUX136 on the input on a low noise input amp where <100nV/sqrt(Hz) noise matters. So far we do not know if the noise come from the mux or the opamp, it may very well be the opamp where the PSRR is only 25dB, apparentlymuch worse than the mux.
    The noise will be investigated and Power filtering will be further improved and.

    We also see some unknown noise frequencies on the output spectrum.
    I will be putting the PCB in a faraday case to rule them out as external from all our different equipment in the building.
    You information about a charge pump in the Tmux136 is very interesting. The frequency of the charge pump may be one of what we see on the output.
    What is the approximate frequency of the charge pump ?
    (So I can rule it out as one of the noise sources)

  • 0 in reply to Peter Soerensen
    TI__Genius 10055 points

    Hello Peter,

    Understood, the frequency of the charge pump is internal information and not something I am able to share. 

    What I can say is the charge pump structure does tend to be pretty noisy. 

    Thanks,

    Nir 

  • 0 in reply to Nir Gilgur
    Prodigy 60 points

    Is it possible we can get the information directly under NDA?

  • 0 in reply to Peter Soerensen
    TI__Genius 10055 points

    Hello Peter,

    I am not sure if an NDA will cover such information. 

    I will communicate this with our internal team and come back to you with a confirmation.

    Please expect a reply by middle of next week.

    Thanks,

    Nir 

  • 0 in reply to Nir Gilgur
    TI__Genius 10055 points

    Hello Peter,

    Mistake on my part, no need for an NDA for charge pump frequency information. 

    The TMUX136 has a charge pump frequency of around 870kHz when Vcc = 2.2V.

    Thanks,

    Nir 

  • +1 in reply to Nir Gilgur
    Prodigy 60 points

    Perfect, Can you also say some think about how much it may vary with temperature and process variations?
    I plan to use a 3.0V supply

  • 0 in reply to Peter Soerensen
    TI__Genius 10055 points

    Hello Peter,

    Here is the information I have: 

    - Charge pump 3V 25C = 750kHz
    - Charge pump 3V 85C = 700kHz
    - Charge pump 3V 125C = 650kHz

    Please let me know if this answers your question.

    Thanks,

    Nir 

  • 0 in reply to Nir Gilgur
    Prodigy 60 points


    Thanks a lot for the detailed information.
    I looks like TMUX136 send out a noise current of approximately 1nA on the charge pump frequency!
    I will verify further after new year, when I am back in lab, but I am 95% sure this is the problem.
    1nA is about a 100 times more than I can tolerate as this preamp must handle signals levels all the way from noise floor of less than 10nV to 0.5Vp, impedance needs to be higher than 50R but even at 50R it is too much.
    My ideal mux would be a TMUX136 WITHOUT the charge pump, higher VCC would be okay, but I seems it does not exist?
    The closes match on capacity, price and size I can find is TMUX1072, It has 3 times higher capacitance but close on price and size. Ron and DC leakage is not important. The TMUX1072 datasheet do not mention a charge pump even though it is specified down to 2.3V supply. Please confirm that TMUX1072 do NOT have a charge pump?
    If TMUX1072 also have a charge pump please propose one without.


  • 0 in reply to Peter Soerensen
    TI__Genius 10055 points

    Hello Peter,

    Can you please explain why the TMUX136 is the ideal mux for you application? What are the parameters it meets for your use case? 

    If you can explain a bit more on what specs you are looking for in a mux (other than not having the charge pump architecture) I might be able to recommend one that will fit. 

    I can confirm the TMUX1072 does have a charge pump.

    Thanks,

    Nir 

  • 0 in reply to Nir Gilgur
    Prodigy 60 points

    2ch 2:1, Very low capacity on analog I/O, low price, tiny package.
    Max Analog voltage +/-0.9Vp
    Supply voltage of 3V is preferred but higher is acceptable.
    Select voltage of 1.8V is preferred.
    Ron can be rather high like <50R,  DC leakage < 50uA

  • 0 in reply to Peter Soerensen
    TI__Genius 10055 points

    Hello Peter,

    After checking our portfolio, our best option we have is the TMUX4053. It is a 3-channel 2:1 with low capacitance, but the Ron might be too high.

    Most of our low Con devices have the charge pump topology so finding a part that fits all of your requirements, unfortunately is not possible. 

    Please let me know if you have any more questions.

    Thanks,

    Nir 

  • 0 in reply to Nir Gilgur
    Prodigy 60 points

    I don't think I can reach my goal with the TMUX4053, I can only use two out of the 3 channels and this means it take significant more board space. Second the high Ron is a going to be a design challenge. Last the capacity is also higher than TMUX1072. TMUX1072 seems a better choice.
    Please double check if the TMUX1072 contains a a charge pump, according to the datasheet is does not, second the datashet say it can be used for data equations.

  • 0 in reply to Peter Soerensen
    TI__Genius 10055 points

    Hello Peter,

    Understood, I have reached out to our design team and they confirmed the TMUX1072 has a charge pump topology.

    Thanks,

    Nir 

  • 0 in reply to Nir Gilgur
    Prodigy 60 points

    Not good, I guess I am not so lucky that the charge pump is turned off if I use max vollage of 5V.
    It seems I muss start this design over after new year.