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.

ADS5407: How to read negative values from high speed ADCs

Part Number: ADS5407
Other Parts Discussed in Thread: THS4511, LMH5401, LMH6401, LMH6521, LMH6559, TIDA-00826


I have a two signals whose frequency is 100-200Mhz

i picked ADC which has DDR LVDS interface and sampling rate 500MSPS with 12bit width and dual channel, and importantly stock availablity which is ADS5407 

Can any one guide me how to use the ADC ADS5407 for below things

1. protection :

At maximum my signal will swing between -1V to +1V ,but there is a chance of +2v and -2v in worst cases, how to protect ADC from this worst cases situation

2. reading negative values:

my signal is of two channels, each channel swings  in range of -2v to +2v, absolute maximum

in such case if i need to read negative going pulse what should be my Vcm ?

i assume the adc can read between 0-Vref only , kindly help how to understand and read negative going signal

  • Shyam,

    Depending on how fast your signal switches from 1V to 2V, a transient-voltage-suppression diode may work for you for protection.

    If you use an amplifier, you can also reduce the gain of the input so that the output is within spec of the ADC.

    If you AC couple the ADC inputs, the device is self-biased so you will not have to worry about the VCM. You can use a negative input and it will be centered about VCM but make sure the swing is per the data sheet (see below).

    If you need DC coupling, each input pin (INP, INM) must swing symmetrically between (VCM + 0.25V) and (VCM – 0.25V), resulting in a 1.0Vpp (default) differential input swing. 

    The high speed amplifier forum can provide more assistance with this if needed.



  • 1. There is a risk of using amplifier to attenuate the signal here, if i attenuate for the sake of wost case signal i will attenuate the important low voltage signals also which makes system detection level poor, so a TVS diode is good option for transient, but in my case signal may not be transient in nature so i may need a zener ?

    but before selecting any protection diode i have below doubt

    2. 1 Vpp is a good parameter for me, if i assume Vcm is 0 the siwing is between +250mv to -250mv, which is 500mv swing total, how did you say its 1Vpp ?

    3. if i check absolute max rating, its rated at 3.3v+0.5 , which is 3.8v approximately, what would be -ve side the maximum value , as per datasheet its 0.5v , so i assume the ADC need to be protected in -ve side only ?

  • Shyam,

    Do you plan on using AC or DC coupling for the analog input pins?



  • I wish to have both options, i am making a prototype

    most preferably ac coupling

  • Shyam,

    Each input (P & N) will swing +/-0.25V about VCM which is a total swing of 0.5V per input. This equates to a 1Vp-p swing differential mode.

    You need to protect against both -ve (-0.5V) and 3.8V.

    If using DC coupling make sure your VCM on the inputs is within spec of the data sheet.



  • I would repeat my requirement again

    my signal is a single ended signal which is of RF type - 250MHz coming from a optical device (Datasheet)

    There are two signals I and Q, one is sinusoidal other cosine

    both vary in the range of +2v to -2v

    i wish to keep both options which are ac couping and dc coupling as done in 5407 eval board, so in this respect i wish to know if my selected ADC fits in or need to select any other ADC ?

  • Shyam,

    Please send your schematic if possible. Your message is confusing. If you operate the ADC with a single-ended input you will lose 1/2 of the ADC resolution. Not sure if this is what you want.



  • i did not design any schematic yet, i am in process of selecting ADC, below is my signal nature

    this is my signal nature when there is no activity at sensor

    the sensor output has two channels I and Q, the yellow one is I which has dc shift inherently up to -300mV on the other hand Q which is green is having a +ve dc shift of 150mV

    my objective is not to remove DC , so coupling will be DC

    i wish to make my data converter Ckt in range of DC-500MHz,

    the signal dc shift may vary with time(temperature effect)

    the signal range over dc is in range of 500mVpp

    what is the way forward, how should i interface this type of signal to a 500MSPS ADC like 5407

  • Hi Shyam, I'm myself am not sure if what you are asking for is possible with only the ADC. I am curious as to why the I data must be 0 through -300mV? Would a possible solution be to shift the input by a high precision voltage reference, such that the AC component of the I tone is preserved (swinging between 0V and -300mV), just that it is shifted to within the ADC input range?

    Please await a response from another team member who is currently out of office. He may be able to assist you better.

    Thanks, Chase

  • Hi Shyam,

    I sort of agree, with Jim and Chase. It would be good for you to at least give us a hand drawn block diagram of your application. It is difficult to guide you with so many requirements. We could recommend several options.

    From what I understand, you have the following requirements, please correct me if I am wrong or in error.

    1) you need to both AC and DC couple the signal

    2) you have a very large analog input signal coming from the output of the sensor, -2V to +2V, 4Vpp

    3) The signals are I/Q coming from the sensor

    4) The signal are 500MHz and below coming from the sensor

    5) The sensor is a single-ended output, not differential for both I/Q

    If this is the case, I would use the following block diagram as a start, you can both AC couple and DC couple the amplifier's interface by inserting a AC blocking cap on the outputs of the amplifier.

    You can insert an attenuation pad before the amplifier, to adjust the level from 4Vpp to ~1Vpp, which the amplifier/ADC can handle on their inputs.

    The amplifier's gain can be set to near a gain of one and will help do the single ended sensor output to differential which the ADC will need for its interface.

    Hope this helps.



  • Dear Sir, 

    Very thanks for your valuable time allotted in answering 

    My signal is 250MHz so i am choosing 500Msps ADC, you are correct the signal is a single-ended

    I have gone through TI 2 GHz OscilloScope design which i felt is a good starting point -

    I have changed the below things and let me know if i am going right

    I have removed input attenuators(Teledyne) in design because i will manage with below point 3

    1. Input FDA LMH5401 to re-place with THS4511 - 2Nos - Comparision

    2. PGA LMH6401 to re-place with LMH6521( Dual Channel) - Comparision

    3. Buffer LMH6559 with no change - 2Nos

    4. ADC12J4000NKE to re-place with ADS5407 - Dual Channel(Sufficient for my application)

    5. Power Solution -

    I feel it's better if I go with PGAs in place if my signal exceeds or nears limits after DC shift i will have the option to improve the signal and also play with the dc shift through the LMH6559 loop.(Correct me if i am wrong in picking a PGA in the chain)

    I will take the ADC output from 5407 and bring it to an FMC to make a Full length "Dual-channel 500MHz acquisition card with DC - Coupling and PGAs " from the above setup

    Kindly recommend if any changes are required, Items on TIDA-00826 are a bit costlier and over-spec for me

    I am playing at the Sub-GHz line and have filtered the above with my limited knowledge

  • Hi Shyam,

    I believe this will work for you as well for your application.

    Please start a new post if you have further questions when you start your design.



  • I am left with few doubts

    1. Do i really require a jitter cleaner as used for JESD ADCs ? Are they required for Sub-GHz ADCs if so ? kindly suggest one

    2. What is the ideal mechanism of clock control, is it better if i generate the clock on my ADC board and give it to FMC ?

    or should i take CLK from FMC and give to ADC ?

    any way the programmer on FPGA side would sample the data based on his input recieved over DATCLK pins of ADS5407

    which will be better option CLK Driven from FMC or CLK Generated Locally

    Kindly suggest a TI Clock source for Input to ADC ADS5407

  • Hi Shyam,

    Typically we would start a new E2E post as these are different questions.

    I will forward you post to the clocking group for comments.



  • Re strcutured the question again to go ahead with LMH5401

    LMH5401: Tuning LMH5401 to handle FSR of +/- 2V or 4Vpp - Amplifiers forum - Amplifiers - TI E2E support forums

    Actually THS4511 is a single rail opamp which can pose problem to my desgin i feel, so i wish to keep the design of example as is for quick solution

  • Hi Shyam,

    I believe, both above E2E thread queries (Clocks & HS AMP) are already taken care. 

    I am closing this thread now and feel free to open new thread or reply on this for further queries.



    Ajeet Pal