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.

THS4552: FDA feedback resistors values

Part Number: THS4552
Other Parts Discussed in Thread: TINA-TI

Hello all,

In my project, I'm using a IQ Modulator (LTC5599) and a IQ modem (CMX7164), and I have two "problems" interfacing the IQ outputs of the modem to the IQ modulator inputs.

1. The DC common mode voltage of the modem's I or Q outputs is 1.6V, while the DC level required for the modulator is 1.4Vdc

2. The maximum load of the modem's I or Q outputs is 20kohm, while the input impedance of the modulator is about 1kohm

Here is the IQ Input, and outputs specs:

Modulator's inputs:

Modem's Outputs:

I thought about using a dual fully differential amp (such as the THS4552), because the IQ interface is differential, and I want to use a single chip for the solution. By using this solution, I can interface the modem to the modulator using a single chip, and the 1.4V common mode DC level can be adjusted using the VCOM pin of the FDA

I do not have much of experience in using FDAs...by looking at application notes of different FDAs, I have noticed that all of them have a relatively low  feedback resistors (1-2kohms), which create a input impedance of a several kohm, while I need to keep the input impedance above 20kohm...at least.

Question 1:

Can I use a larger values of feedback resistors? (20-100kohm) - what would be the consequences on the FDA performance?

Question 2:

Can someone offer a different/better solution for this matter?

One more thing worth to mention - My baseband bandwidth is relatively low, only a few KHz wide.

Thank you all!

David

  • Hi David,

    To answer your questions:

    1. Yes, it is possible to use a larger value of feedback resistor for the THS4552. The trade-off is going to be some roll-off in the frequency response and increase in 4kTRf noise at the FDA output. If the base-band bandwidth is relatively low on the order of few kHz wide, then I think increasing the Rf value to 20kohms (or higher for G>1V/V) should not be of significant impact in-terms of bandwidth.

    2. I think using the THS4552 FDA for level-shifting and buffering operation is the best approach for I/Q modulation. You could use the below THS4552 TINA-TI circuit that is optimized for your application in-between the modem and the modulator. The THS4552 operates on single 5V supply with an output CM of 1.4V because the device has a rail-to-rail output. If the desired output swing is higher, then it might be better to center the THS4552 supplies (say +3.9/-1.1V) around the output CM for maximum output voltage swing.  

    Here is the TINA-TI circuit for the same: THS4552_IQMod.TSC

    Best Regards,

    Rohit

  • Hi Rohit,
    Thank you very much for the detailed answer, and for the simulation!
    The output swing that required for the modulator is 1Vp-p differential (or 500mV single-ended) so working with a single supply of 3.3V over 1.4Vdc offset should be just fine...
    Thanks again, You've been a great help.
    Best regards,
    David

  • Hello again,

    Following my last post regarding using FDA to interface between a IQ modem to a IQ Modulator:
    I need to adjust my output common voltage around 1.4V, while the IQ outputs of the FDA are centered around 3.3V/2 = ~1.65V.
    I was wondering if I can use a simple resistor in parallel to the voltage divider bottom resistor, instead of connecting an external DC source of 1.4V to Vocm pin?
    /cfs-file/__key/communityserver-discussions-components-files/14/5633.3365.1.pngThanks,
    David
  • Hi David,

    You could probably set the Vocm input voltage to 1.4V using only a 180k resistor to GND. However, I would not rely on the accuracy of the internal 60k ohm resistor to set the Vocm input voltage. Instead, I would recommend to use the below voltage divider circuit in-order to decouple or minimize the impact of the internal 60k resistor while setting an accurate 1.4V Vocm.

    Best Regards,

    Rohit

  • Hello again Rohit,

    I will try this solution on the EVB.

    Thank you for your kind help!

    David