THS4531DGKEVM: Convert single ended to differential output

Part Number: THS4531DGKEVM
Other Parts Discussed in Thread: THS4531

Hi,

I see that this Eval Board takes single-ended input and with unity gain converts to single-ended output (with some resistors and a transformer). Is it possible to use the differential output?

thanks,

Saurabh

  • Hi Saurabh,

    Welcome to the E2E forum, 

    By default, the THS4531 EVM is configured for single in and single out. We can very easily configure the EVM to be differential output with a population of R26, R29, R30 = 0Ohm. R27, R28 = DNP, and J11 be populated with an SMA jack. 

    Please let me know if you have any questions. 

    Best,

    Jacob

  • Hi Jacob,

    Thanks. Is it easy to change those resistors after I buy them? Or, can I order in such a way that it is set for differential output? For my application, I need single BNC/SMA to be carrying the differential output.

  • Hey Saurabh, 

    Super easy to swap these components on the PCB. Unfortunately we do not currently a different assembly version, but again this is an easy conversion for diff-out. 

    Full SMA input and output will be easy to implement, you just need to add the one extra SMA jack. 

    Please let me know if you need any other help with getting this test board setup.

    Thanks,

    Jacob

  • Thanks Jacob. Good to know it is easy change with resistors. Last thing - with that extra SMA, a single SMA line will be carrying the differential output?

  • Hey Saurabh, 

    The added J11 will be Vout- with the outer conductor being GND, and the existing J10 will be VOUT+ with the outer conductor GND. 

    VOCM will act as the voltage that the outputs are centered around. 

    Please let me know if this makes sense. 

    Thanks,

    Jacob

  • Thanks Jacob. It makes sense. What I would like to have is not 3 different SMA output lines but a single SMA/BNC output carrying +ve and -ve of the differential signal and centered around the mid of single ended input. As in the circuit I shared in the email.

  • Hi Saurabh, 

    As is, you would have two differential outputs (J10, J11)

    If we created a single SMA jack, you would lose your reference to ground. Is this intentional? Plugging this into any oscilloscope or other test equipment would effectively short the negative output to earth GND. 

    What is the intention with this setup? Can you share this email that you are referencing? 

    Thanks!

    Best,

    Jacob

  • Hi Jacob,

    It makes sense. I was confused. I can take SMA jack out that is at J10 and make a new connection with the center pins of J10 and J11 to make a differential output (or even center pins of J5 and J6?). I will have a 3-wire output, two for the differential and a ground.

  • Hi Saurabh, 

    Yes you could remove J10, but understand that the SMA cable interface is provided as an easy means of maintaining the characteristic impedance of a 50Ohm transmission line. The most reliable way to transmit this differential signal is via two coax cables like J10 and J11. Is the three wire cable a twisted pair with an outer shield? 

    J5 and J6 are the direct outputs of the amplifier. These can be used as well, though the output of the high-speed amplifier will not want to see direct connection to reactive impedance without some amount of series resistance (use R22,R23). 

    Thanks,

    Jacob

  • Hi Jacob, very useful. I am trying to lock (stabilize) a laser where the error signal is applied to the laser's piezo port. A lock box outputs a single-ended signal, 50 Ohm, that controls the piezo, and I would like to convert it to a differential one to make it more robust to noise and tighter laser lock. The laser piezo connection can accept a +ve, -ve and a ground line (laser has a DB-9 connector). So, if the signal from lock box (single-ended) goes to a differential amp (just two differential or 3 with ground) and then to the laser piezo, would be great. Right now, I am connecting the lock box output (+ve and ground) to the piezo +ve and ground only.

  • This is the circuit that the laser manual gives, 

  • I see, 

    Thanks for the diagram! In that case, we could direct solder the 3-wire DB-9 connector to J6/J5 and use something like a 10-100ohm Riso for R22/R23. 

    I agree with you, this should reduce some noise by running the signal differentially. 

    Thanks,

    Jacob

  • Hi Jacob,

    thanks for the help. Would I be able to solder standard resistors at R22/R23 just directly above the board? and I guess I can pick ground lines near J6/J5

  • Hey Saurabh. 

    Yes, you could solder these on the footprints on top of the board, access is pretty reasonable. Yeah, we can pull the GND from the outer conductors of the SMA footprint. This strategy does not require any removal of components, just adding of Riso and soldering of wires. 

    Please let me know how the modification goes. 

    Thanks,

    Jacob

  • Hi Jabob,

    Sounds great. Thanks a lot for the help!

    Saurabh

  • Anytime!

    Best,

    Jacob