THS4524EVM: Adding up to 4 RF frequencies with THS4524EVM

Hello Dmitri,
The THS4524 only has around 150MHz of bandwidth, so it isn't ideal to use it in a 100MHz application....the amplifier is not very linear close to its gain bandwidth product. The THS4541 may be better suited for such an application.

Are the signal you want to add fully differential or single ended?

You could certainly configure the FDA like and "Op-amp Adder" or "Summing Amplifier" in order to achieve this. (You can google this term to see what an op-amp adder is in case you don't know.

-Samir

Dear Samir, thank you for your reply. The signals that I'm adding are produced by a DDS board, same ground for all of them. So I suppose they can be called fully differential.

I must confess I prefer an off-the-shelf solution. I have trouble finding an op-amp adder as a standalone product that matches my requirements.

Earlier I found this article discussing the use of differential amplifiers for summing frequencies that discusses some disadvantages of classic op-amp adder, that led me to this THS4524EVM evaluation board. http://electronicdesign.com/analog/high-speed-noninverting-summing-amplifier-operates-220-mhz

Quoting: "The circuit in Figure 1 has many other disadvantages, including low input impedance, different input impedances for positive and negative inputs, narrow bandwidth, and the need for tightly matched resistors."

Should I be concerned with these "disadvantages"? The most concerning part for me is "narrow bandwidth".

Thank you for your help!

Dmitri

Dmitri,
Not knowing the details of your driver/DDS makes it difficult for me to suggest a solution. I can tell you that TI does not offer any off-the- shelf adders. ...you will have to purchase resistors and amplifiers and build one.

The THS4524 solution is not going to work if you require a high input impedance. If your DDS can drive a resistance of a couple of hundred ohms then that should not be a concern.

Different input impedances - If you want to add the signal with the same scaling factor then this is not a concern. For example if you want V1+V2+v3+v4, then the resistors will all be equal. Also if your DDS has a strong drive capability then again this shouldn't be a problem. The different input resistance only becomes an issue if the DDS cannot drive the load and in that case each paths linearity/performance will be compromised because of the DDS drive capability.

Tightly matched resistors - Yes depending on accuracy you will need that.

"Narrow Bandwidth".... Here is where you will need an amplifier about 10x your bandwidth of interest to ensure linear performance (low distortion)

Also, why do you need a quad amplifier like the THS4524... a single one should do. It would be very helpful if you could download TINA-TI (SPICE simulator) from the TI website and the SPICE model for the THS4521 and test your proposed circuit there with the frequencies and input levels that you are thinking about.

-Samir

Dear Samir,
The DDS board that I'm using at the moment is AD9959.
Re: "Also, why do you need a quad amplifier like the THS4524... "
I just imagined a different use of this board. Let's say I have 4 harmonics V(f1), ..,V(f4). 1 channel differential input gives me V(f1)-V(f2), 2nd channel V(f3)-V(f4), and 3rd channel gives the difference of Ch1out-Ch2out, which is V(f1)-V(f2)-V(f3)+V(f4). Clearly +/- sign does not matter. The channel inputs/outputs on this board are 50 Ohms, so no issues with impedance matching.

Dmitri

Hello Dmitri,
I am not familiar with the AD9959. I would look to get a SPICE model and see if the circuit you desire simulates and delivers the correct results.
-Samir

Dear Samir,

I'll try it. Thank you for you time and tips!

Best regards,

Dmitri