Replacement for AD630 for the purpose of Lock-in Amplification

I am making a lock-in amplifier which works in the megahertz range(500MHz to be exact). AD630 does not work in this range. So can anyone please suggest some IC which does exactly what AD630 does but at a higher frequency range? Or do I have to use some other approach entirely? I am also attaching a circuit which I am working with (I got it from Instructables).

John. I can't see you beeing able to use this circuit without considerable redesign, the AD620 and 630 are for use in the kHz range, not MHz.I am not an RF guy (and at 500 MHz you should have posted this in the RF category) but a balanced mixer I have seen that works up to 1.5 GHz is the TFM-2 from Mini-Circuits:https://www.minicircuits.com/WebStore/dashboard.html?model=TFM-2%2B
Just a double-balanced modulator, 2 transformers and 4 diodes, no amplification as in the AD630, but at 500 MHz you can't be picky....I'd suggest you re-post in the RF category if you want those with more expertise than myself to look at this.

Mini-Circuits also makes amplifiers for 500 MHz signals.The Mini-Circuits passive (2 transformers and 4 diodes) double-balanced mixers DBM's have three ports: RF, LO, IF. All three ports expect 50-ohm source or load. Look at the schematic of the double-balanced mixer. You apply your small 500-MHz signal to the RF port. You apply your large local-oscillator signal to the LO port. The synchronous-demodulator output appears on the IF port. The IF port is DC-coupled, though it may have a few millivolts of DC offset due to mismatch among the 4 diodes and due to asymmetry of the LO signal.Analog Devices does make DBM's for use at 500 MHz. Study their datasheets for DBM's and for IQ (In-phase and Quadrature) mixers.
The DC offset of the DBM output limits the small-signal effectiveness of the lock-in amplifier. Gain before the DBM reduces this problem but also reduces the large-signal dynamic range.You can eliminate the DC offset by making your LO generate both I and Q LO 10 kHz away from your signal, using two mixers to give you I and Q components at 10 kHz. Run these into a stereo audio A/D converter (ADC), and do the final synchronous demodulation in software, where DC offset is tiny numerical round-off error.
You could also run your 500-MHz signal into an RF ADC running at 2 GSPS, and do the synchronous demodulation in an FPGA.The oldest reference I know of to IQ demodulation to a complex signal in the modern sense, is D. K. Weaver's paper, "A Third Method of Generation and Detection of Single-Sideband Signals", pages 1703 to 1705  in the December 1956 issue of the Proceedings of the IRE (now the IEEE). Modern DSP'ers will find Weaver's method remarkably familiar.