Digital Clock Generator IC used as Analog Mixer LO input?
I need an analog frequency for mixer LO as 600-1200Mhz. The ultimate frequency of DDS is about 400MHz (AD9910 by ADI), but the CDCE62005 (by TI) ranges from 4.25MHz to 1.175GHz. so I pan to adopt the AD9910 to synthesize a low frequency (about 40Mhz), and applied to the input of CDCE62005, by programming the CDCE62005, the output can achieve the required frequency.
the question is, CDCE62005 is a digital clock IC, does it fit for analog mixer lo sources?
Thanks to anybody.
Yan
Any analog mixer in principle functions as a DPDT switch, alternatively sending LO current here and there through a pair or a quad of diodes. By doing that, adding a RF signal generates two or more sidebands by this nonlinear multiplication action.
With any analog mixer you can use any signal source generating a sinusoidal or rectangular LO signal. Then such mixer will generate linear combination of LO and RF or LO and IF frequencies. The remaining task is to choose one or more resulting frequencies or bands and use a band-pass plus band-reject filters to obtain a desired spectrum at output.
Using a "digital" LO source often means that not a sinusoidal signal is used as the LO in an analog mixer- then it may be sometimes more demanding to apply analog filters to generate a clean spectrum output.
I don't think it's a good idea to use a digital IC as a analog LO source.
The digital signal has many harmonic. Such as for 600M, the 2nd harmonic is 1200M. To get analog signal from digital signal, normally use Low pass filter.For 600M digital signal, you must filter 1200M, 1800M, etc.
But 1200M is your passband, so you can't filter it.
You can use some wideband VCO, such as Mini-Circuits ZX95-1300+, 400M~1300M, output power is +8dBm, and harmonic is more than 12dBc. You can check the phase noise specification whether meeting your requirement or not. And you can find a amplifier to amplify the LO signal if power is not enough for mixer.
The question is better stated this way: does the mixer in question prefer a square-wave like LO waveform or a sinusoidal LO waveform? The answer is that it depends on the mixer. The vast majority of diode mixers DO NOT like square-wave inputs. This includes most double and triple balanced varieties.
The only diode mixer that clearly PREFERS a square wave LO drive is a T3 mixer from Marki Microwave. This is a proprietary design that acts like a classical "commutating mixer". As Jiri correctly states above, mixers are best thought of as switches (aka commutators). Thus, the switching action tends to be best when it is instantaneous. Curiously, the majority of mixers do not benefit from the fast switch time of the square LO, this can be partially attributed to the finite bandwidth of the balun structure. A T3 mixer, for example, can go MHz to many GHz, give the possibility of yielding extremely fast rise time (i.e. many odd harmonics of the LO).
Marki Microwave has written up some of these results here:
http://www.markimicrowave.com/menus/.../t3_primer.pdf
If you do not care about the performance details, you can drive the mixer with either a square wave or sine wave. As pointed out, some mixer prefer a square wave for spurious intermods, while other mixers dread such an LO drive! You would have to measure the specific mixer carefully and see which is better for your particular system requirements.
There is no question that if you use a square wave device to generate an LO, it would be a very simple matter (3 pole lowpass) to filter it to look almost like a sine wave. You might have to reamplify, since the LO drive level in dBm might not be met.
Thanks to everybody.
The mixer used here is balanced mixer (consisting of four diode).