harmonic reject mixer

I am working on harmonics reject mixer, down-convertion, working from 150kHz to 30MHz, using 6 of 8 LOs(pi/4 from each other, we use 0, 45°, 90°, 180°, 225°, and 270° LO) and us 0.7 as the ratio close to 1/√2 (0.707) with 1% error,

I used pss+pxf simulation, when the LO freq is 1M, the harmonic rejection ratio are: H1/H3=55.6dB, H1/H5=60dB, (match well with the hand calculation),

But when the LO is set to 30MHz, the ratio become H1/H3=48dB, H1/H5=49dB, the LO is ideal LO(50% duty cycle, rise and falling edge is 0.01T), from the equation of the harmonics reject it should be frequency independent, but it do decrease with the frequency, I don't know why.

Can anybody who has the similar expirence in Harmonics reject mixer design tell me why or give me some idea about it? thanks a lot.

I never designed a harmonic-rejection mixer, but I know that its performance is very sensitive to component matching and delay accuracy (the same as in image-reject mixers).
You are on simulation stage right now, but I think manufacturing a high volume circuit which tightly controls the phase matching over wideband frequencies will be challenge.
Maybe it’s easier just to place a filter after the mixer :D

Thanks vfone!
The problem is the LO I used is ideal LO, and the performance is monotony with the frequency, and the peak value is shown at 1~2MHz, for frequency lower than 1MHz, the HR ratio goes down again, I think is it because the new induced phase error by the parasitic capacitance or somethink like that,

Thank you for your help!

For sure the reason is that you mentioned.