Hi Marco,

?

Your filter is indeed well aligned now. And the ripple in a 6MHz range will be less than 1dB.

?

The fact that the 0Hz peak is at 4.5MHz is really very strange.

The 0 Hz is mainly related to LO-RF leakage, or mixer balance.

If you have -11.5 dBm, I would say wait before changing the diodes, but find the cause why the 0Hz is shifted to 4.5MHz.

The DC offset caused by the mixer will always be 0Hz.

So, if one of the PLLs is incorrect, you will have a frequency shift.

?

For example:

Here is a simulation of this problem with my 8591A.

I replaced the 600 MHz source to the second converter, with my E4432B generator.

The frequency of 600 MHz being multiplied by 3 to give 1800 MHz, an error of 1.5 MHz at 600 MHz, would give me an offset of 4.5 MHz...as shown in the image below. I set my generator to 600 MHz precisely and the center frequency of the 8591 is at 0 Hz. I set my generator to 601.5 MHz and I get an offset of 4.5 MHz.

?

But if the external 10 MHz reference is disconnected, the offset would not be large enough to move 4.5 MHz, there could be up to 10 kHz or so, not counting the REF UNLOCK error message, so it would have to be the first LO…counterlock problem? It’s possible to measure first LO at J5 first converter.

?

Send me a picture of this offset from 0 Hz to 4.5 MHz with and without CAL CORRECT.

?

Yves

?

?

Your 8591E

?