problem with a high IF mixer design

[danding]

Hi
I am designing a V band singly balanced mixer, which the IF bandwidth is close to the LO bandwidth. Should I use bandpass filters both at the LO port and IF port? Or is there any good method to solve this problem? In addition, I want to design the passive diode mixer using waveguide-to-stripline transition. If anybody can suggest something useful, I will be grateful.
Thanks.

 

[tony_lth]

I think the bandpass filter is needed for LO and IF.

 

[danding]

Thx, tony_lth. What if I use a broadband coupler, which the input ports are defined as LO and IF?

---------- Post added at 16:55 ---------- Previous post was at 16:52 ----------

I think the bandpass filter is needed for LO and IF.

Thx, tony_lth. What if I use a broadband coupler, which the input ports are defined as LO and IF ports?

 

[jiripolivka]

I do not think you can design a good mixer for wideband use with one diode only.
One mixer diode cannot be matched over a wide band in principle.
Please look at "www.spaceklabs.com"; Spacek Labs. has a wide variety of mm-wave wideband balanced mixers . They combine waveguide, coaxial and stripline for the best results. Our M-180 series even offers a full RF and LO operation over 18 to 40 GHz. IF is typically DC to 15 GHz wide but can be expanded to 23 GHz.
If you do not want to use a verified technology, I would rather recommend to find a MMIC mixer; Hittite possibly has a suitable mixer.

 

[danding]

I do not think you can design a good mixer for wideband use with one diode only.
One mixer diode cannot be matched over a wide band in principle.
Please look at "www.spaceklabs.com"; Spacek Labs. has a wide variety of mm-wave wideband balanced mixers . They combine waveguide, coaxial and stripline for the best results. Our M-180 series even offers a full RF and LO operation over 18 to 40 GHz. IF is typically DC to 15 GHz wide but can be expanded to 23 GHz.
If you do not want to use a verified technology, I would rather recommend to find a MMIC mixer; Hittite possibly has a suitable mixer.

Thank u very much for your advice, jiripolivaka. I've found beautiful mixers in the web site you recommended. However, I have to design the mixer I need by myself.
And now I want to design the singly-balanced mixer with a branch-line coupler having two sections for high IF bandwidth and LO， because they two are very close, but the problem is that isolations between them are poor. The attachment is a schemic circuit of the singly balanced mixer I designed in ADS.
In addition, any other good method such as crossbar singly balanced mixer to design the mixer I need?

 

[jiripolivka]

Dear Danding:
I think your schematic is correct. But- I never used models like this, so I cannot offer you an opinion or advice.
Mixers we use were patented in ~1972 and since then up to now no other design created a better solution. A crossbar mixer can work but is possibly not wideband enough for your application.
Our mixers are a combination of waveguide and stripline, IF circuit is coaxial. This way they can offer as wide bandwidth as a particular waveguide is while stripline is a TEM line, wide from DC to infinity.
Maybe you can check MMIC mixers by Hittite, they are also good but not as good as Spacek models.

 

[tony_lth]

you can try this schematics.