Generating a 100GHz signal - how you'd do it

Dear microwave friends,

I need to generate a PLL locked signal at around 100GHz, power between 10-100mW. PLL comparison frequency, freq step, bandwidth, etc... are all free parameters for now. The cost will surely go beyond $10.000, so let's forget the budget and concentrate on the technical stuff for now.

I don't want to reveal my idea for now, so I'd be very grateful to hear about your ideas and how would you do it.

Another big issue is which manufacturers produce such components in the 20-100GHz range? Packaged devices and connectorized modules are preferable over chips. If we would compile a manufacturers list that would be great (oscillators, amplifiers, multipliers, dividers, etc...).

Any input is greatly appreciated!

Thanks,
rfmw

Just off the top of my head, I would use a vco (fet, gunn, etc) at 100 GHz, split off some output power and send it to a harmonic mixer. I would drive the mixer at maybe 10 Ghz, and use the IF output (after amplifying) to drive a standard PLL chip.

Spacek Labs, Inc. "www.spaceklabs.com" offers two solutions for <10 k$:
a) a Gunn oscillator locked to a harmonic of a PLO source, e.g. at 25 GHz. This PLO is locked to a 100 MHz OCXO reference. Output power ~+10 dBm.
b) A 25 GHz PLO as above, with an active/passive frequency multiplier. RF output is 0...+6 dBm with a passive doubler, or, +10 dBm with an active doubler.

Thank you both!

Good idea, Biff, worth of investigating some more. Jiripolivka, my idea is more or less same as yours. Good that you've mentioned the company, I would have overlooked it for sure.

My plan is/was to use the Gunn (HXI) at 50GHz or so, split the power for a divider (SHF) for the PLL and take the rest to a powerful amp (NEL) and multiply it (VDI) to 100.

If I do the project, I would select a Hittite VCO :Hittite Microwave Corporation - Products

Then freq Multiplies to about 20~35GHz,Hittite Microwave Corporation - Products

Then freq multiplies to about 100GHz:https://www.cernex.com/PRODUCTS_NEW/Multiplier-PDF/Active%20Multiplier.pdf

At 110GHz, the power output is about +4dBm.

Dear RFMW,

your idea is correct, I only do not know if a divider exists for the 50 GHz signal. I have tested some frequency dividers, the highest from IFR worked up to ~26 GHz only.
This is also why we use the 25 GHz PLO which is commercially available.

Tony, thanks for the Cernex, I've never heard of them, but they have very useful devices. I know Hittite well, but unfortunately they cover more or less only lower frequencies (for this project).

Jiri, see shf.de, they have a 60GHz divider by 2. I agree that's a good idea not to rely on a single manufacturer with a very specific piece of hardware I guess your idea to start at 25GHz make more sense.

Dear RFMW,

thank you for the SHF.DE, I did not know about them!
We do not use the dividers so often, multipliers and mainly mixers is what we do.

Getting some power at 100 GHz is still difficult and expensive. If you need to design a LO source for a mixer receiver, we make DC-biased mixers which can operate with 0...+3 dBm LO input while meeting all specifications.

Hello Jiri,

I am also designing mixers but at 30GHz. can you give some idea about the DC biased mixer which uses 0-3 dBm LO power. What is the configuration of the mixer? Is it waveguide based or planar?

Regards,
akr

Dear akr,

our company, Spacek Labs. Inc,. has two mixer versions for Ka-band (30 GHz), and we can modify any for DC bias. With DC bias you can operate such mixer like a "standard" mixer, but with LO input of 0 to +3 dBm. Also all specifications are met except of high-power linearity and the LO/RF isolation can be set for a narrow band of LO frequency.
Please see "www.spaceklabs.com", you can see all specifications and outlines of our mixers.
I would welcome your requirements to be sent to "jpolivka@spaceklabs.com", we can make you a quote. Orders welcome!

Thank you, best regards, Jiri Polivka