If it's a high sided LO, that mean the output frequency is being swept backwards too. So the GD (derivative of phase wrt. freq) still comes out positive.

R&S has a two-tone GD measurement technique. I have never used it, but they say it can even be used to measure satellite links. In this they talk about using a reference mixer fed from the same LO as your mixer, bringing it back up to RF, and I assume that would take care of the frequency inversion as well.

https://cdn.rohde-schwarz.com/pws/dl_downloads/dl_application/application_notes/1ef98/1EF98_1e.pdf

I can tell you what not use use; MS OneDrive. It would work great IF they added a .gitignore type feature to prevent syncing of files or folders. Despite requests from thousands of users, they have not implemented it. One Drive works well for backing up work, but say you have a project folder with GB of 3D EM data cache, it will attempt to back that up. So what I'm having to do is keep my working projects outside of "My Documents", then remember to copy them in periodically for backup. That defeats the purpose of having a cloud backup.

Well I'll pay you $100M to design a 4 way splitter with under 5 dB insertion loss.

I use MS Word, just copy and paste stuff from Microwave Office, and paste in photos from a pocket camera. I use the Testwave software (within MWO) heavily in the lab. Each project has it's own document; nothing shared like a lab notebook. This is more for my own purposes than any patent stuff, because invariable I'll have to dig up something a year later, and I can't remember how I did it. It's easy to give the report to customers too.

I figured out where that loop fixture is from. The oscillator class by Jeremy Everard and Simon Bale. It is shown in the recent MTTS webinar. I'll have to update the video.

Video 7 details a coupled loop fixture designed to measure the quality factor (Q) of surface mount (SMT) resonators to 1 GHz.

The fixture has adjustable loop spacing to null the magnetic coupling between the loops. The design and analysis is performed in Cadence AWR Microwave Office using the Axiem and Analyst electromagnetic simulators.

The initial design has poor high frequency isolation due to common mode currents on the interconnects. The design is revised using EM simulation to show additional isolation may be achieved with the addition of ferrite loaded baluns.

One reason would be reliability. The smaller the connector, the more likely it is damaged, so if you don’t need to go to 67 GHz, keep it at 50 GHz. Less expensive cal kit too.

I have a 67 GHz PNA at work where I ordered 1.85 mm NMD to 3.5 mm cables, and a 50 GHz PNA with 2.4 mm NMD to 3.5 mm. Of course I have full 1.85 mm and 2.4 mm cables, but only hook those up when I need the extended frequency. Otherwise keep everything at 3.5 mm.

Years ago some idiot at work force fit a 3.5 mm onto a 2.4 mm brand new spectrum analyzer. The threads are not even compatible (for that reason), but a monkey with a wrench can pull it off.

I’ve used through-hole, vertical mount GPO at X-band. In that case, I just calibrated at 3.5 mm and used the adapter. I then tweaked the matching network on the PCB for good return loss. I was looking back into an isolator, so didn’t need a load.

If you have enough distance (on the PCB) between the GPO and the next discontinuity, maybe use time domain mode to measure the GPO, or gate it out.

If you know the GPO has good return loss (which you can do with time domain), and can get a short circuit established after it, then use port extension to get a phase corrected measurement. Though YMMV with port extension; I put a marker at the frequency I’m most interested in, then rotate the delay for an exact short at the marker. The phase typically diverges above and below the marker frequency, but at least I know the phase is correct at that one frequency.

GP3O and even WFL are “good” through 60 GHz, but all that means is they don’t mode. The WFL has piss poor return loss even at 2 GHz.

Start with tightening it by hand. You turn the large nut on the equipment (it’s a 2.4 mm NMD connector) while keeping the 1.85 mm adapter stable. Once it is finger tight, hold the 1.86 mm with a wrench and use a torque wrench on the NMD.

Video 4 shows how to use FlatCAM to process the Inverted F Antenna (IFA) gerbers into G-Code, which is then used on the PCB milling machine.



Video 5 shows how to tune the IFA using a VNA, and discusses use of a calibration kit.



Video 6 shows testing of the antenna by measuring C/N0 using a GNSS receiver module.

just need to read the intro documentation on VSS. It's very powerful, but need to get your head wrapped around the whole analytic signal method and try some experiments. It makes good sense since you separate out the carrier and envelope. Though VSS is more than time-domaim, real and analytic signals. The cascade budget analysis and RF inspector are great. With the latter, you can pinpoint the sources of all the intermods through multiple conversion chains.

Yes, ADS, but I'll tell you in terms of Microwave Office, which is what I use. Much more integrated than ADS.

I have done resonator measurements and simulation such as these, but typically you would use a two port (S21) measurement with two loop probes, as you will get much more dynamic range than a reflection measurement. You need very weak coupling to measure the unloaded Q of the resonator, and you probably won't get that with a 1 port measurement, and get an accurate measurement..

Your loop probes must be electrically small and operating below their self resonance, which it looks as if they are, given the geometry.

You haven't stated the boundary conditions; open or closed. Your say it is in a box, but I assume you mean embedded in a brick of dielectric, which has the sides open to free space, rather than in a conducting box.

So you can simulate it in 2.5D, which for MWO would be Axiem (the ADS equivalent is Momentum). You are on an infinite substrate, of multiple dielectric layers, with top and bottom boundaries. The boundaries can be open (sheet resistance of 377 Ohm/square), or conducting. You need to get the open boundaries 1/4 wavelength away (i.e. out of the reactive near field) from you circuitry. This looks to be a UHF resonator, so that can be quite far away, compared to the thin layer between the parallel plate capacitor. This is drastic size difference is not an issue for 2.5D simulator since you only mesh the conductors, not the volume, which is why you start there before moving to 3D. If you need conducting side boundaries, surround it with vias, but again, I assume this is open boundary. You would use differential ports (+/-) to drive the probe, or a gap port.

Another 2.5D in MWO is EMSight; ADS does not have an equivalent. Now you are in a box with conducting sidewalls, and can still set the top and bottom boundaries. Sonnet is very similar. The advantage is you get much more dynamic range with the closed boundary, but you are in a box. The mesh is also on a grid, whereas open boundary is gridless, though sonnet has some conformal meshing tricks.

For 3D, the MWO simulator is Analyst. The ADS equivalent is EMPro (I think). Now you can set arbitrary boundary conditions; open & short (E and H), and PML. You still need to get 1/4 wavelength away for open and PML, but not restricted to an infinite substrate. Though now you need to mesh the volume, including tiny volume between the capacitor plates.

If you are in a conducting box, you can use eigenmode analysis with Analyst, dispensing with the driven probes to solve the resonant modes within the box.

CST has a characteristic mode analysis, which I believe is an open boundary eigenmode analysis.

All of these programs support DXF import, though you ought to be re-drawing it in the native environment, so you can parameterize it.

It's not an issue of accuracy, rather application. The cos(wt+phi) is a real function while the e^j*(wt+phi) is an analytic function, or phasor (typically used in RF texts with the wt (time varying term) dropped.

https://en.wikipedia.org/wiki/Phasor

In general, I'd read up on analytic signals and get used to thinking in those terms, as they have practical uses in I/Q conversion, SDR (e.g. GNU Radio), and RF system simulation (e.g.) AWR VSS. It's really a method of representing signals by separating out the carrier and envelope (i.e. the baseband, modulation, information).

I'd also recommend Microwave Office. I've been using it for 20 years for board level RF design (antennas, PA) and system level stuff using the VSS product. The base product is linear simulation and layout, then you'd probably want harmonic balance and Axiem for planar EM. I have been using Analyst a lot for antennas, though unless you are doing 3D structures, you can probably just use Axiem if everything is on a PCB.

I know for high frequencies, it’s a loss issue. It order to stay TEM for coax, you have too keep reducing the diameter, then the loss increases substantially. Using coax past 26 GHz for more than a few meters length is problematic.

Waveguide has much lower loss per unit length than the equivalent TEM coax, but now dispersion becomes the issue.

So back in school, my Microwave professor said Bell had manufactured miles of circular waveguide for use in telecom, which was obsolete overnight by fiber. There was a certain circular mode that has very low dispersion, compared to rectangular.

This looks like a good discussion of the elliptical waveguide.

http://www2.rfsworld.com/RFS_Edition4/pdfs/FLEXWELL_Elliptical_Waveguide_347-369.pdf

Thanks hagster. Our purchasing dept was able to make contact with them to get a support quote.

I have the iFilter for MWO, but I haven't done any filters in years. The advantage of the integrated packages is that they can spit out schematics, with pre-populated tolerances, into the native simulator.

Thanks for the confirmation. Guess I better make double backups of the instrument control software.

Someone replied on Twitter that the entire engineering team was bought out. I've had one of their near-field antenna scanners for several years, and last year bought the EMI scanner. They made good stuff.