Anybody have have 'practical' experience working in 'broadband' microwave? 1 GHz thru 200 GHz, or even up to 300 GHz?

It does -- sort of. The boundary between dielectrics will support evanescent modes or total internal reflection, stuff like that. It's not an exponential decay with depth, because it's not dissipative.

A consequence is, if you have dielectric boundaries, you get mixed TEM modes, which causes dispersion (differing velocities). You're better off with buried stripline than half-open microstrip, for example. But even better still with micro hardline coax and stuff, at least until even that becomes useless (due to losses and such).

Judging by the latest Lecroy video that's going around, micro coax is good for as few inches as you can spare before the converters. They even use waveguides for the upper (diplexed) bands.

Tim

Not much (recent) experience, Ku band spectrometer, in grad school. If the wide band goes near DC then I think it has to be a coax, or something similar.

George H.

Thank you for your reply. I knew miniature was going to be the only way to go.

Any 'obscure' papers or AppNotes out there that cover wide bandwidth 1GHz up through 200 GHz?

Uh, my 'experience' is DC to 18GHz, the requirement is 1GHz though 200 GHz, but the connclusion is the same. tiny.

A few Ham Radio guys build stuff for that band, If you have specific questions the Microwave Mailing List might help.

I'm not going to post the URL, the list is tightly moderated.

Steve

sounds perfect. would like to post questions there.

is alright to send URL to me directly? if difficulty finding on this email...

robert DOT a DOT macy AT gmail DOT com

Beats me. Anything good in BSTJ? Might be some good reference info if you shell out a few $k for various microwave/high speed design tools, but...

Oh, and, multiple modes aren't necessarily a bad thing, but for oscilloscope sorts of purposes, the dispersion and "weirdness" you get will probably preclude most use. So, yeah.

For example, regular old waveguide looks like a bandpass in its design frequency range (low frequencies are quite literally shorted out, since the stuff looks like coax without the "co", or stripline without the strip!), with moderate dispersion in that range (a property of its physics; phase velocity is above the speed of light while group velocity is below).

The band stops for a little while, until the waves fold in on themselves and you get 0,1 TE/TM modes (with different velocities, couplings and bandwidths).

Pretty quickly, the modes begin to overlap, as you go into the 0,2, 1,0 and 1,1 modes, and so on. How close they are depends on proportions of the guide.

Even if you're working an otherwise pretty stable mode (like, I think it's the TE/TM 0,0 modes of cylindrical waveguide comes to mind -- they're symmetrical in nodal pattern and equal in frequency response, but orthogonal -- polarized), small asymmetries in construction can perturb energy between them, leading to power loss (into lossy modes), notches and peaks (interference due to standing waves of different modes superimposing at electrical connections), echos or high Q resonances (energy slowly couples into a low-loss mode, rattles around a while, comes back later), and so on.

Tim

Google WA1MBA and you'll find it pretty quickly.

For mm-wave stuff, spend some time looking at papers from NRAO and their international counterparts. Everybody talks about "DC to daylight" but those guys actually do it.

-- john, KE5FX

Dnon't know what BSTJ is.

ARRRGGG!

Looks like construction geometries are going to have to be VERY small. Which, as pointed out, means lossey, low power.

Bell Systems Technical Journal, .

It was the first hit on Google - too lazy to try?

Thanks, found it, subscribed.

DC to daylight? That's EXACTLY what I need, well almost.

Thanks. I know some people at Alcatel.

Somehow 'searches' and I don't get along. [Even Ms. Macy comments! "What's wrong with you?! Everybody else can find things!] After hours, yes, hours, I usually find worthless stuff only. Actually, for me, as I've been complaining over the past years, google searches end up retrieving garbage. I get unrelated results [where google in their benevolence thinks I 'meant' something else] pages of ads, followed by titles that sound like EXACTLY what I want, then when I click on them, find out the page no longer exists! You'd think google could put in 'bots to check for the 'value' of their searches.

Latest example: I was looking for the extremely cheap ADC/DAC chip for digitizing audio. something like CSVD, or something, control slope variable delay. (from the 70's) The same chip digitizes audio and then can reconstitute it. Google gave me 20 pages of CVS Pharmacy, even when trying to exclude, Never found the chip's part number, but spent over 1 1/2 hours looking.

Worse, after searches and 'test' visiting websites, I have to go through a process of purging my system of 'invisible' garbage. Like a multitude of 'no information' pictures and partially downloaded videos these people deem are necessary to enhance the web experience.

Thus, I always ask for EXACT URLs to cut all that wasted time out.

update: subscribed, accepted, received first day's 'chatter'

first impression, great forum, thanks!

second impression, nice balance between practical and math, exactly the source looking for, thanks.

third impression, their applications may be TOO DIY, but we'll see, will try posting the same question there.

Curious: Perhaps simultaneous or causal, don't know. But AFTER registering for the uWave forum, my spam went from once a month to over sixteen a day. And, this is 'specifically addressed' spam, not 'splatter addressed'

Was this coincidental, or do 'bots go through such forums for the addresses?

Its not the forum, its the path it passes thru to/from the List. I get no spam tied to the uWave list.

Steve