K&S rectangular hobby tubing has been used as waveguide by Hams at 47 Ghz t o build sub-harmonic mixers. It helps if you silver plate it to lower the l oss from skin depth issues. The largest size rectangular tubing that K&S se lls should be a ways below cutoff at the frequencies proposed for 5G.
The SBMS microwave ham radio boasts a file called 47ghzmxr2.pdf that is wor th a look.
A friend of mine got 138 Ghz working using very small diodes soldered to th e end of UT-141 coax with conical horns made of brass sheet. He used subhar monic conversion as well. However he had a 23 Ghz spectrum analyser as a de tector and a tunable microwave source at 22 ghz as the excitation. He achie ved 30 cm range.
It's do-able, but you need a local oscillator to bias the diode into conduc tion.
Me? I don't buy anything. They give me samples. ;-)
The company buys enough that we have one dedicated analog FAE (who moved locally and works out of his house) and something like a third of digital/processor FAE. They just added another analog FAE to the territory (we probably get 1/3 of his time), so business is good. I don't know if the sales guy is dedicated to us but it wouldn't surprise me.
As far as Atmel/Microchip goes, I was an early promoter of their ARM processors and have standardized on them for my piece of the business. My designs also use tons of Microchip opamps.
Of course they aren't the only ones that come through, though I'm a hard sell for LTC (too expensive) and Maxim (you gotta be kidding).
I wonder if they'll still do that. ;-) We do get both the National and several of the TI groups in here selling us stuff. There is a lot of overlap.
Five years ago, I worked at a small company that used TI DSPs and some power stuff. Our volumes were fairly small (a few thousand units a year) but we still got some service from TI. Most of it was through a disty but TI used to fly bigwigs through, mostly because we found some really nasty bugs and they kept trying to placate us (maybe they were afraid of getting their ass sued). The part never worked and AFAIK, the problem was never fixed. I know as long as I was there, the bubblegum would break loose every once in a while and customers would get pissed. Again. I knew how to fix it but was never allowed to spend any time on a real fix, just more patches. It would have meant a board spin, and worse, some BSP changes.
I think the big, x-million-chip, users get dedicated app engineers who might even have access to the chip designers. The little companies get bad data sheets and forums, or support people who can only read us the data sheets.
Interestingly, LTC sends factory people to see us about once a year. They are probably eager to hear my ideas for goofy\\\\\ brilliant new products.
Sure, but what's the resistor value? I'd like to run one external resistor from a 0..+3V DAC into that pin to tweak the gain. Or maybe I need opamps and stuff. The data sheet does say "do not load this pin" so I need a circuit that will drive it without loading it.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
Sure, and they're always asking us what features we want on the chips, too. The "Lunch-N-Learns" are mostly for their product managers to show us their new (and planned) toys to get feedback on our needs. Of course they can't afford this for 1K sorts of Digikey customers.
That's why LTC has such good tools. They can compete in the low volume, though high margin space because they have great support through their tools, rather than expensive people. It'll be interesting to see how this integrates in with ADI's business model.
LTC is around frequently and there are designers who use some of their parts but it's in some pretty specialized applications. Maxim even has its adherents but they burned me with a part about five years ago. I had prototypes built and running when they cancelled the part. The first words had said to by boss was "I told you so".
Diving it sounds a lot like "loading" it, just to something other than
P.S. I'm designing in a TI AFE chip - it has an internal clock for the auto zero or external clock, which can be a crystal or a digital input. if you use the digital input, nowhere does it say what to do with the other pin from the crystal oscillator.
Upon request, they say I can either leave it open or ground it, which seems.. odd. I guess it's an input connected by a few M to some internal node that would have been the crystal drive, but would like to see what's actually on the other side of the wire bonds.
--sp
--
Best regards,
Spehro Pefhany
Amazon link for AoE 3rd Edition: http://tinyurl.com/ntrpwu8
That's the usual situation with MCU clocks, but not on this chip.
The pin is listed as an input *only* on the datasheet, which is why I was concerned. The other pin- the crystal drive (clock output) can be configured as an input using an internal register and that is where you feed the clock signal in.
The input must have a weakpullup or maybe they leave the Pierce oscillator bias resistor in place, who knows?
--sp
--
Best regards,
Spehro Pefhany
Amazon link for AoE 3rd Edition: http://tinyurl.com/ntrpwu8
That situation sounds like the standard weak-inverter semi-reliable crystal oscillator. The unused pin is the inverter output, and you drive the input.
Grounding it is a little weird.
Chip schematics (dream on!) or at least block diagrams sure help.
--
John Larkin Highland Technology, Inc
lunatic fringe electronics
My guess(tm) is that WR28 would be the most common for 5G frequencies:
0.280 x 0.140 inches. Yeah, that's tiny. I've been doing silver plating and using electroless silver at home for quite a while. However most everything I've done is at lower frequencies (L,C,X bands and wi-fi).
Looks like he uses round tubing for waveguide instead of tubular. Looks nice, but I would feel better if there were some photos of the actual mixer and some numerical test results. Hmmm... dated 2003.
Nicely done. I couldn't do that with my pile of antique pile of test equipment. My HP141T/HP8555a spectrum analyzer will go to 18GHz. If I purchase the external mixers, it will go to 43GHz. (Notice the burnout limit = 1 milliwatt). Actually, I do have some suitable mixers, but the diodes are all fried. Yet another project.
My highest frequency source goes to only 4GHz, which rather limits what I can do. 138GHz is currently out of my range.
Yep. That's the problem with upconverters and downconverters. One needs to have a local oscillator source to feed the mixer.
--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
With such applications put the local oscillator in the middle of the band of interest and feed it to simple mixer which folds the upper and lower bands to the same baseband frequency. Feed the mixer output to a broadband 0 - 2.5 GHz signal indicator. This will cover 5 GHz (Flo+/-2.5 GHz) of the band. Folding the upper and lower bands doesn't hurt, since we are only interested in total RF power.
To generate 28 GHz LO signal use some oscillator (e.g. DRO) in 9-10 GHz range, overdrive it in a MMIC stage and filter out the 3rd harmonic to feed the mixer
Yep, that would work. The LO and mixer are a bit complicated due to the 30GHz frequency, but the IF amp and detector can be quite crude. I've done something like that using a cheap satellite TV signal meter: Apply 12VDC to the output port through a 75 ohm load resistor to power the meter. The input port goes to the mixer. Inside the meter is a rather crude MMIC 950-1950MHz amplifier and detector to drive the meter. Most of the gain is at the 950MHz end.
If you feel link playing, put a cellular yagi antenna on the input port of the meter and use it to find cell sites operating in the
850-950MHz region. I was doing that for a while until I build something better. It doesn't work too well at PCS frequencies
1900MHz.
The problem with such schemes is that it's difficult to tell what the detector is hearing. It could be some other service on adjacent frequencies. For example, it could be responding to 24GHz speed trap radar. With a simple mixer, minimizing the LO radiation out the antenna might be a problem. Besides, the OP wanted "cheap" which to me means a horn antenna, RF bandpass filter, diode detector, and DC amplifier. In other words, a 30GHz field strength meter.
--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
The classical 1N21 mixer diode might be usable at those frequencies, but the noise figure would be quite bad, but who cares, if you are only interested in the envelope power.
One should remember that microwave cellular communications doesn't benefit from line of sight point to point communications. You are fully dependent on in-room reflections as i the IR remote case.
Today, these are speed radars and satellite services. In the future, there are other services too.
A horn might have too much "gain" or directivity for usable measurements.
Waveguide filters on the RF as well as on the LO ports will keep most of the low frequency (below 20 GHz) crud away, while any post mixer filtering will keep away any interference far away from the LO.
Older US KA band radar detectors with the stepped ridge waveguide horn are quite broadband, especially the ones that have fixed 11 Ghz LOs. In the n ear field, a few meters of range, they should quite fine for this. The on es that try to identify the band or emitter by sweeping the LO and IF may not be good candidates for this task.
Ebay is full of the things, but the trick is identifying the internal topol ogy. There is a class of older radar detectors with little input filtering that will sound off just fine in this band. Sadly they are the ones that were ea sily detected with counter detector devices in some states... Thus they wer e only on the market a few years, maybe five.
They're all over eBay and Amazon in a few configurations. Search for "satellite meter" or "satellite finder".
I guess a warning might be useful. I've own about 10 of these meters. They're quite useful, quite cheap, but not really quality instruments. The interior construction and solder are abysmal. This looks like some of the better meters: Some meters were unstable and would go into oscillation for no obvious reason. Others are seriously lacking in gain. One was dead on arrival but fixed by resoldering a few ugly connections. When I need one to aim a dish, I usually grab 4 or more meters just in case one fails. I had one installation where 3 out 4 sat finders didn't work quite right. I can also testify that dropping a satellite finder onto concrete from about 20ft will cause it to self-disassemble. Caveat Emptor.
--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
No way. At 30GHz, the cat whisker inside the 1N21 or 1N23 diode makes a dandy inductor. At 30GHz lead length is critical and must be minimized. Microwave diodes come in much much much smaller packages.
Really? I've done quite a bit working with 802.11n systems. I've found that wireless routers spend most of their working lives at
802.11g speeds and modes, without the benefit of 802.11n MIMO features. The only time 802.11n speeds are achieved is when I either force the router to use them by disabling 802.11b/g, or when I have a nearly ideal situation that includes both a direct line of sight an a highly reflective environment.
802.11ac did it right. I can force an 802.11ac router to only use
802.11ac and it will slow all the way down to a 6.5Mbit/sec rate. No need to switch to 802.11a rates. Anyway, 802.11ac acted the same as
802.11n on 2.4GHz. Anything that increases the BER (bit error rate) will cause it to slow down, switch modes, drop out of 802.11n MIMO, and only use the direct path.
That depends on the shape and size of the horn.
Rectangular WR28 waveguide would be 0.14 x 0.28 in (3.5 x 7.0 mm) and is a 2:1 rectangle. My guess(tm) a 3 x 6 cm horn would be about right. Plugging into a handy horn antenna guestimator: I get: Gain = 22.6dBi Beamwidth vertical = 17 degrees Beamwidth horizontal = 11.7 No problem. That should work quite nicely and is much better than a dish, which ended up with a too narrow beamwidth of less than 1 degree. Construction of a horn is also much easier.
--
Jeff Liebermann jeffl@cruzio.com
150 Felker St #D http://www.LearnByDestroying.com
Santa Cruz CA 95060 http://802.11junk.com
Skype: JeffLiebermann AE6KS 831-336-2558
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.