TNP #11 - Teardown & Analysis of a 75-110GHz (W-Band) Waveguide mm-Wave Balanced Mixer Module

[@Thesignalpath]

*Minor Correction: Upon closer inspection it appears that all ports (RF/LO/IF) are shared on the center common port of the diodes. The top and bottom connections to the chassis do not carry any RF signal. The mixer is balanced only in that it uses two diodes. No differential stimulus is created.*

[@luc7luc7]

It is truly a balanced mixer. in fact it is a single balanced mixer. Differential stimulus is seen from the perspective of the diodes. RF can be seen as being applied differential to the diode path to IF since one diode is referenced to the bottom wall of the waveguide while the top diode is referenced to the top wall of the waveguide the latter will see the RF signal 180deg reversed relative to the first. Diodes are in the same direction and driven at the center by the LO therefore each diode conducts on each half of the LO cycle, this way the bottom diode is ON and connects the +RF signal (note that "+" is because it is referenced to the lower side of the WG) to the IF port. When the upper diode conducts connects the -RF signal (note that "-" is because it is referenced to the top of the waveguide) to the IF. Truly balanced action results. The IF is the result of switching, at the LO pace, between +RF and -RF into the IF port.

[@Firerexify]

Whenever I'm feeling blue, I just watch these videos and immediately cheer up when I hear you call what I do every day for a living *Magic* THANKS!! 😊

[@MachiningandMicrowaves]

I can appreciate the machining work in this unit, I've been making 122 GHz parts with WR-06 guides. Very interesting to see the connection of the opposed pair of diodes and the tapered end of the 2.92 connector pin. To get tapered-height guide matching into the diodes, I'm using a 0.5 mm diameter end mill on a rotary table. Doing this sort of machining in a home shop is certainly an amusing challenge. Once I get it to work, I'll have to do a video on the process. Thanks very much for this, it has saved me a lot of trouble trying to work out how to make the diode connection.

[@LazyTurtle1988]

Every time I see a waveguided instrument I feel being scammed :) but they are really cool and magical at the same time.

[@antiikadad917]

RF electronics is not a scam :) it is real and works exceptionally well. Almost everything in a communication satellite is waveguided electronics due to high frequencies.

[@SirMo]

Also the equipment required just to be able to test and characterize the mixer is also probably astronomical.

[@huanhosedubo]

Signals paths in these frequencies behive like mirrors and light!

[@Electronics-Rocks]

Love the black MN magic of wave guides / microwave

[@HeyBirt]

The real trick to machining parts like this is the fixturing. You need a way to keep the part(s) aligned to a known reference and securely held. On this particular piece some of the machining operations were done to the individual parts, then these parts had to be aligned and fixed in position, alignment pin holes drilled and then other operations like drilling/tapping the split threaded holes. I part I have spent days designing and machining fixturing to make one 'simple' small part like this which had machining operations on 5 sides. Complexity of design, manufacturing and low volume equal high price...

[@gregorymccoy6797]

I would love to see you do a white board demo of wave propagation fundamentals that us low-freq people don't normally understand or encounter. I am fascinated by high freq devices and the magic coupling and filters one sees that are dictated by cavity dimensions.

[@todayonthebench]

That there seems like a nice mixer. Surely makes my own hand soldered diode ring mixers seem crude in comparison. Though, I so far only went up to the really low GHz...

[@ikocheratcr]

Does the metal it is made off have any effect on the operation of this device? I guess it is Au plated, but does the base metal helps in any way?

[@jayzinho10]

Please do a video on waveguide-reflection type amplifiers.

[@scratchpad7954]

Seeing this kind of device leaves me to wonder what the applications are of EHF radio in the W band at 75-110 GHz.

[@paolomonai9511]

Good question..basically applications span from receivers for astronomy to seekers (autonomous radar guided missiles) and so military equipment. Regards from Italy.

[@stargazer7644]

The adaptive cruise control and collision avoidance/auto brake system on my car operates at 75 and 122 GHz. Water and oxygen absorption in the atmosphere is high especially at 122 GHz, so long range applications are not common.

[@paolomonai9511]

@@stargazer7644 You are right! I forgot to mention those commercial applications!! Regards.

[@zebo-the-fat]

@@stargazer7644 That may explain why the adaptive cruise control in my car fails in heavy rain!

[@johnnycash4034]

This is amazing. Thanks. Can you do more of these please? Where can one learn about these things?

[@alejandrosilva7520]

Incredible, thanks!

[@johnyoungquist6540]

I'll guess you couldn't design such a thing without precise modelling. I wonder how many prototypes they had to build to get it right. Looks like a lot of precision machining and maybe EDM. (electrical discharge machining erodes with sparks) Some tool marks are visible on the flat surfaces. Simple in principal but fascinating. Great video.

[@stargazer7644]

It looks like a standard milling machine job to me. Nothing fancy, just precise.

[@gabiold]

@@stargazer7644 Nah, that two-dimensionally conical waveguide channel is not that of a simple milling. Definitely not with manual work, but probably not that everyday work with a CNC either.

[@robertw1871]

It’s amazing that 10 dollars worth of material can be fabricated into something worth so much.

[@metatechnologist]

My guess it's military sponsored.

[@saintfather7757]

Its gold plated

[@yoctoflop]

See: CPUs and other specialized ICs

[@benmodel5745]

Definitely a ton of value added

[@Leonelf0]

It's also partly because of the machining methods used. This is definitely made using EDM, so it's slow to manufacture and has a small market niche. In my job we ordered two sets of a harmonic mixer and a frequency multiplier to reach 300GHz, combined worth a quarter million apparently..

[@phrtza]

Fascinating! Subscribed :)

[@thekaduu]

I imagine they would have to wire EDM the wave guide parts as opposed to CNC machining. I can see machining marks on the outside which does not need to be perfect.

[@thekaduu]

It looks like the inside of the part made in Japan and shipped to China to "finish" the outside :P

[@Thomas_Lemmey]

It is just machined with an endmill, at 7:40 you can see the round marks in the bottom of the waveguide from the cutter.

[@robertw1871]

EDM won’t work on that shape, so it’s got to be an ultra precision probably extremely high speed (100,000 RPM) milling machine…. I’d love to see the machine, I’d imagine the tolerance here has to be in the micron level probably 10 millionths maybe less? Whatever the process I’m sure it’s expensive…

[@robertdixon8238]

Thanks. Can you draw a simple equivalent circuit?

[@paolomonai9511]

@@hGUImain You are right. This is the most effective building techique for waveguide components. Commonly used also for filters (E plane type). Regards from Italy.

[@gouthamsh1]

Why does the waveguide taper near the diodes? Isn't it supposed to follow WR-10 dimensions throughout the signal path?

[@gouthamsh1]

I mean... A shrinking waveguide may act as high pass filter because the waveguide cutoff frequency increases in that section. Or does it?

[@user-ig8pp3st4v]

Shrinking along the y direction (the shorter dimension of the WR10) does not affect the cutoff frequency, and it pbb helps further localize the field.

[@vaibhavzadokar7221]

So cool!!

[@RixtronixLAB]

Nice video, keep it up, thanks :)

[@djmips]

If not repaired by a talented person like yourself, could it be repaired by the original company? If the metal waveguides were fine then shouldn't the be able to apply a new diode with whatever technique they use to build a new one?

[@paolomonai9511]

Not a so easy task...Diodes are surely high barrier Schottky diodes, prefearably a matched pair to have a perfect balance. They are thermally bonded onto the metal body. So you need a bonding machine, a very expensive equipment that is commonly used for wire bonding chips to substrates. Pure microelectronics...regards from Italy.

[@stargazer7644]

Of course, anyone who could make one has the equipment to repair one. Assuming replacement diodes could be sourced.

[@ljubomirculibrk4097]

Prices of high RF parts in US is just nuts.

[@AgentOffice]

Can you teach what the diode does

[@tcarney57]

You can look up "balanced diode mixer," but in short the diodes switch on and off at the LO frequency, chopping (a.k.a., "commutating") the RF signal. Mathematically, this multiplies the two signals together, resulting in the sum and difference of the two frequencies out the IF port. You usually only want one of them (the sum for up-conversion or the difference for down-conversion), so you either filter out the mixing products you don't want (the harmonics are mixed as well), or you ignore them if doing so doesn't cause problems. A _double_ -balanced diode ring mixer uses four diodes and it's used to get better isolation between the ports. If you don't need maximum isolation, two diodes are used, especially--as in this case--the mixer is already costly.

[@AgentOffice]

@@tcarney57 wow a simple component doing some crazy physics

[@xDevscom_EE]

EIP mixer?

[@SirMo]

Could you repair it? :)

[@Thesignalpath]

Sadly the diodes have to be replaced which is very difficult.

[@SirMo]

@@Thesignalpath So you're saying there is a chance! :) I understand though, would probably need one of those specialized semi packaging machines.

[@luc7luc7]

Yes it is repairable and it is not too difficult to repair providing you have diodes to replace, and a micro-gap-welder. (done under a microscope obviously)

[@asmahoseini4393]

سلام اقا ..از اسم شهریار فهمیدم شما شاید ایرانی باشید. آیا هستید ؟ ممنون میشم اگه جواب بدین

[@Thesignalpath]

Hello! Yes, I am originally Persian. But I have been living in Canada/USA most of my life now.

[@asmahoseini4393]

@@Thesignalpath خیلی خوشحال شدم .میشه بگویید که که در چه رشته ای تحصیل کرده اید

[@Thesignalpath]

@@asmahoseini4393 Sure! I did all my degrees including my Ph.D. at University of Toronto. I focused on mm-wave and high-speed ASIC design. I graduated in 2009.

[@chuxxsss]

Love waveguides and Taps to mix frequencies.