EEVblog

[@swilwerth]

DC: that's a short, you cannot pass. RF: hold my beer ...

[@matthewellisor5835]

That's an Effin' t-shirt!

[@tomsmith5893]

Bought a radio from Micro Air once. They wrote on the customs declaration that it was being returned after repair (saved me some import duty, thanks lads). Their reason for repair was "radio sounded like a frog in a pipe". Most Australian sentence ever 👌

[@Thesignalpath]

Its a diplexer Dave. Not to be mistaken with a duplexer.

[@_hackwell]

Hi Shahriar ;-) Do you or Dave plan to make a video abou this ? It would be very interesting

[@hamedzaheri7050]

چقدر جالب. دمت گرم.کاش یه ویدئو راجع به این بذاری.

[@LaserFur]

@@esepecesito Each rod has inductance and the tip is a capacitor. so it's a parallel resonant. Then they couple to each other capacitively and magnetically.

[@goamarty]

@@esepecesito The 5 rods are quarterwave resonators. The reason for the coupling points beeing close to GND is to have lose coupling and do not load the resonator down with your 50 ohm input and output impedance. That would ruin your Q and thus selectivity. The milled cavities allow defined field coupling.

[@BruceNitroxpro]

@@hamedzaheri7050 , جالب کلمه ای ملایم برای این است. این یک قطعه کوچک از تاریخ است ، اما تاریخ مدرن نیست.

[@flymypg]

Designs like this are clearly short-run labor-intensive items that are nearly impossible to upgrade. At the aircraft instrument manufacturer where I worked, we called them "bricks" (where the old-school electro-mechanical instruments are called "steam gauges"). Instead, we put 100% of our effort into the RF PCBs, so the chassis could be generic and used by multiple different instruments. Yes, our instruments were double or triple the cost of this one, but ours were readily field-upgradable as standards or needs changed, and we went to great lengths to make the upgrades affordable and easily installed. A huge part of the costs of a cockpit replacement/upgrade is the labor associated with bending the sheet metal and routing and fastening the copper: Well-designed instruments that ensure the cockpit won't need to be torn apart for a long time pay for themselves rather quickly. We took this a bit further: About half of our designs were drop-in replacements for older instruments, where we would customize the back plate to precisely match the legacy instrument, then add a metric ton of digital upgrades. Just the savings of not having to tear the cockpit apart made our higher price an easy bargain overall. We pioneered a full "glass cockpit" upgrade that had installation costs that were 80% less than our competitors (such as Garmin and Honeywell), where the first step after stripping out the old instruments was to use a saws-all to connect the outline of all the existing holes, then dropping in a pre-fabricated overlay that held our instruments and also made use of all the existing wiring. A massive upgrade that took our competitors weeks to do we did in days, or sometimes hours. Our installations were so fast that we could do entire business jet fleets in the time others needed for a single aircraft. Our engineering process was guided by targeting the cockpit as a whole while also considering the lifetime of the aircraft and the evolution of private and commercial aviation needs and regulations. Then working backwards from there to create "big picture" products that sold themselves. We were a small company of under 100 employees (including those on our PCB in-house pick-and-place machines and product assembly line), and we were extremely aggressive in the market. Which annoyed the Big Players to no end. One of whom tried to sue us out of existence using their patent portfolio. They expected us to instantly fold, but instead we bet the company and hired literally the best intellectual property law firm in the US and waded into the battle. Of the seven patents they sued us over, we had 6 invalidated by the court, and the seventh was preserved only by the Big Player quitting the suit just to preserve that patent. We won not just our legal fees (which equaled a huge chunk of the value of our company), but we also got the original suit labeled as the business equivalent of a SLAPP (en.wikipedia.org/wiki/Strategic_lawsuit_against_public_participation), which yielded a punitive award that multiplied our profits for that year. That lawsuit affected all parts of our company, costing us customers and delaying product introductions and upgrades, and despite winning big, it still took us years to regain our reputation and position in the market. We weren't content with just "winning" that lawsuit, so we decided to beat them up a bit: We added a very specific feature to one of our instruments that we gave to all our customers as a free upgrade, then asked them to request the FAA make it mandatory, which the FAA agreed to do. It was a simple software upgrade for everyone but the Big Player who sued us: It turned out this specific feature would instantly obsolete a profitable "brick" product of that Big Player, which made them fight the FAA, which our marketing department used to beat them up in the industry and market. Then we offered an extremely aggressive "trade-up" discount to their customers. We actually lost money on each installation. But it was well worth it! We got a ton of follow-up business from that tactic. This is one reason why it is important for engineers to know how business works, not just in theory, but in the dirty reality as well. We wouldn't have won the original suit if our engineers weren't great in depositions and on the witness stand. And we wouldn't have had the payback victory if our engineers hadn't identified and ruthlessly exploited the weakness of "brick" instruments.

[@rkan2]

You're begging us to figure out the company and products :D Please... I guess it has to be Avidyne, since that is the first smaller avionics manufacture that I know (just because of watching Martin Pauly's videos) :) Very interesting!

[@flymypg]

@@rkan2 No can do. This happened nearly 20 years ago, and they're still in business. Though I left them 15 years ago, I still wouldn't want this story to resurface and cause them any possible PR issues today. I joined them after a decade of making instruments for the commercial nuclear industry, space experiments, and other safety-critical systems. The shift to commercial aviation was a real eye-opener in many ways, not the least of which was the immense burden of the FAA certification process. The company also targeted this as a core competency, so much so that internally we called ourselves "A test and certification company that happens to make aircraft instruments."

[@flymypg]

BTW, my experience at that instrument maker is why I'm still so ragingly angry at Boeing over the 737-MAX debacle: They made it a business strategy to find ways to weaken the FAA certification process without "technically" breaking the rules. This was the exact ethical opposite of what we did: We used the FAA certification process to make our products better! We could have had some of our employees certified as DERs, but we intentionally always chose to contract with the best of the independent DERs, from whom we would wring every ounce of benefit from their outside perspective and deep experience. If a DER wasn't holding our feet to the fire and making us sweat, we'd replace them with one who would. Then we'd hire a senior-as-God DER to review both our FAA submissions and our internal processes (including our other DERs). The FAA considered our process to be a model for small businesses: We certainly worked the rules to our favor (there are always multiple paths that meet both the intent and letter of the regulations), but we also gave the FAA feedback that helped correct weakness in the regulations and/or their enforcement. We considered the FAA an ally, not an adversary.

[@JFirn86Q]

Sandel?

[@absurdengineering]

This is like a mini business lesson right there! Thank you for sharing this - what a story!

[@Sailingon]

The shorting is the bit I've never been able to get my head around. Been a radio ham for 30 years and still gets me.

[@han5vk]

I'm fairly convinced RF electronics are just magic.

[@LaserFur]

FM = freeking magic.

[@planker]

only in FM

[@johneckert1690]

Can we go back it time to vacuum tubes. There's no magic with vacuum tubes, at least for us old timers. Up until a few years ago I was running tube type, 6 meter FM radios on my cars.

[@LaserFur]

@@johneckert1690 go play thimbleweed park. it's all about the tubes.

[@pherdantler707]

Your 'comb' filter is a combination bandpass-bandstop filter. The removable part with longer elements is very limited bandwidth 1030 MHZ bandpass, the shorter elements integral to the case are bandstop, 1090 MHZ (Tx freq) and 3270 MHZ (Tx freq 3rd harmonic)

[@graemezimmer604]

Actually I think the shorter elements are to control the amount of coupling between the tuned sections.

[@pherdantler707]

@@graemezimmer604 With all due respect to Graeme's statement, I offer that both the receiver (Rx) and transmitter (Tx) appear to be directly coupled to the antenna in order to support full duplex operation. While there is some PCB trace filtering as Dave pointed out, it is woefully insufficient to inadequate to reduce the transmit level that would be seen by the Rx input, leading to severe receiver overload and damage without an additional large amount of Tx attenuation somewhere before the Rx input. The shorter elements in the 'comb' filter displayed is the only place I noted in the video where sufficient Tx attenuation could occur. I further offer that the same line of thought has been suggested by another as well as myself.. Hey Dave, why don't you drag out on of your fancy spectrum analyzers with a tracking generator and put this 'quandary' to a rest in a video? Great way to show the value of such a unit!

[@Derundurel]

@@pherdantler707 I'd like to see that too!

[@barrieshepherd7694]

For those interested in ADS-B transponder 'RADAR' you can use a USB TV stick and a Raspberry Pi to make your own receiver, no RF knowledge really necessary and happily track the aircraft around your vicinity - up a range of around 200km depending on your antenna / location.

[@gblargg]

9:30 Capacitor can be soldered at any point along those parallel traces, to fine-tune the relative lengths of each side.

[@MarkNewtonFlies]

I’m really pleased to see this tear down, because John got it from me: This transponder used to be in my aircraft, a Vans RV-6, callsign VH-SOL. Microair is, as your research suggests, an Australian company. They’re not really much interested in avionics these days, though: Through a process of repeated acquisitions and refocusing, they’re now a company that produces in-flight entertainment equipment, and the support they offer for their old products is, shall we say, poor. This T2000SFL developed a fault, and the Bundaberg folks said that they couldn’t repair it until they’d received their next shipment of logic board PCBs, which was supposed to take three weeks. Then six weeks. Then four months. Enquiries within the pilot community suggested that I wasn’t alone, and the correct answer was almost certainly “never,” so I removed this instrument from my panel and replaced it with a new ADS-B-Out transponder, leaving this one surplus to requirements. Given the fact that it was faulty and the unlikelihood of repair, I didn’t feel like on-selling it would be ethical, and when John said he knew someone who did youtube tear-downs I offered it up to him, whereupon it remained safely in one of his desk drawers for about three years :) ATC radar doesn’t generally use primary returns (“echoes”) in the way that WW2 military radar did. Instead, ATC radar facilities have an antenna system which rotates several times per second. The system consists of a dish-shaped directional transmitter and an omnidirectional receiver. The transmitter sends a tightly focussed beam of pulses at 1030 MHz which causes transponders which see it to reply with a squit of data on 1090 MHz. The direction the dish is pointing when the omni receives the 1090 MHz reply is used by ATC to obtain the direction to the transmitting station, and the delay between the transmission of the excitation pulse and the reception of the reply yields distance. There are multiple framing formats for the data package, and the format called “Mode C” includes a binary-encoded representation of the aircraft’s altitude, rounding out the 3D picture ATC needs to stop airplanes from crashing in to each other. (ADS-B uses Mode S, which consists of a validly formatted backwards compatible mode C reply followed, after a short delay, with a variable-length “extended squitter” containing lots of other data, including high-resolution GPS position, which means ADS-B ground stations can obtain a 3D position without the expensive, maintenance-hungry, analogue, difficult to calibrate rotating dish assembly. The 1090 MHz Extended Squitter response is called 1090ES) The fault this transponder exhibited: When run on a test bench with a simulated 1030 MHz strobe, the transponder generated 1090MHz mode C responses unreliably. As the 1030 MHz power level was ramped through its range, there was a notch a few dB wide where the mode C response would fail to be sent between 20% and 95% of the time. Because ATC radar needs at least three transponder returns in a row to agree with each other (for error correction reasons), a 20% failure rate would cause my airplane to disappear from radar intermittently, and a 95% failure rate would cause me to not appear at all. The power levels which exhibited this behavior corresponded to a distance of between 20 and 45 nautical miles from the radar ground station. Which meant a ground check of the transponder would pass, and ATC could see me when I contacted them and asked for clearances, but as I drew to within 20 and 45 nm of controlled airports I’d drop off their screen, and, without radar observance, they’d have to cancel the clearance and instruct me to leave. Unsurprisingly., that was not a tolerable situation. And so the doomed T2000SFL ended up on your workbench being pulled apart. Given the trouble and expense it caused me (its replacement was A$2500), watching it being meticulously destroyed was an unexpectedly pleasurable experience. If you’d consider producing another video where the remnant parts are tipped into a shredder, I’ll be sure to like and subscribe :) Incidentally, you’re right about standard dimensions. The round part of the faceplate of this device fits into a 2.75” hole on an instrument panel. Generally speaking, the six primary flight instruments are 3.25” diameter, and secondary instrumentation is 2.75”. There’s no rack, as such; The sole physical mounting for the instrument is via the four screws in the faceplate. There’s also no standard for the shape of devices behind the panel (although there’s obviously a volumetric envelope constraining them so that standard hole spacing will work). There’s no limitation on depth, but instrument designers need to deal with practical limitations, because there isn’t usually much space between the panel faceplate and the engine bay’s firewall.

[@rkan2]

A lot of international and other larger airports would still have primary radar btw, even after all the ADS-B. The rotating antennas usually have both primary and secondary on them. In addition these airports usually have ground movement radars, so that you can do instrument landings in the soup.

[@MarkNewtonFlies]

@@rkan2 The primary radar is only good for maybe 15 or 30 nautical miles range. It also doesn't identify targets (primary returns don't have squawk codes). In Australia, only the large class C airports have primary. I'm not even sure that Adelaide has it anymore. Ground movement radar uses ADS-B these days, positions of aircraft reported as GPS coordinates. Instrument landings use localizer and glideslope transmitters which are entirely separate from radar (unless you're talking about GCA approaches, which I don't think we've ever used in Australia). Autoland-equipped aircraft augment ILS information with differential GPS for accurate guidance down to about 50cm resolution at 120 knots.

[@absurdengineering]

Thank you for this in-depth recollection, and insightful details!

[@rkan2]

@@MarkNewtonFlies I still see a few misconceptions here... ADS-B usually still only provides enhancement to aircraft tracking accuracy, but is only the primary method for ATC in very few, albeit increasing amount of situations. Some airspace use it for decreasing spacing, such as oceanic and a few airspaces in Europe at least. The same goes for ground movement radar, which will remain essential for ILS Cat II and lesser minima approaches, until your flock of birds, deer, or any other fact of life will be equipped with ADS-B. I don't know of a country that would allow ILS CAT IIIabc without an actual ground movement radar. AFAIK ILS also doesn't have that much to do with GPS for guidance, since it still preceeds any GPS/RNAV type approach. Although it is true that GPS-equipped aircraft today often augment navigation accuracy using GPS (such as IRS on an Airbus), an aircraft certified to do ILS landings, should be able to perform them without GPS assistance. Apparently there actually are some ILS approaches that require GPS, but they're so close to an RNAV procedure, that they will probably soon (if not already) be non existent: bruceair.wordpress.com/2017/10/17/an-ils-that-requires-gps/

[@rkan2]

@@MarkNewtonFlies Since you're a pilot, you might already subscribe to Martin Pauly, but I recommend these episodes especially, even if they mostly tell how they run stuff in the US: kzbin.info/www/bejne/b52zdaGhqrSdjcU & kzbin.info/www/bejne/pGjdqK2FpaxorJY

[@gglovato]

i'd love for The Signal Path to take a look at this device and do a detailed analysis and experiments on the filter with his equipment!

[@frogz]

he's seen it so he may?

[@PilotPlater]

Good explanation at the beginning on how it is operated. Speaking of the transmit modes, it's fairly simple: Standby: transmitter disabled but some electronics powered up, left over from old analog units that required a warm up time On/Mode A: transmits the 4 digit squak code which is enough for crude position and identification, but not altitude ALT/Mode C: also transmits 4 digit code and crude position. This mode uses the altitude encoder input to transmit the barometric pressure ("pressure altitude") to give precise altitude (in 100 foot increments) as well. Newer/more expensive units have mode S/ADSB to add various other information which may include aircraft registration, gnss/gps position etc.

[@rkan2]

Yeah, only baro altitude with this thing.. No GPS stuff

[@Dave5281968]

The best teardown you've done in months! Thanks, Dave.

[@rene0]

Waiting for the reassemble video and, after getting it to work, illicitly broadcast your studio as hovering over Australia.

[@74HC138]

The need for that comb filter - the transmit and receive antennas are connected together, and the receiver probably won't like 150 watts of power being dumped straight into it, so you need a really good filter that will pass 1030 and block 1090 (and its harmonics).

[@sirflimflam]

Man I love that machined aluminium housing.

[@f.hababorbitz]

I never saw a single avionics box that needed that process. It is wasteful of expensive material that is paid for, but thrown away as scrap. It could have been made from folded/pressed sheet aluminum. But it was made, and sold the way it is. I was not impressed by the effort it took to take it apart. It is not a simple maintainable design.

[@sirflimflam]

@@f.hababorbitz Pretty much all milling setups I know recycle the scrap for reuse either in house or through a third party, especially considering the waste is almost always more than the actual product in the end. It wouldn't make financial sense to just throw away the scrap.

[@f.hababorbitz]

@@sirflimflam It is scrap, and it goes to salvage. And the original 'billet' cost 8-30X that scrap that was made. For avionics the part rate is small but the material costs high, as it is all process approved by certification authorities. There are so many QC people in this loop. So it is foolish to recycle material for a part that could have been made with pressed sheet stock, and very little scrap. And it all costs money in the end to build it this way, and the energy to haul it back to be recycled. And it has so many processes to do that, and energy to separate the alloy materials, so it can all be made into a big block again. Then there are castings, that used to have magnesium for light weight, but the insurance industry shut that down from the fire risk. I recall my company redesigning servo motors to remove the magnesium, which changed the weight, and this flowed through so many document layers all requiring revisions. And then the certification authorities have to approve of the weight change, as it is stamped on the nameplate, as well the end customers.

[@Boffin55]

It's not part of TCAS, that would require a different (mode S) transponder. This type of transponder basically amplifies the response ping to (secondary surveillance) radar (generally a ping of about 200W in response), and encodes on the response the pressure altitude (in hundreds of feet), and a 4 digital octal number. (0000-7777). Aircraft are assigned a squawk (transponder) code when in positive ATC control. Some numbers have special meaning too! 7700 - emergency 7600 - Radio Failure 7500 - I've been hijacked !!!

[@Boffin55]

and it's not GPS altitude, it's 'standard pressure' altitude, ie the altitude it would be if the sea level pressure was 29.92" / 1013hPa

[@BicyclesMayUseFullLane]

@@Boffin55 Isn't it just whatever the altimeter outputs instead of QNE, or am I just stupid?

[@ovalteen4404]

@@BicyclesMayUseFullLane The transponder does not have access to the altimeter. It's coupled with a dedicated altimeter/encoder that is not adjusted by the altimeter setting. ATC does the calculations themselves based on known atmospheric pressure at the time.

[@CrobbDawg]

This transponder will respond to TCAS interrogation. ADS-B is a fairly recent invention.

[@airwood99]

Agree not part of an onboard TCAS , but would allow TCAS equipped aircraft to see small aircraft fitted with this but only give a TA response. Mode S is required in most countries so this transponder would be replaced by a new Mode S transponder in same size format, hence the tear down candidate we have here.

[@aamiddel8646]

The traces right behind the antenna plug most likely makes up a transmitter/transceiver switch (so the transmit energy is not blown into the receiver and the receive signal is not shorted by the transmitter). All to do with trace length and layout in relation to the frequency.

[@aamiddel8646]

Sorry I meant transmitter/receiver switch..

[@michaelslee4336]

Worked for an engineering company, they did a lot of fully machined aluminium enclosures from solid stock for a large comms company. Absolutely beautiful.

[@kay110]

Don't forget that those RF power transistors usually contain Beryllium oxide which is toxic in powdered form. If the top of the transistor falls off or gets damaged, then extreme care must be taken when handling the item. May be prudent to mention this while doing a teardown that contains these devices.

[@xjet]

While you were doning the "ohm" mantra I was thinking of a line from Kath and Kim: "It's Kardonay you shunt" :-D

[@gcr100]

Always glad to see you here Bruce :D

[@xjet]

@@gcr100 Always glad to be here mate ;-)

[@13FPV]

Hey Bruce! You here! How’s the ADSB chooching?

[@xjet]

@@13FPV Just trying to find the time to make the vid. Been pretty crook recently but coming right now.

[@muctop17]

All this "ohms" are marked "TL" so they are Inductors in power supply lines, to block out HF !

[@markrodenbiker2969]

I worked for a few years at an avionics company for a few years and this is very much not top of the line kit. There wouldn't be any soldered wires or a shady D-Sub connector on the back. But that's why Garmin and others could underprice on the general avionics market. Even taking out the NRE costs, the component, PCB, assembly and test costs would not scale down to that price.

[@stevec00ps]

03:55 the state of the soldering on the resistors lower left! And bottom right at 06:13, and two shorted pins on the top of the PIC Micro?

[@leonerduk]

Yup, I spotted those too. Maybe the reason this is in the mailbag was that it was a reject? Perhaps that's where it failed?

[@EEVblog]

That shorted pin on the micro would be a track joining the two pads and lack of solder mask between pins.

[@rkan2]

I doubt these things would be certified for very long if something detrimental like that would allowed to leave the factory.

[@Fluxkompressor]

I worked for the automotive industry. Pick and place machines and stuff. If we had a short like this, the board had to be fixed. Even if the pins were intended to be connected. The automatic optical inspection machine (AOI) would go wild on this board. Also the sloppy resistor placement would have been a no go. And that was just for your car infotaimentsystem. Imagine a lack of quality in a thing that prevents planes vom craching into each other!

[@RogerCillion]

Could also be a revised product. I always see handcrafted revisions on PCB on a quite big company who makes their own PCBs

[@dokta9532]

Thanks Dave, this brought back memories as I used to repair the military versions back in the 1970s. And YES I used to tune them and I AM grey bearded but NO virgin. Where is the 1090 oscillator? I would have thought that the tuned cavity was part of the Tx circuit but much has changed (I didn't see a single valve!). Back in the day the transmitter was a tuned cavity with a pickup loop to the antenna or a TWAT (Travelling Wave Tube). I would love to lay my hands on a modern schematic....

[@TheHuesSciTech]

Maybe that super fancy filter is to get the attenuation required to prevent the 200W transmission at 1090MHz from blowing the receiver up while letting 1030 MHz incoming signals through. (Edit: in other words, a diplexer [thanks Shahriar!]) You're not going to achieve that with a 2nd order Butterworth filter!

[@ammocraft]

Yeah, I was thinking the combination of rejecting the TX signal, and having good selectivity on the RX frequency might make for some nice RF design at 1GHz. I always smile when I see what is a DC short, that is nothing of the sort at RF too, like a standard LC match on an antenna.

[@Kirillissimus]

That was the first thing I though when I saw the single antenna connector - how did they manage to have such a powerful transmitter and such a sensitive receiver operate on a common line. If I had to design something like this I would just have a couple of separate off the shelf SDR modules with 2 separate filters, separate connectors, separate cables and separate antennas placed at different sides of an aircraft to minimize interference. A lot of time, swear and effort went into the two simple looking boards, that is for sure.

[@2string861]

So a duplexer? Or am I thinking of something else lol

[@tamiamibusch]

Their is no active tx/Rx switch in this type of unit hence the 1/4 wave cavity. Its a passive switch

[@michaelclayton632]

Allright guys . I was a Radar Systems Tech in the RCAF and during my basic trades training I learned the APX77 system. This is like 40+ years ago,but if my memory serves me correctly, the was a blanking pulse sent to the Rx every time the Tx fired and it took the signal to ground in the Rx. The same blanking pulse that goes to/ from the TACAN. IFF, TACAN , DME, Transponders(civilian) and ACAS all work in the same band so you need blanking.

[@pbyfr]

I remember these MicroAir transponders, it was the first or one of the first small format model. But, if I remember correctly, it needed another unit for altitude encoding. The small format is especially needed in glider (no place for the bigger rectangular format common in airplane). In the end, we managed without transponder until Becker came with a one unit small format transponder (it was a bit longer). On fiberglass sailplane, the installation wasn't easy, we needed to add a ground plate around the antenna, otherwise the emission was very bad (not good at 20 km!). And these antenna were expansive but easily cut off! Switched to more expansive shark fin antenna.

[@BikerCaf]

It's just a basic band pass comb filter, a nice one though. The milled elements/parts under/between the 5 tuned elements alter the coupling between each tuned element (hence why they are different lengths), the coupling between the tuned elements sets the filters pass band width and pass band shape along with the distance between each of the 5 tuned elements. It's all really very simple.

[@jerryfankhauser2959]

Years ago I worked where they built transponders, back then all of the logic was TTL, the readout was individual LED's, code selection was BCD coded switches, the transmitter was a coaxial triode in a tuned cavity with 1500 volts on the plate. Hurt like heck if you put your finger in the wrong place.

[@SeanPearceUK]

"What happened to the TO220?"... "The front fell off" :-D

[@davidararar]

Just like the tanker?

[@randycarter2001]

Shine a UV light on the circuit boards. They'll glow blue purple.

[@olafzijnbuis]

At 04:14 Google Automatic Captions: "...you know i'm a milf fanboy.." :-)

[@cornflake75]

it gets even funnier at 11:33 😁

[@ThislsYusuf]

@@cornflake75 😂😂😂

[@noomerical2330]

interesting. thanks for showing this. that chip cap connected to the gnd with the vias is for tuning. slide the cap along the line. it's labeled tc for tuning cap. spotted a couple more.

[@graemezimmer604]

This. Very much so.

[@linuxguy1199]

21:34 Quarter wavelength shorted stub looks like an open! Amplifier is probably so fancy since it needs to act as a duplexer, blocking out the TX with serious isolation!

[@MrJef06]

The number of screws!!! Spared no expense ;-)

[@cambridgemart2075]

The 'rigid' coax is actually semi-rigid as it's bent to shape from a roll of coax; the convoluted form there is less rigid than standard semi rigid, e.g RG402, and is designed to be hand formed rather than requiring forming tools.

[@scowell]

It has to transmit and receive at the same time... just like a radio repeater... hence the diplexer. L-band magic! Thanks for the teardown.

[@akhtarkh]

It is very interesting to see that the whole case is machined out of solid aluminum block, which seems to be the most expensive part of this unit.

[@graemezimmer604]

Definitely not a Cavity filter or a Comb Filter. It's what is known as an "Interdigital Bandpass Filter". The machined stubs between the resonators control the amount of coupling between stages. see www.google.com/search?q=Interdigital+Bandpass+Filter&tbm=isch And it's definitely not a "short to ground". The aluminum stubs are quarter wave sections at 1GHz, and the input is soldered to a low impedance tap towards the bottom of the stub.

[@BikerCaf]

No it's not an interdigital, it's a comb filter. Interdigital filters have the tuned elements alternating from which end they are ground, comb filters have all the elements grounded along the same side. This will help you learn about the difference between them .. lea.hamradio.si/~s53mv/cavity/cavity.html

[@Vict20]

Comb filter is required to protect the sensitive receiver electronics ( this can pick up secondary radar piulses from over 200 miles away) from the 100 watt plus transmissions of the transmitter. It won't let any signal pass at 1090 MHZ but will be a clear path for 1060Megs

[@shaunkruger]

Based on the product page in the description I can see why this would have ended up in the mail bag. This is only a Mode C transponder so it doesn't take a GPS source and respond with or otherwise transmit a GPS location like newer ADSB-out transponders do. Mode C transponders that only reply with barometric altitude have become interesting paperweights with the ADSB-out requirements in some areas. All that being said, it's interesting to go through something like a transponder and see a design with so many individual boards that I could see avionics shops swapping out parts from salvaged equipment to do repairs.

[@AngDavies]

18:21 Impedance inverter?- turns the capacitance into its inverse- a capacitance(times a constant Z_0^2). About the right length -line needs to be quarter wavelength which would be ~7.5cm, times speed of light in the coax as a fraction of c. Probably easier making a good variable capacitor than a good variable inductor, and they needed the latter for whatever reason.

[@dokta9532]

68.76mm

[@mikefochtman7164]

Not an RF guru, but remember my first experience seeing 'waveguides'. Rectangular shaped tubing going from antenna to receiver with big warning stickers, 'DO NOT STEP, HANG, HIT'. Definitely some UHF VooDoo.

[@Zandraccoon]

The position that the signal wires connect to the filter is dependent on the input and output impedance required. The impedance along the element is a gradient starting at zero ohms where it is shorted to ground and is effectively infinite impedance 1/4 wavelength away.

[@mattwillis3219]

Always amazing to see electronics produced to this quality, built to last for as long as they are needed.

[@emilmckellar4932]

Oooaaaaaah that duplexer, So well done for cost vs effectiveness to split 1030 and 1090. Good design work. I need alone time

[@davidmenasce6614]

It would probably have gone through all the different test cases in DO168. There are 25 different test categories for certification. One of the categories is fungus growth and another is salt fog. Conformal coating is one of the ways we try mitigate for this. There's a bunch of other stuff as well.

[@FesixGermany]

Would have been interesting to see the spectrum of that combline filter.

[@fredfred2363]

I thought I was pretty good at electronics design. Not now... What a cracking tear down. Thank you 👍🏻

[@spacenomad5484]

Speakin of how long you've been doing this... I've been watching since the ESD mats were blue :P

[@UpcycleElectronics]

Oh yeah? Who else is a warranty card shredding member of the DS1052E hack club here? And still using its smaller than a pocket scope screen?

[@redsquirrelftw]

I have an antenna in my attic to receive these signals. Sites like Flightaware actually work by people like me getting the ADS-B data and then sending it over to the sites. I'm actually one of the few in my region that has it so it's kinda cool to know I'm practically responsible for that section of the map. They sent the equipment for free. Interestingly enough the RX is all done through a SDR and a Raspberry Pi. I guess the tolerances don't need to be as serious since it's RX only, and it does not really need to meet avionics standards. The RF voodo is probably being done at the antenna though.

[@bigjoeangel]

Spotted the solder spike within 3 seconds of "We're in like Flynn". That's a bit how ya doin'!

[@benbaselet2026]

And the second one a little later.

[@mcflapper7591]

maybe it's been "revisioned"? or another tech wiz had it's go at it already..

[@shyammohabir8283]

A transponder is really a transmitter-responder to RF frequency. Pilot set a 4 digit "Squawk code" on the transponder so that the Air Traffic controller can identify the aircraft on their ground radar. It helps with collision avoidance. The mode C & S types also provide barometric /altitude info. I was wondering if the "cavity filter" part of it?

[@Momchil0]

It was made in Australia as the label reveals. At 6:29 the 0 Ohm resistor in the right corner doesn't look right.

[@mikeyhau]

It's a room temperature superconductor!

[@TheMatgaw]

So glad to see this teardown. Now I better understand why I can't afford one of these ;)

[@dave809garcia]

"ohm" mantra my dog started Barking like crazy

[@trcostan]

Transponders operate at high peak powers abs the T/R spacing is only 60 MHz so you need major isolation between the TX and RX!

[@johnathanasiou9284]

From memory, the ID button is pushed by the pilot if ATC (Air Traffic Control) request "squawk ident" which flashes their target on ATC screens for quicker identification in airspace. The 4 digit code is known in aviation as the "squawk code" & is uniquely assigned by ATC.. If you want a great explanation, see Captain Joe YT channel.

[@G6EJD]

The antenna connection point will be at the RF stage output impedance say 50R typically and that point is at a voltage minimum just like a yagi antenna where all the elements are connected to the same mounting pole apparently shorting everything out unless you look at the impedance and not resistance.

[@m.k.8158]

Dave, are you sure that what you are calling "hot snot" is actually "hot snot" Looks more like RTV Silicone.

[@Broken_Yugo]

Yeah, hot melt on a power transistor seems like a bad idea for the hackjob crap I throw together, nevermind aviation hardware.

[@danosdotnl]

@@Broken_Yugo the housing looks ok..but..yeh, wow... looks like my hot snot bench prototypes...and these stay on the ground.

[@AndrewFremantle]

Now reassemble it! How many fasteners will you have left? :D

[@muzzaball]

Dave did say he'd have bits left over lol.

[@simontay4851]

That is not going back together.

[@frogz]

...........more than he started with probably

[@craigs5212]

Yup that 's a relatively high Q narrow pass band diplexer. While ATC does use rotating skin paint classical radar it's a secondary system to the primary transponder based system. ATC transmits a rotating pattern interrogation beam at 1030Mhz, when the aircraft transponder receives the beam it replies on 1090Mhz. You can see that a couple hundred watt echo from the transponder greatly increases the SNR of the system compared to skin reflection. The reply can encode the OCTAL squawk code or the aircraft pressure altitude normalized to 29.92" Hg. The interrogation data encoding can tell the transponder which data is to be returned from the aircraft. To prevent false replies from ATC antenna side lobes ATC also transmits a pulse on an omni-directional antenna. The aircraft transponder compares the omni-directional pulse to the normal interrogation pulses and only replies if the main interrogation pulses have greater amplitude indicating you are on the main lobe. Range is done by round trip delay and azimuth by antenna pattern angle. A new system called ADSB has been put in place these days. It uses GPS and data links. Basically ACT asks the aircraft where are you and your aircraft automatically replies. For collision avoidance other aircraft can see the replies and know which are potential threats.

[@SkypowerwithKarl]

That may not be “shorted”. Those parts may have an extra thick anodized coating. Anodized surfaces are none conducive. Usually hard black anodize is the better insulation. They may have ground out a small patch to attach the leads.

[@marcellinden7305]

That RF power board looks like a Teflon dielectric board developed by Codan Microwave division (Brisbane) in the 1990's. Very stable substrate electronically for microwave boards that can have a solid metal back plate for cooling. Below 1GHz fiberglass is usually fine but above 1GHz the Teflon substrate board is ideal the way to go.

[@gustavlicht9620]

1 GHz is practically DC in high speed fiber optical communication transceivers :)

[@smunaut]

One of the reason you need a good filter at RX with lots of rejection because you're TX in at 100+W in the same antenna port ... you don't want your TX power to go burn your sensitive RX stuff. Also, that little alone cap with the long trace grounded, it's probably to be able to adjust the position of the cap for matching.

[@mfx1]

"1.0 INTRODUCTION It would be a really good idea if you read through this section of the manual BEFORE you start installing your Microair T2000SFL Transponder. If it is too late, and you are reading this message after the fact, perhaps the information that follows can help you sort things out. "

[@adamlumpkins2000]

I was wondering if you noticed the resistor that look like it was pulled off the board by one of the mounting screws?

[@jimmycrider8677]

so much fun tuning cavity filters, you have to tilt head and stick out your tongue to get it right .

[@tuttocrafting]

Dammit, rf is really black magic. Especially for me... I'm learning electronic by myself...

[@TheBinarydeity]

me too.

[@Darxide23]

I have two electronics degrees from university and RF is still black magick voodoo to me.

[@MrBanzoid]

Even if you do understand some rf theory, it still looks like black magic. Short circuits acting like opens, opens acting like shorts, straight wires acting like inductors and capacitors... and it gets even more magical as the frequency increases.

[@lmamakos]

I wonder if that short hunk of rigid coax in the transmitter section is in a stub configuration and tuned to notch at the RX frequency? Of course, you need the much more elaborate filter on the receiver to avoid de-sense of the receiver. Nice video. Anxiously awaiting part 2 of the video where you reassemble it and do a bench test.

[@russoft]

The highest I work up to is about 700 MHz, but generally below 450 MHz. It's always interesting to see how quickly RF circuits change as you get up to 1GHz +. That comb filter baffled me. Now I need to go brush up on my microwave circuits.

[@morgannahyde7502]

I have a casual interest in civilian aviation and always am put off by the cost of everything. After looking at this tear down I'm convinced that the gear is worth every penny. That case and tuned filter alone is worth the price of admission.

[@morgannahyde7502]

Oh. And why that wound up on your bench is that the rules have changed. It's called ADS-B. Briefly, the IFF transponder is now required to transmit digital information about the planes position in space. There is ADS-A which allows the transponder to receive that info from ATC or directly from other planes nearby. The info can be used to locate the plane accurately as GA (and commercial) aviation is transitioning to GPS based direct routing rather than established 'corridors' in the air based on VOR fixes.

[@MomirPeh]

I believe that the U-shaped piece of coax on the bottom of the RF transmit board is a high quality capacitor with a little trim cap at the end poking from the board. I think that the coax's isolation is teflon based and it's of really high quality with excellent temperature stability and exceptionally low leakage. And since I didn't see any kind of XO in there, I think that the mentioned capacitor is a part of the main oscillator circuit. Not sure if the crystal based oscillators would be reliable in an airplane with a lot of vibration...

[@EEVblog]

Could be, it's very specific.

[@SchlonzKlug]

Looks more like a stub hat didn’t had enough room on the pcb. You just transform a super small cap over your length matched transmission line to create a different value. And also a very neat way to creat yourself a good q “inductor”

[@PaulSteMarie]

To answer your questions, I missed the exact frequency, but I assume it's either X-Band or K-band. Guessing K-band from the size of the filter. At any rate, way back when I was working in radio astronomy in those bands, and typically they had a wave guide at the antenna focus that fed to the actual receiver element, typically the gate of a gallium arsenide FET or the end of a ruby maser. I assume they're doing something similar here. Avionics tends to be vastly over-engineered in comparison to anything short of military or space applications.

[@AIM54A]

That trace with all the vias on the RF section is probably a ground plane for that trace next to it.

[@cambridgemart2075]

Withe the single capacitor on it, I wonder if it's a Select On Test CL network.

[@ches74]

Interesting bit of kit. Hot snot is good enough for avionics in Straya? Looks a bit how ya doing. We wouldn't dream of not screwing down a TO-220 package. If something required extra support it would be in mastic or epoxy, something that's going to withstand heat without sagging.

[@jimBobuu]

That part of putting the output signal to ground... Yeah that's freaky. That's why I specialized in autopilots! If its above 400hz, I don't want deal with it!

[@MrJef06]

Haha you should watch The Signal Path, basically everything under 6 GHz Shariar calls it low frequency ;-)

[@EEVblog]

In my former industry anything over 2KHz was RF

[@Digital-Dan]

You need at least 440, to get a nice A.

[@RogerCillion]

Where is the line between HF and LF considering shielding of cables, regarding single and double shields ? Do you know. Is it at 100Hz or 1kHz ?

[@RN1441]

Sweet fancy Moses, a PIC17. 'Hens Teeth' comes to mind.

[@PartTimeLaowai]

"Made in Australia" - sadly a phrase rarely seen or heard nowadays.

[@edwardwright6961]

same as Made in Great Britain

[@cozzm0AU]

6:15 look at that bodge resistor !

[@WeBeEngineers]

Sometimes when you look at RF components like that Filter / Diplexer, you can imagine that the shorted end is very low impedance (maybe < 10 ohms in the RF realm), and the other end of those bars is open, so very high impedance. Therefore, somewhere in between, is an impedance that you may be looking for, i.e. the proper match for the input circuit, etc. So, when you find it, you can tap into it with a wire at that point. Sometimes thats 50 ohms, sometimes it different, depends on what its connecting to.

[@JamesPotts]

I'm surprised to see the hot-snot. I realize it's lower-end gen-av, but glue seems like a bad idea. The (albeit considerably more expensive) avionics that I see at work would have a TO screwed in place.

[@JamesPotts]

Though I suppose if it survives the environmental chamber, it's all good.

[@louis2760]

To the electronic Hobbyist (me) this RF Stuff looks like straight up Alien Technology

[@yurikirsanov8763]

Nothing to do with TCAS, this one just transmits the code selected. ID is used to send a 'ping' to ATC so they could see you on their screen immediately. TCAS uses transponders of nearby aircraft, but it receives their signals from transponder like this and displays aircraft around you.

[@supercompooper]

I think the Rx filter has to deal with very high magnitude spikes from lightning. Could some of that comb filter be for handling that?

[@rkan2]

A lighting would almost definitely fry this if it got to it.. Which is why planes are designed so that the lighting doesn't get to them (usually)...

[@tHaH4x0r]

Are you sure those are test jumper links? And not inductors? Also the terminated capacitor on the end of the coax most likely acts like a tuned inductor. Transmission line can act as an 'impedance transformer'. It is most obvious looking at a smith chart, where you 'rotate' around the chart with it. With a 1/4 length coax you can make a shorted output look like open and visa versa.

[@CATA20034]

From the build quality I would say it's not even automotive rated. Many caps not glued properly, allot of wires hanging in the breeze ..

[@rkan2]

Fair, then again the warranties in these things are rather different, and when this thing was manufactured, you would've been able to still fly without a working transponder in some airspace.

[@Broken_Yugo]

Yeah, I would have expected potting or a lot more glue/silicone, I bet small aircraft vibrate a lot.

[@michaelclayton632]

As a former RCAF Radar Systems technician I can say that I have never seen hot snot used in any avionics I maintained. The temperature, G, and vibration levels in military AC warrant better design I expect.

[@absurdengineering]

Yeah. It’s kinda marginal. The quality of soldering is also in the “much to be desired” category - probably wouldn’t pass muster with most avionics’ inspection books. I’d say that Dave gave it lots of Aussie leeway… I’d have been mostly “what were they thinking?!”

[@captmulch1]

Dave - generally flying around in uncontrolled airspace the transponder code is set to 1200, and the mode is set to 'on'. If flying into controlled airspace most light aircraft set their code to 3000 and mode 'on'. This just gives a general ping/paint of where the aircraft is on the air traffic control/flight service radar screen. If flight service or air traffic control is wanting a more detailed identification they will ask you to set a particular code e.g. 4235. They may also ask the pilot to set the mode to 'Altitude' (Alt) so it paints the code and altitude on the radar screen. The transponder is always set to 'standby' when not flying (i.e. taxiing) and when changing codes. Instrument flight rules flights and the Big Boys always get a custom code (OK, IFR class G is a standard 2000 code) as air traffic control is responsible for keeping them separated from other aircraft. The transponder is also used for emergency signals such code 7500 (unlawful interference), 7600 (communications failure), 7700 (emergency). These are useful as a pilot may have lost their radios, so air traffic control can be alerted with the transponder. A hijack can also be drawn attention to with a subtle change of the transponder to 7500. vfrg.casa.gov.au/general/radar-transponders/transponder-operation/

[@RobertBardos]

fantastic dave! thanks alot from USA!

[@DanielLopez-up6os]

I particullary Track 1090mhz frequency locally to track planes flying withing 100 miles :D

[@radarmusen]

Fantastic to see all in one panel, normal there is a black box and a panel .

[@stevetobias4890]

The antenna being grounded to the unit housing which is then connected to the aircraft body, wouldn't this act as a ground plane for the antenna to operate?

[@isettech]

At those frequencies the antenna ground plane can be a simple soup can lid, if the antenna design needs a ground plane at all. The antenna could be a dipole design, or other design not needing a counterpoise.

[@stevetobias4890]

@@isettech cool, thank you

[@absurdengineering]

@@isettech 1GHz is 30cm wavelength so that doesn’t quite check out unless it’s an institutional size soup can :) But a paint can lid would probably work OK.

[@Graham_Langley]

Surprised to see a BNC connector - back when I worked on fixing Bendix avionics in the 70s I was told they fretted and failed in aircraft and the threaded TNC version should be used instead.

[@maxvideodrome4215]

For how much I paid for mine, I figured it would be made of gold!

[@matthewmiller6068]

If the DC short passing RF makes your head hurt consider that at RF some insulators can conduct!

[@rubikmonat6589]

Month late, but as I explained to an electrician once concerning waveguides, we transmit big power through insulators rather than conductors, more efficient.

[@kenjewell3889]

I am really surprised they are using hot melt glue. Having worked in Aerospace electronics design for years, designing EMA for players like Boeing and Airbus, hot melt would be a no go component in any actuator, not even used in ground testing.

[@stevec00ps]

03:16 Is that "No intentional spins"? Might as well say "No buzzing the tower" :)

[@bp1carnal]

That's why I always plan an accidental spin lol

[@mcflapper7591]

my first read was "no intentional sins". :-D

[@FesixGermany]

We have trolleys for 48h burn-in of devices with 24V 1000W MeanWell power supplies who fail after some time most probably due to vibration because these trolleys are moved through elevators and across the production hall... these are just not meant to be used like this, can't blame MeanWell for the failures.

[@LaserFur]

Be careful with disposing of those transistors since they could contain Beryllium oxide.

[@electronicsNmore]

Amazing how expensive aircraft and boat parts are. Not too much there for $2500. Must be paying a lot more because not that many are sold compared to consumer electronic devices. Great video!

[@ThePoxun]

Remember its not just the components you are paying for and this certainly isn't something any semi-competent engineer could design in a few days using the reference designs of a couple of off the shelf components. These are life or death avionics parts that take a significant amount of effort to design, test and certify to the reliability needed. It's not like consumer electronics where if something breaks the only consequence is someone can't listen to music for a bit.

[@ovalteen4404]

Plus all the regulatory BS that comes in getting them approved in the first place. Then the requirement that all parts come from high-assurance testing. If a chip goes out in that transponder, the tech can't get a legal replacement from AliExpress. Instruments for homebuilt/hobbyist aircraft can be installed much cheaper, and usually have more bells and whistles.

[@electronicsNmore]

@@ThePoxun True

[@andrewsnow7386]

Not much there? That's the opposite of what i was thinking. 🙃 So I don't claim to understand any of the electronics, but just that machined metal box is amazing. Not only did they spend the machine time to machine away maybe 90% of the aluminum, it had to be machined from all 6 sides, guaranteeing it has to be re-fixtured at least once. And there are 5 other complex machined parts! I'm a little surprised they can include any electrical parts at all for $2550. 😁

[@Spirit532]

It's $150 of instrument, $2000 of certification, and $350 of profit.