#1926

Episode 1926
this filter has amazing performance
Be a Patron: / imsaiguy

[@pieroc91]

You should take a look to a cellphone tower power amplifier, Nokia has the wildest filters, on older models they used a stepper motor to reposition the poles on the cavity and fine tune the band pass. Also the power stage is really wild, huge transistors and all sorts of RF goodness, even circulators!

[@tedivester4947]

Well done presentation. A good example of how a VNA works and how to use it to analyze a circuit.

[@davidbuchanan1440]

Bird is best known for their wattmeters, but I know they did a lot of low volume products for the military.

[@bigjd2k]

That is one neat old-school box of magic right there!!

[@DimasFajar-ns4vb]

peace be upon you sir

[@RomDump]

Did a little google sleuthing. Nebraska Surplus Sales sells an identical filter from Bird Electronic with some additional part numbers. Bird P/N: 615027 aka 84460A. Rockwell P/N: 241-0009-00. NSN 5915-00-503-0051. Google Books said it was used in General Dynamics T-29 aircraft.

[@johnwest7993]

That's a rugged, very sharp filter. I'm guessing it was used on the aircraft communications gear to keep radar pulses from blowing the com gear front ends and allowing them to hear and 'see' at the same time.

[@chriswalford4161]

I’m curious what kind of angular and dimensional precision would be necessary to make that?

[@RideGasGas]

Could identify the system it was from with the part number up top. For example, the satcom systems started with 5925-.

[@steveschulte8696]

The NSN starting with 5915- designates a filter or network assembly with resistors, capacitors and inductors in combination. The 11 digit code possibly indicates a part number assigned before about 1975, (13 digit codes started appearing in about 1975). 5905- are resistors, 5910- are capacitors, and 5950- are inductors or coils. My favorites are 5960- and 5961- tubes and transistors.

[@barrybogart5436]

UHF is pipe fitting.

[@Dieseleux]

I have one like this!

[@nickcaruso]

Hey. You have machine tools! You should (heh) try to make one and see what we can learn. Presuming it's an empty box except for the tubes, I guess, otherwise you'd have to tear the thing apart. I am assuming the box is soldered shut? Still, might be interesting to just make a box with some tubes and experiment.

[@nickcaruso]

Or, you know, doesn't PCBWay offer 3d fabrication??? Maybe even in metal?

[@tvelektron]

Very nice filter, actually i am thinking what to do with for ham radio? A filter for the nearby cell tower? But for that 400MHz is not so great looking at the 70cm Band ...

[@allananderson5840]

:-) Or when your neighbor or field day is running 1280 EME. How much power can something like that handle? I assume it would get hot filtering harmonics on a megawatt transmitter but perhaps ham power levels.

[@lunalangton5776]

Would need to plot a wider span to make sure the stopband attenuation doesn't start getting worse as it approaches GSM frequencies at >2x the cutoff (plot cuts off at 570MHz here) but yeah keeping >UHF QRM out of your RX with this should generally work. Though you're right, wouldn't want to use this as your TX filter on 70cm, you'd be in that yucky bit :(. I'd be curious whether... just plugging off some of the "organ pipes" might make the passband flatter, at the cost of less steepness.

[@gretalaube91]

I dunno, but me thinks the "order" refers to the polynomial degree that would design/express the filter function. With an LC filter, that usually matches the number of "tanks". Resonant cavity, strip, and stub filters, not so much. You could look at the steepest slope, probably at the transition,... dA/dF's in dB, and say real confident-like: "A polynomial of degree N would approach this slope value, so this is an N pole filter"... yeah, but mo = bettuh. 73 de W3IHM

[@ruudslaman6966]

HF wizardry right there

[@rfburns5601]

@ 4:52 - Kinda looks like a crystal with the parallel/series resonance thing going on.

[@johnwest7993]

You don't want to ask the price. If I was still 14, instead of in my 70's with part of my career working on spy satellite gear I would cut it open to see what's in it. :) BTW, the CB amp builders measure power in 'Birds'.

[@ludmilascoles1195]

Hmm pre NSN and post JAN so I say that short time that the USAF had its own PN system so maybe 51~56

[@MrXenon1977]

Do not switch it on, take it apart! (No criticism... I love this channel. But I´d also really be interested to see whats inside this thing)

[@johnwest7993]

There's essentially nothing inside it, or at least very little. Those tubes do the filtration job.

[@MrXenon1977]

@@johnwest7993 Do you know any link to an image or description of these things? Is it a type of waveguide or some intentionally mismatched coax cable?

[@lunalangton5776]

That is steep as heck, but as someone who's just learning I'm a bit confused by filter responses like this. Like, if you need it that steep, you're probably interested in signals close to the cutoff. But the passband starts rolling off with a kinda meh slope, then it peaks again up to about -6dB before the steep part. Even if you were interested in just that bit where it peaks again, and didn't mind that the cavity filter's exact response probably moves slightly with temperature, you've lost about -6db. Though, it does attenuate the unwanted signals quite well, so, I guess you can just compensate for the loss by adding gain, if you can tolerate noise. I guess there's always tradeoffs and someone had an application where this was good enough.

[@al_lazy3519]

Perhaps in the contrract stamped on the filter there's more info about what's needed and why

[@joeteejoetee]

WOW !!! I'm glad that you did not open it too!

[@byronwatkins2565]

Each resonator is two orders; seven resonators is 14 orders. I am impressed by the performance for the age.

[@IMSAIGuy]

any texts that explain this design?

[@byronwatkins2565]

@@IMSAIGuy Several, but it is not complicated. RC or RL are first order filters but do not resonate. LC (rather RLC) are second order filters and do resonate. Resonance (power) spectrum is the Lorentzian frequency distribution regardless of the reason for the resonance. In the stop band, each resonator contributes -12 dB/octave.

[@IMSAIGuy]

@@byronwatkins2565 yes I understand that but there is still magic on the sizes and spacing of the tubes and I expect some baffles inside. I have not seen any texts that show this

[@byronwatkins2565]

@@IMSAIGuy A single resonator is well understood; however, multiple resonators interact. (Energy flows back and forth.) There are special cases where pairwise interactions can be approximated rather well; however, computer modeling is required outside these special cases and when more than one interaction of a pair is significant. Before computer models were available, engineers used trial, error, and revision. With seven resonators this must have been a lengthy project. Incidentally, electron wave functions in molecules also have Lorentzian resonances, but the quality factor of these resonances are intrinsic and not controlled.

[@kpnconsulting8739]

I love it except for the "whoop de do". I wonder if that's a termination issue? It looks like something that should be critically damped but isn't.

[@bjj6708]

Could you please do it on a nano vna? Most of us can,t afford a hi end network vector analyser

[@IMSAIGuy]

I've done many many videos of using a NanoVNA measuring filters (playlist NanoVNA). It looks the same only slower, smaller, and harder to photograph.

[@lunalangton5776]

1:52 Explaining with your hands like Bubbles explaining what happened to Orangie to Ricky in Trailer Park Boys 🤣