The Signal Path - 2022-03-13
In this episode Shahriar repairs an OPTOELECTRONICS APS-104 tunable band-pass filter. The instrument provides continuous tuning from 20MHz to 1GHz with a constant 4-MHz band pass response. Since the repair proves to be fairly easy, the video focuses on comparison between various band-pass filter architectures. This includes microwave cavity filters, PCB-based filters with varactor tuning, YIG filters and frequency-translation filters. A range of measurements highlight the differences and pros/cons of the aforementioned filters. The APS-104 unit is fully reverse engineered and characterized as well. www.TheSignalPath.com www.YouTube.com/TheSignalPath www.Patreon.com/TheSignalPath
You know it's going to be a good day when you see a new TSP video
Those cavities remind me of the RF voodoo recently shown in CuriousMarc's Apollo communication equipement restoration. I'd love to see a crossover episode where Shahriar explains in-depth the complicated RF parts of the restored Apollo hardware.
This is very interesting. I've always been envious of your 2GHz tunable mechanical filter. This teardown may mean this thing could be reproduced for home gamers.
[edit] Do you have any tunable Yig bandpass filters? That would also be an interesting video. LOL, I just got to 14 minutes and you did discuss Yig filters.
I am so excited that you're working on a noise figure video! I've always struggled with that. Over the years, you have made an enormous difference in my understanding and enjoyment of the hobby. You've really multiplied my abilities. Thanks, Sharihar. You rock!
Fantastic video as always! You are doing the field of engineering huge service by sharing your knowledge with us all!
Wonderful episode, thank you Shahriar for putting the time into content like this!
I'm reminded of the first time I saw a microwave cavity filter, back in the late 80's. I was fresh out of college, and went on a trip with a senior engineer to Red River Army base in Texarkana. The product they were having trouble with was our PPX3 interrogator used with the Stinger missile. A whole batch of new PPX3s was failing self test. We narrowed the problem to the diplexer module and found that if we banged it on the bench, it would start working again. Back at HQ our failure analysis team did something that really impressed me; they were able to cross-section one of the tiny cavities and look at in with an SEM. They found tiny dendrites shorting the screw and the wall of the cavity. The amazing part was they didn't break the dendrites during the sawing process. Anyway, it turned out the problem was caused by a new lubricant the cavity manufacturer was using, which they got from some supplier. The lube guy wouldn't tell us was was in it because it was proprietary. Switching back to the old lube fixed the problem.
I remember seeing a General Radio catalog from the 1960s or 1970s that had an audio spectrum analyzer built on this principle, intended for frequency response and distortion analysis. I've always wondered how they kept the linearity good enough for meaningful distortion analysis.
Nice video. Similar concept to the Wadley loop used years ago in the Barlow Wadley radio.
Have you considered doing a video on passive intermodulation distortion? It's a relatively new problem area in high power multi-carrier RF systems.
Very nice video as usual !
Can you consider making a video on N-Path RF filters? They are very interesting and quite surprising
Thanks for this in depth video. I really like this kind of content from you!
Loved this video! It took me back some 45 years when i was playing with coffee cans screws and 2K25s
Good episode. Like the design. Wondering if I could use something similar to make a tunable notch filter. Need very high Q at 108.0 to 108.2 MHz - not easy. Using the manually tuned "oxygen tank" filters now. Again, not easy and clearly not able to automate that.
all those hand-tweaked vanes on the tunable bandpass filter.. wow. And was it my imagination that the center disc of those capacitor elements seemed to be slightly tilted such that the relative distance to the left/right side changes during the rotation as well?
Thank you for this video. Takes me back to my lab days at NASA.
Great video!, I'm loving these new tutorials
Great video, as usual! Out of curiosity, would you share the DOI for the microstrip filter paper you mentioned here?
Great video as always! Only disadvantage is that I don't have it and all of your instruments as well :D Noise figure measurement video will be appreciated so much! And also please focus on possibility how to do that in amateur shop. Mine is not so much amateur but far away from your. Therefore keep in mind different possible setups.
Looks like half your RF lab instruments participated in this tear down. Wow!
Would also be interesting to see what, if any, phase noise degradation is added along the signal path. The phase noise of the LO should cancel since it’s coherent up and downconversion. What was the LO frequency?
Very interesting high quality top notch video as usual in the signal path. Waiting for the noise figure video and the keysight fieldfox teardown also.
A simple way to build a fixed bandwidth filter. Rather elegant from a theoretical standpoint as long as its downsides aren't problematic. To a degree similar to how a lot of radios and spectrum analyzers filter.
Mixers are truly one of the RF engineer's more important tools. All though one with tons of downsides.
Though, if one adds a variable bandwidth filter instead of a fixed one, and perhaps a more stable local oscillator, it could turn into a very interesting tool. Especially if fitted with a simpler display for easier adjustments.
This was typically connected to a telescoping antenna and Optoelectronics frequency counter to find the transmit frequency of two way radios in a congested radio environment. The architecture was obvious but I had never seen the filter. Thanks.
I can donate very sharp superconducting RF filter assemblies, used in celltower GSM amplifiers, designed to operate at 77K temperature. I bought unit years ago to get cryocooler unit, and have no use of RF parts. Did some S11 measurement at 77K and higher temperature too :).
Ah yes we had a little fad for cooled amplifiers in the GSM world a couple of decades ago. Then found it was more economic to just place a normal LNA directly at the antenna before the coax run to the base station.
Over 50 years ago I built a single tube short wave regenerative receiver that used ‘reaction’ ie positive feedback for bandwidth control. If the feedback was adjusted just below the point where oscillation sets in the bandwidth was tightened and tightened. A common technique at the time. There was also a version that did run with oscillation occurring but used an ultrasonic ‘squegging’ to keep on killing and then restarting the feedback giving increased gain. These were called Super(sonic)-regen receivers and were commonly used at VHF and UHF. I’d be interested in Shariah’s explanation of why the Q increases as it goes into oscillation!
The explanation is relatively simple. Q factor tells you how much energy is being lost in the oscillator in a single cycle. The oscillator with finite Q is going to loose some energy in each cycle, the oscillation will eventually stop and the bandwidth is finite. The oscillator with infinite Q losses no energy at all, the oscillation will last forever and the bandwidth is infinitely small. Now if you have a positive feedback loop around an oscillator with finite Q, the energy lost in the oscillator gets replenished by the loop increasing effective Q and reducing the bandwidth. The higher effective Q, the closer you are too oscillation. The final point is where effective Q is infinite, which means you have oscillation running forever, infinite Q and infinitely small bandwidth.
@Adam Turowski Great explanation - thanks. That also explains ringing in high Q filters I guess.
@RichardInCancale Correct
Love the new addition of the cat in the TSP logo too :)
That is the TNP logo! :)
That was a grate episode!
thanks for the video! i didnt understand - first filter -do the 4 rotate individually to make different superposition of capacitor values?
Now this is a sufficiently advanced technology for me to call magic. As CuriousMarc would say it, "the blackest of the black arts in analog electronics". Beautiful design and build quality.
I'd say microwave engineering is a little more mysterious than analog electronics. :)
@The Signal Path You're not kidding. I just learned today that the spinning boom for ship's radar is nothing but a microwave wave guide. With my father in the US Navy, I've been looking at these things all my life and had no idea. 😳
That is a pretty nice idea actually.
I guess something like this might come in handy as a preselector for an SDR radio.
I might have a new summer project, I guess.
Thanks for the interesting video, I would definately like to see more obscure RF and microwave components or maybe also devices with rather unconventional principles.
EDIT:
Maybe some more cellular stuff would also be cool for a mini series.
Like getting your hands on some GSM/UMTS/LTE base station boards and just showing the architecture, explaining the siganl path and maybe do some reverse engineering and get some low level stuff working / reused.
Well using an active preselector for SDR is not a very good ideas. The active components in the preselector are likely to get into non linear parts of their characteristic due to saturation and generate intermodulation products - as Shahriar shown in the video. Passive filters are way better as SDR preselector.
@Adam Turowski I agree with you in general but do you think this might actually be a problem for some over the air signals? By getting the gain of the unit way lower like closer to 0dB it should be less of an issue? In my experience I basically don't ever get a signal stronger than -60dBm which I guess should be fine.
However as I said earlier you definately got a point to keep in mind.
@Da_Stier Preselector is only needed to get rid of the signals which amplitude could push SDR input into non linear region - and this applies to all signals being present across whole spectrum coming from antenna. If you can guarantee that there are no such signals present at SDR input, then you don't need preselector at all. If you are dealing with very weak signals hidden in the noise, you rather need LNA - with the same caveat that you have to guarantee that it is not going to be pushed into non linear region by excessive signal amplitude coming from the antenna.
Thank you; I always enjoy your videos
Would a passive tracking filter with a non-constant bandwidth before the fixed bandwidth filter be something which would be used to improve the linearity problems?
It would have to be before the input of this entire module. And of course you'd have to tune them in tandem.
@The Signal Path I understand and that's what I meant with "tracking"
I made UHF self-tunable bandpass filters with varactors and coaxial resonators
What kind of coax cables are those thin gray colored ones you used for the VNA? Are they semi-rigid? Thanks
Nice vídeo!
By adding the 20db attenuator, didn't you drop s11 by 40db? Pity you only showed smith chart.
I didn’t talk about the measures S11 of the unit because (as you said) in my setup I was basically measuring the return loss of the attenuator. The return loss of the instrument is not bad, better than -10dB for most of its rated BW.
@The Signal Path thanks for sharing.
That was fascinating.
Every time I watch a TSP video I need to get a towel to wipe up all of my drool on the floor from the various pieces of test equipment shown in the video... <<<Sigh>>>
RF is beautiful and magic 😆👍
nice work thankyou.
I live in a senior citizens building with a cable outlet in each apartment. Once we switch to the cable input on my TV we can access predetermined channels like 22, 70, 96.1, 98.546, etc. but the only signal that we can access is channel 96.1 which is a screen split into four sections showing the laundry room and the entrances. Can this cable system be used to create several channels where news bulletins can be released to the apartments providing various information like vaccines, or water turn off, etc. health issues which are customized for the residents on a regular basis? Any help would be appreciated. Thank you kindly.
Is it upconvert - fixed filter - downconvert?
I shall watch the rest of the video and then no doubt feel pretty silly.
Thank you kind Sir.
So the bandwidth should always be fixed?
Well it depends on what you want to do. If your application needs fixed BW, then yes! :)
Nice...!!
Thank you.
2-port ass parameters 🤟
Great Video, great Lab, but where is the cat?
👍🙏👍
Optoelectronics is hardly a high end equipment manufacturer. What you got is fairly typical.
w2aew - 2022-03-13
Nice video - really cool to see the performance on all of the different instruments. Note that you could connect the LO and the output of the up-conversion mixer to additional channels of the scope and look at the spectrum of each of those (and you can change the CF of Spectrum View on each channel individually). I recently did something like this to show how a superheterodyne shortwave receiver works.
The Signal Path - 2022-03-13
Absolutely! There will be a lot more of the 6-Series. :)