Вставка
- Опубліковано 21 лип 2024
- In this episode Shahriar investigates a failed R&S FSL Spectrum Analyzer. The instrument shows an unusual spectrum trace and fails a wide range of self-test functions. The LO Unlock failure is likely the most important item to address which can cause a cascade failure of other functions.
The teardown of the instrument shows a single-board RF design. The system architecture is analyzed along with an X-Ray of the main board showing various RF/LO routes and interconnects. The core synthesizer is examined and measured along the signal path to verify the operation of each part. The failure is traced to an open 1-Ohm resistor which feeds the power supply of the divider IC feeding the PFD of the PLL. A ceramic capacitor failure is also observed which may explain the damage to the 1-Ohm resistor. After these components are replaced the instrument passes all self-tests. The functionality of the instrument is verified by measuring tones from 9KHz - 6GHz using a Keysight MXG.
www.TheSignalPath.com
/ thesignalpath
www.Patreon.com/TheSignalPath - Наука та технологія
I am amazed you managed to find the fault so quickly on a complex device without any documentation. Well done.
When the failure comes to passive element in such a complex circuit and you can track it and fix it...nobody can say you cheated the exam on the front panel connector😉
If videos like these do not encourage people into electronics and RF, I don't know what would do. Your ability to trace down the fault is no joke! Thanks a lot for these educational repair videos!
Whoa - that X-ray machine is new, huh? Nice debug and repair...
Hi Alan! No, I have had it for several years. :)
Amazing. Few minutes, without manuals, and voilà. Unbelievable.
Nice to see my probe getting some use! I'll send you some pogo tips for it which should improve the landing quality. I can send you one or two of the solder-in sibling, the AKL-PT2, too.
Thanks! Improvement to the tips would be hugely welcome!
Will the AKL-PT1 be back on sale at some point?
@@BaconbuttywithCheese It never was sold openly in this form. I did a kickstarter for the 2.5 GHz first revision then massively improved it to the 6 GHz bandwidth prototype Shahriar is using.
I'm still making slight tweaks to improve flatness and ergonomics before a full launch. Its solder-in flex sibling, the AKL-PT2, is available now.
Very good find. The excellent knowledge is the key to locating the spot. Without you have no chance cornering the cap. Brilliant !
I lear alot from your thought process when you tackle a faulty device.
Much appreciated.
Wow impressive. I love watching your video's even though sometimes your discussion goes over my head. Wonderful analysis to watch
really nice analysis! to find that tiny cap. Great job and very interesting to see all the process.
Absolutely excellent diagnostics. Fascinating for an electronic engineer here who only ever built low band stuff.
Nice to see the detective work, going from the resistor voltage drop to the capacitor causing the resistor to fail.
You know, at the start of this video when you stated "no, it's not a front panel connector issue" I jokingly said to myself, "ah, probably needs a recap then!"
Sorry, too many Mr Carlson old radio videos in my system. :)
you make this looks so easy...thanks for the video
Awesome, as always. You are the HF Master!
Outstanding repair, as always :)
Really nice repair with excellent explanation! Thank you!!!
Awesome diagnosis and repair. I wouldn’t know where to start. You still got lucky that it wasn’t one of those custom IC’s that failed. 😺
IC's tend to be reliable now, as the early failure modes were researched and understood, and fixes applied to the process path for the more reputable manufacturers. The cheap ones know how they will fail, and roughly how long that will take, so just build good enough, to last that long for the vast majority of the devices. I would say R&S, and most of the other top end manufacturers, know the equipment might be in use for decades, so go for the most reliable IC's in there, and good quality for the rest as well.
Excellent job methodically trouble-shooting this fault and repairing it!
Great video as always!
I believe these decoupling capacitors shorts are the main culprit for many PC boards malfunctions.
Fascinating stuff...cheers.
Superb reasoning. Pretty sure I would have been lost. Reverse engineering a complex board looks easy when you're doing it.
Great repair 👍
Great video
Great work 👌
So the bypass capacitor developed a partial short?
A common fault of MLCC: cracks which produce a short. commonly on MLCCs with higher number of layers, i.e. high capacitance. That one looked like a fat one.
I find these all the time
MLCCs will crack easily if heated unevenly, and silver migration through the cracks courtesy of lead-free solder will ultimately do the rest (note the Pb-free symbol on the board). An unholy combination.
To think that an instrument like this (few things R&S are ever cheap) can be taken out by a single bad one like ordinary consumer electronics is wild. I hope R&S have at least considered going with automotive grade MLCCs in more recent times. This model appears to be from the mid-late 2000s.
@@LordPrecision but all this eco-posturing makes purple dyed hair people feel superior about themselves
Eco-posturing? Pathetic.
o-my... i really thought that when you looked under the R&S hood there would be an arduino hanging on cable-ties...:/ this is so exciting, let the show go-on...:) thanks smart guy..
As usual, great video!
I have one question, though: isn't the noise floor quite high at -50dBm? I would expect at least -80dBm from an R&S Spectrum Analyzer. Even the "TinySA" has about -80dBM noise floor!
Hi!
I think, it's because the reference level is 9dBm (it's too much for preciece measurements) and large RBW (3MHz).
Or maybe cheap 5% accuracy capacitor... Who knows😀
I was actually hoping to see another early-'90s one with Transputers...but people tend to hold on to those things due to their performance being so good.
I have a R&S VNA with the transputers, it’s pretty great. But I never knew that people wanted them.
bonne représentation et explication fort analyseur
Where did the analyser come from, what did it cost to buy as a broken unit?
How far is that cap from the shield fixing screw? If it's pretty close then root cause is a design error; too much stress on an MLCC. Awesome repair video, outstanding presenter.
I have seen this a few times when caps close to mount hole would crack due to mechanical stress.
Hello!
Very interested in your X-ray machine!
can you provide x-ray information as your device is called?
I feel hard done by if I have to fix a Betamax video recorder without a schematic, and you fix a spectrum analyzer!
A1 Mate very nice
Thanks for this video , you didn't use thermal camera to tell us that with basic tools we can do a repair.
The PCB of magic
"Unsurprisingly it passes it's self test". Well of course, as the Master has laid his hands upon the instrument.
Another unit rendered inoperative by a few cents worth of parts. Unbelievable 🤔
I love a great $0.008 repair.
So often a defective cap or resistor causes the unexpected failure.
Nickel barriers are a huge issue, poor solder ability at times, or they magically debond and create high resistance.
Nice repair. I just wonder why the noise floor is so high. It's around -35 dBm at 6GHz.
3 MHz of RBW and a fairly high reference level with 30dB frontend attenuation increases the noisefloor quite a bit. Additionally this SA might not be the lowest noise instrument
It is from the entry line series. Low performance, low cost.
Super
👍👍
Daily driving such a FSL :)
>googles price of Rohde & Schwarz
"Oh... okay... so it's worth more than my car. So that's something I'll never own in a million years."
FSL is from the low cost (low performance) series.
me as a junior embedded systems engineer watching this video thinking to my self "what got me into this!! probably i will never be this good"
I have spectrum analyzer has fault sometime show LO unlock, I can open transit case to check main board, please show me hownto open it?
did you actually record this video less than 48h ago and already did all the editing?
I do almost no editing at all. It is just recorded sequentially and stitched together. I normally don't spend more than 20 minutes on editing.
so it's almost a recorded talk
9:00 X-ray that only stings once? ;D
"We probably won't even be able to find the data sheet for them"
*Louis Rossman has entered the chat*
Haha... Only if it were as easy to get this schematics as it is to get an Apple laptop schematic.
@@Thesignalpath you mean Paul Daniels doesn't have board view software for that? :)
Why some traces that connect resistors\capacitors or inductors don't have solder mask?
These are the RF traces. Solder mask is a poor dielectric material and is often avoided on RF paths. It is also generally not included in the EM simulations.
@@Thesignalpath Thanks. Can you explain more about this "Solder mask is a poor dielectric material"?
Hello, Can you calibrate CMW500?
Amazing how a 1 cent part can ruin such expensive equipment
Nice reverse engineering and fault diagnosis, it is surprising how many times a fault comes back to a bad ceramic capacitor.
Your into- and outro- music is cool, I guessing you play electric guitar lol?
Too bad those probes are not available! They only have available a Passive Solder-In Probe which is just an expensive piece of transmission line, the others are still vaporware!
What is the I-V characteristic of a catode?
Almost all physical processes, but especially all electrical characteristics of semiconductors (diodes, transistors and similar components) or vacuum tubes, are "non-linear".
That means, linear change of a parameter, e.g. applied voltage or resulting current do not show a linear, uniform change of the corresponding result. In order to represent this optimally, these values are graphically displayed in curve form in a so-called "4 - quadrant characteristic curve field" - "The I-V Characteristic Curve".
This enables the simultaneous representation of 4 characteristic curves, in 4, horizontally and vertically adjacent fields, registered on a +x -x +y -y axis cross (Field number: I - V), with change of their values (voltage, current, resistance etc.).
Sources:
(Excellent) translation program : www.deepl.com/translator
duckduckgo.com/?q=4+quadrant+characteristic+transistor&t=ffsb&atb=v96-1&ia=web
fytronix.com/PHYSIC-EQUIPMENTS/Characteristic-curves-of-semiconductors-PHY2013.pdf
www.sciencedirect.com/science/article/abs/pii/S0042207X19304555
duckduckgo.com/?q=4+quadranten+kennlinienfeld+transistor&t=ffsb&atb=v96-1&ia=web
duckduckgo.com/?q=4+quadrant+characteristic+transistor&t=ffsb&atb=v96-1&iax=images&ia=images
duckduckgo.com/?t=ffcm&q=the+I-V+characteristic+of+a+catode&atb=v96-1&ia=web
Greetings ... :)
@@christophbrand i feel bad now, because I was making a joke... i know what I-V characteristics are and I know what cathodes are
but not what CATodes are...
@@GeorgeTsiros
Hmm ... ok... well... I'm not angry with you. I hope you had at least a little bit of fun, otherwise it would have been all for nothing.
Well, it was also a little bit fun for me too, rummage in old memories of my education, which are already a little bit "dusty". ... ... ...
I think, the way you meant it is probably the ode (poem form) to a cat was meant ... oh what ... is also no matter.
You have led me, once again, clearly in mind, in any situation not to lose the "view of the whole", and at least to keep the/my humor.
For me, humor, fun and jokes are a very unique and (and especially related to the "present time" ... an over-) vital essence of life itself. It helps reliably to get over "bad times", and I had more than enough of them in this life ....
With kind regards
christoph
@@christophbrand look, _i_ may not needed your explanation, but others quite certainly do! Your effort was absolutely worth it, because anyone here can read it and get some information!
From where do you even buy these?
eBay!
@@Thesignalpath What do you do with repaired units? Do you sell them back?
When ypu speak everything seems easy. XD
Repair of defective R&S FSL Spectrum Analyzer: 1) remove covers, 2) remove PCBA 3) analyze circuit 4) find defective parts 5) replace parts. Equipment required: 1) X-Ray, 2) VNA 3) special probe (fund on Kickstarter) 4) soldering iron, solder wick, hand tools, etc.
Or buy a new one from Siglent; >$100,000 test eqpt. not required :)
Great stuff Shahriar. I always learn a lot from your videos and am a big fan. Thank you.
A narrow vision...
en.wikipedia.org/wiki/Electronic_waste
@@Thesignalpath Yep. And if right-to-repair legislation makes providing schematics (and hopefully layout) mandatory the X-Ray could be skipped.
@@Thesignalpath You should send this to R&S. No schematics, no manuals, no parts. Pure arrogance creating land fills of dead R&S products.
@@Thesignalpath Repair and Reuse, Shariar is doing his utmost to get the last one, Recycle, out of common use, as repair and reuse is the most use of the rare elements that were used, and gives the lowest impact on the environment as well. That he also gets some impressive use out of them is also a bonus. Then of course the equipment that is tossed for an utterly trivial fault is the best value as well.
@@shazam6274 First of all, R&S does not perform component level repair. No big T&M company does. The board would be tossed out and replaced. We are not even taking into account the cost of shipping and repairing it. Aside from that, it represents an opportunity to offer education on top of the repair benefit.
ora mudeng bahas opo
Hello, your work is fabulous, well done. I would like to ask you for advice because I have a Rhode&Schwarz FSL3 and the tracking output on the spectrum analyzer is defective. I don't have any schematics or electronic plans. My problem is the following: from the tracking I output a 0dbm, then from 9khz to around 100khz the tracking is not good at all. After 100kz it is at -2.5dbm up to 1ghz and tends towards 0dbm very slowly up to the max frequency 3ghz. Can you help me please ? THANKS
Seems like component problem. But R&S may charge you $3000 for repairing.
No, they will simply swap the board with a new one, and charge you $15k for that, and extra for the calibration afterwards.
@@SeanBZA EXACTLY what happened on my RTO scope. Two channels just dead suddenly one day (seems like communication problem between front end board and main board). They swapped the main board and calibrate, which costs $6,000! I still don't know why it's broken or even maybe no hardware is broken?
now the question is, would it be more or less expensive than 6k to pay an engineer to actually find the fault, fix it, and verify everything to a level that you're confident in the fix as a company?
@@tommihommi1
Why employ engineers full-time when you can short-term contract them to do the initial design, and then vanish them with a draconian NDA in their back pocket?
All of these companies are the same now - They have impressive research departments, but are mostly just marketing companies that hire the odd bit of part-time talent from time to time.
Compare the above to a robot in China spitting these boards out at 10 per second.
@@digitalradiohacker the engineers are busy designing the next generation, from what I know Rohde Germany is a decent employer
It honestly feels like you're rushing through these videos recently, you're hardly giving yourself a chance to take breathe.
Limited time to record.
№2 ))
Okay so, two years ago, i let a cat stay at my place.
My troubleshooting skills did not improve at all.
I come to the conclusion that it is not the cat.
very expensive...
10cents of parts bring it down
Whenever I have an RF fault in something, I imagine an awesome rack of high frequency test gear, and how I could work through the RF stages . Then I look down at my old 40Mhz scope and realise, this isn't going to go well :-(
hi thanks for posting your knowledge !!! Could you activate subtitles in Spanish ??? I am a follower of your videos, from Argentina !!! Spanish subtitles??
I have enabled the Auto Caption. Hopefully you can set it to any language now.
#1
You have good RF knowledge, but your pace and your voice make it difficult for me to watch the video.
I am sorry. But it is a matter of lack of time.
UA-cam has a playback speed feature. Watching at 0.75x or 0.5x speed may help.
Great video