ฝัง
- เผยแพร่เมื่อ 6 ก.ย. 2024
- In this episode Shahriar repairs an Agilent E8257D PSG Analog Signal Generator generously loaned by AllTest. Please visit their website for all your measurement, calibration and service needs:
sales@alltest.net
alltestchamber...
alltest.net
The instrument does provide an RF output signal. However, there is also a large DC offset voltage present at the output RF port which changes depending the frequency band. The DC voltage is present even when the output RF signal is disabled. Furthermore, the OCXO of the instrument is defective and does not produce a 10MHz output signal.
The block diagram of the PSG is examined in details with emphasis on the final attenuator, coupler and doubler RF decks. Interestingly enough the output offset is traced all the way back to the Modulator Filter block. The teardown of the module reveals a series of PIN diode switches and after some investigation the fault is traced to a damaged PIN diode on the sub-3.2GHz path. Since the diode can't be easily replaced, the control voltage to the diode is disabled instead which removed the DC offset problem.
The OCXO teardown reveals that during a prior repair, the oscillator module has been damages and torn off the PCB. New pins are added to the oscillator module which returns the crystal back to working condition. The performance of the PSG is verified using a spectrum analyzer and frequency counter.
www.TheSignalPath.com
/ thesignalpath
www.Patreon.com/TheSignalPath
![POLYCAT - ข้อความรูปยิ้ม [Music Video]](http://i.ytimg.com/vi/lBhIgPGCgbI/mqdefault.jpg)
Love the new logo with Pooch :-)
Stay safe in these crazy times!
Ah yes, the classic common source cat amplifier, I think I'm going to use my cat as a wifi extender as well.
Those two power transistors inside the oven are actually the heater elements. They operate in linear mode, and are easy to control how much power they would dissipate. And the white wiring inside probably goes to a temperature sensor close to oscillator itself.
Interesting. It could be! I should have measured it.
It makes sense
I'm mystified how you could get thumbs downs for a video of this quality. Just what could anyone find to complain about? Good luck with getting and using a wirebonder, I last got to play with one of those around 1988.
There is a simple explanation for this; the like and dislike buttons are very close together. That's not a very good UI decision and you'll always get some misclicks.
@@JimmyAlzen I have accidentally clicked the dislike on some videos when using my laptop with my fat fingers but you can always correct it. But I agree why anyone with even no brains would ever leave dislike on his great videos !
I think Pooch wants to tell you that you should definitely get a wire-bonder ;)
For future reference on repairing that crystal. Those broken wires could be welded to new wires by a good jeweler with a laser welder. I had some work done on a set of diamond ear studs that had one missing mounting stud, they welded a new one on in a few minutes.
I worked for HP/Agilent in Labs, and supported the Santa Rosa site for a while. Congratulations on fearless troubleshooting and repair! Not many are willing to crack open an OXCO to look for problems, but that part is easy after digging into one of the RF decks!
That make the quarantine a little more bearable. Thanks!
all videos about " Repair of an Agilent ... " are very nice, thanks, please continue ...
Best RF/Microwave repair video series on TH-cam of all time!
The beauty of the troubleshooting. With thousand and one parts , you single out one poor diode and two wires detached. I really like to watch the process. Stay safe!
I like cats as much as the next guy, but connecting a static generator directly to the gate if a FET...? :)
Brilliant video, so professional. Fascinating to debug it in my head along with you, regards, Brian.
Another excellent video, thank you! I'm suprised there is no key/notch/pin/set screw to prevent the crystal mount from rotating when the screw is tightened.
Consider their idea of anti-tamper - touch it and we'll have to sell you a new one! :)
I enjoy to watching these repair videos
Wow a real black art, I didnt even recognize the parts you were talking about, great to have this Hi Frequency stuff
Excellent repair! Looking forward to you getting a wirebond machine!
Hi, do you know which kind of wirebond is used at 27:05 on the output line?
I really appreciate your videos and how you outline your train of thought. Invaluable.
Great video, thanks!
Hey Shahriar, your episodes are getting better and better (I didn't think it'd be possible).
Love the repairs and explanations! Also I want to compliment your work at Nokia Bell Labs, I've been through some of your papers and they're amazing!
Cheers
Thank you.
Nice work Shahriar. Looking at the way the crystal pins were twisted suggests the damage was done during reassembly. Sooo close :-)
I was thinking the same. BTW, how about long time oscillator drift after the thermal shock of soldering the crystal's pins?
A nice and useful repair video. I watched every second with pleasure.
This is all interesting. I even understand (at least I think, I understand), I studied electronics at high school (I would not even think about repair without a Masters degree). But I can not imagine, I could do the repair. even if I had all year for repair. I would 100% made even more damage to the instrument. My respect for all people who can do these repairs. It's quite depressing for me. But at the same time, I am enjoying them all. Thank you for your time and effort.
EXACTLY what I was waiting for! 😍
Impressive. Good Job !
Great video. One vote for a wire bonder!
these YIG-Modules are a really common failure point in HP/Agilent/Keysight devices.
Earlier instruments didn't have the option to switch to a different reference, I think they added that because the YIGs fail so often
There was nothing wrong with the YiG. There is a faulty PIN diode and a broken 10MHz TCXO module
Keep in mind that the YIG does not replace the OCXO. In fact, the YIG always needs to be in a closed loop (using a PLL) in order to have the desired frequency.
Awesome as always. All your videos are very informative. Thanks a lot for sharing these.I learn a lot from your videos
Cool stuff!! It so nice to see the debugging process!
Though it is not my field but I appreciate your job highly good luck
Wow, how can I just say it's amazing?!
Your videos are absolutely amazing. Keep up the incredible work!
PSG!!! I wanted one so bad , but even borked ones cost like my calibrator :(
Me too. :( I'd love to have a good PSG (with vector modulation!)
I had a ribbon cable with similar manufacturer design on a weight scale. It had an internal short down to .5 ohm. I looked around with a magnifying loupe but could not find it. I didn't want to continue poking holes in it trying to trace it down so, I replaced the ribbon with wires. I have a working scale again.
just amazing
Love the new intro!
VERY VERY NICE VIDEO, VERY THANKS ... WOW
Nice job 👍
I want to see the wire bounding unit on your bench ;)
Thanks again
Another excellent video with a new error, another great learning experience.
It's a MacGyver surgery ! Very nice. Thank you.
Amazing video. Thank you for it. What are those grid-like structures coming from the ICs in 24:09 ? Is that some kind of low-parasitic bonding technique ?
Tomáš Thúróczy yes! :)
Absolute specialized laboratory equipment and the repairings of these are not for common mortal technicians.Still happy even just take a look inside them.
Nice video! Yup, a wire bonding station would be a nice addition and produce some great content. Then you need as assortment of replacement chip diodes and transistors. Wonder if there is a supplier that sells chip components in less than 100K quantities.
Neat ! Considering there was a high DC bias through much of the doubler too, how did it not get damaged ? Or can those usually handle the puny -13V bias ?
Love the new logo - Approved!
This was really interesting! Nice repair!
Cool, interesting episode, thanks!
What is the best way to tell if an RF board is ceramic or Teflon, like at 24:43? I would have thought that board was some type of Teflon base material. Thanks!
If they can get away without flexing, like they can in a rigid metal enclosure, it's going to be ceramic, because the performance is better. If you have doubts, you can bang on it and see if it goes "click" (ceramic) or if it goes "thunk" (teflon).
@@jjoonathan7178 Thanks for the info. Yeah! I was thinking you could lightly tap it with a plastic spudger and judge by sound. Thanks!
And here I was, wondering about the LAST schematic symbol you added to your logo.... the little pussy cat! I'm SURE she must add something to the logo!
amaizing! that was a great discover 👍
I remember seeing a wire bond machine on ebay not long ago, it was for parts/repair
It was interesting to see the little coils of wirebond wire in the module. Are those there for some kind of trimming? Adding those little coils in must have been pretty tricky.
If you get a wire bonder, a few words of advice on shipping. Make sure the seller properly secures all moving parts. I had bought one a few years a go and the seller did not secure the arm. The head assembly and several switches got the living daylights beat out of them. Also if it is an older west bond, like I bought, make sure they secure or remove the PCBs from their slots. I had to replace a broken 40 pin edge connector because of that.
I thought the coiled wires were inductive elements..?
Yes they are inductors designed to be attached using the same wire bonding techniques.
Oh my god what just did i watch!!! this is absolutely amazing sir i appreciate your hard work you are legend absolutely!! How did you learn these things sir?
I have to second the hp vs others service manuals. HP and successors have excellent service manuals but Rohde & Schwarz has a terrible service manual even when I got one from the vendor. I received a service manual from Rohde & Schwarz for an ESIB40 EMI Receiver and even that service manual did not have enough information to diagnose the problem.
Additionally, HP and successors user manuals are fantastic vs many other companies.
That is called a repair tutorial.. Awesome sir
My question is what is the component connected between psg and Spectrum Analyzer to connect DMM.. Can we measure dc volt directly from PSG output port on DMM
Did you pick up Electroboom's arc shenanigans in your lab or do you watch his videos anyway? 😁
Am I the only to think that curve 38:36 looks a little bit bent towards the right ?? Especially at peaks ?
Considering that the bias leakege was happening before the ALC detector, you could put an AC coupled path to the next module. Any signal loss would be compensated by the ALC. Would an AC coupled V thru solve the problem?
I did consider that option. But remember that the broken diode is constantly stressing all other diodes in that module by forward biasing them. Removing the broken diode from the circuit completely is better in the long run.
The Signal Path I would have etched my own pn junction into the pcb
@@Thesignalpath couldnt you put an SMD Diode ontop of the dead diode(after disabling it) and wirebond or even solder it into the circuit?
@@srmeister1 The diods are some very special very high frequency ones and only the chip it self which you can't buy but must be replaced with, and ordinary smd diod on top would toally whack everything way, way of its spec since among many other things cant handle the frequency. And you do that by hand with stuff that is less than 0.1 mm wide and bonding wires that are less than 0.05mm, won't happen even under a microscope. And the 32GHz impedance and everything has to be hold so just the fact that one thing is 0.1 mm higher then the others will make that change. And if you looked at the video he clearly say he don't have wirebonding machine.
at 25:15 where the trace jumps over the isolation it appears to be touching, could that cause the voltage you were seeing?
Nice new intro!
"Keysight classic design" - That's one way to spell HP
Is UNLEVEL problem be solved with same troubleshooting..if not please guide.. And where can I find the block diagrams of keysight vector signal generator
Is there a reason why you did not run self test to help diagnose the problem?
Plz I need the RF output module schematic,and datasheet for M11 1005 SMD package
I was assuming a copling capacitor had gone short circuit. Perhaps a DC block would have been a quicker temporary fix?
The DC voltage constantly forward biases other diodes in the same module. DC block would have solved the issue for the next stage, but no the same stage.
It looks like the V connectors go straight into the can and have a nice flat binding post that is wire bonded to the strip lines. Where can I find such a connector?
May I please ask you what cable you are using? I do need some thin pliable low Hz coaxial cables. :-)
Isn't Alltest the company that threatened to sue EEVblog members a few years back?
👍👍
Here is a sincere question: How do you afford all of those equipment? Honestly I can financially manage myself - amen for that - but the thought of buying an entry level VNA, or even a simple Spectrum Analyzer overwhelms me. Are these equipment simply not available for hobbyists? If so, then how do we measure stuff at GHz frequencies?
Creative thumbnail had me thinking it was a flow oven.
Triple keyboard...wooo hooo
Simon Østergaard There are 14 of them actually. ;)
Great video, great repair. Don't forget to add a PayPal Link that we can sponsor your Lab.
Regards from Germany
Thank you. I will.
considering the fact that the unit is now functional, but missing a diode and thus technically not up to spec, is there anything they can do with the instrument? Can they rent it, or will they just have to sell it as an as-is kind of thing? Or just wait until they get the correct parts anyway.
It depends on where it is used and for what purpose. If you want 100% factory performance in all conditions and corners, probably a new module is required.
how does one replace the diode? (as in how do you get it off without destroing things around it)
Those coils made of bonding wire look like they've been tuned by hand...
The bonding wires in some situations can and are used as adjustments! As an example, positioning the bonding wires in the Hp/Agilent 848x RF power sensors is an important part of calibration.
A few years ago, going upto 31,8 GHz was some of exotic....
But today with 5G phones (25 - 39 GHz in Canada, 27,5 - 28,5 GHz and 37 - 40 GHz in USA), such frequencies wil become banal...
...but good luck for coverage, believe me !!!!
3Gig isn't that bad, you could probably replace the diode.
What could have happened if the -13V would have reached your spectrum analyser?
He would have had two instruments to fix. As he says at the beginning, spectrum analyzers often can't tolerate any DC voltage at their inputs. It can destroy the input circuit.
I think it lets the magic smoke out!
I can make a generator of 50 ghz with 12v ? Thank you
Putting up to 13v dc into the attenuator surely is bad thing? But obviously it survived
There is a biasing resistor in series with the diode and 13V source. That’s what has saved everything.
Hi guys!
Help me upgrade the firmware E8251A 20ghz
Does your head ever hurt ?
Pooch says "give an oscilloscope to this poor guy"
I love corona :D That's what I was looking forward will happen on this channel :D I'm so happy :D (yes I'm, but guys take care) thank you!!! Now you are ready for my donation;)
5min later: I'm your patreon allready ... keep going , I love those videos ;)
don't you have spare components??? at least diodes jeeeez
Hi, are u persian ?
hamedtheunforgiven Yes.
@@Thesignalpath we are very proud of u man , keep learning .
@@thevoidedwarranty Thanks!
Weird thumbnail.