TSP

In this episode Shahriar attempts the repair of an Agilent PSG Vector Signal Generator. Aside from excellent RF performance, the PSG series generators are also equipped with built-in baseband generators allowing for a wide range complex RF modulations. The instrument has been loaned to The Signal Path by AllTest:
alltest.net/
This particular unit fails the internal I/Q Calibration routines. This routine calibrates for image rejection, LO rejection as well as a wide range of I/Q routing, offsets, gain/phase corrections. The calibration errors points of the I/Q MUX board and the summing nodes inside that assembly. The measurements confirms that the Q signal exhibits large distortion.
The teardown reveals the complexity of the I/Q MUX board which comprises numerous op-amps, selectors, amplifiers & multiplexers. The block diagram of the I/Q mux is presented in details along with various measurements. The fault on the I/Q MUX board is traced to one multiplexer on the Q path. The IC is replaced and the waveforms return to their linear forms. The instrument passes the internal I/Q Calibration routine after the repair. The RF output of the unit is also verified and presented showing the correct operation of the vector modulation path.
www.TheSignalPath.com
/ thesignalpath​​​
www.Patreon.com/TheSignalPath

[@lbgstzockt8493]

I must say that I simply love your videos. I have recently started my bachelors degree in electrical and computer engineering and your videos were a big contribution to studying this field. Thank you!

[@jq4t49f3]

This channel is an island of rationality in a psychotic world.

[@KeysightLabs]

Wonderful video as always!

[@kungfumaster8171]

I would have killed for video's like this when I was a young engineering student. Instead all I got was crusty bored old men that fell asleep at that desk mumbling about how great tubes where. As usual awesome video.

[@arrbam02]

Great video, we're "glad" that it was not the front panel connector ;) Thanks!

[@brianford6729]

Brilliant video - thank you for taking the time to piece it together as you figured it out! Like the block diagram you took something insanely complicated and made it accessible. Appreciate ya and your cat!

[@w2aew]

Awesome repair - isn't it great when the block diagram nearly matches the functional layout on the board?

[@artursmihelsons415]

Great repair and excellent video! 👍 Interesting decisions from Agilent side with board layout. It's always nice to see internals of expensive equipment..

[@neonkev7866]

Nice work! Always fun to watch the thought process behind honing down onto the specific problem area inside of complex units. I just finished repairing a 20GHz E8257D. Typical Frac-N problems below 3.2 GHz, the infamous Agilent 1GC1-4210 GaAs multi modulus divider MMIC. I've got 2 of these that went bad out of E8257Ds I've fixed. Would you be interested in xraying/testing them? These are famous for going bad in PSG and ESG signal generations.

[@Thesignalpath]

Sure! That would be interesting to see. How would you like to arrange that?

[@neonkev7866]

@@Thesignalpath I'll reach out to the gmail inbox on your KZbin "about" page for details.

[@xDevscom_EE]

Nice old-school analog repair and great troubleshooting procedure. It is very similar to what I usually do when fixing DC/LF metrology gear like Fluke 5790A I've repaired last week. Inject signal to input and follow how it propagates thru instrument, until determining point of failure. Now, let's step up a challenge for TSP and have a good DC only instrument repair :-) It can be harder to troubleshoot, since one needs to know what is "good" and "bad" levels are, instead of easy to see distorted AC signal ;-)

[@jianhuiluo]

That is a great work! Thank you for sharing your experience.

[@OctavMandru]

And you say you are not a patient man.... After the third one I would have started from the other direction:) Very interesting repair. Thank you

[@TheConnorGames]

I hate to be "That Guy", but it looks like a MLCC near U314 flew off at some time during your repair.

[@Thesignalpath]

Nice catch! It was a decoupling cap on one of the power supplies of that particular IC. I replaced it before I put the board back.

[@keresztesbotond740]

Came here to write this exact same comment, but you beat me to it :D

[@petrutarabuta5617]

Excellent repair. Thanks!

[@McTroyd]

The green-encircled mouse pointer got my cat's attention. Suddenly my monitor was blocked. 🤣 I know from a past interview with Dave Jones that you usually buy these broken (because working = 💰🤑💸) and hope to fix them in videos like this. What do you figure your approximate success rate is? From what you post, you seem to have a good eye for buying broken, but I'm sure there's more "behind the scenes" we don't see here.

[@paulpaulzadeh6172]

can you use your x-ray inside the the broken IC , why it get damage ?

[@akosbuzogany2752]

What I'd like to see? The faces of the iPhone engineers trying to make their phones irreparable after having watched this video.

[@electronicayexperimentos]

😁😁😁 It must be very funny.

[@lasersbee]

Like the logical hunting for the defect.... and repair🤓

[@WolfmanDude]

Very interesting video! Random video idea: Maybe you could look at some very early microwave test gear? It would be interesting to see how they made things in the GHz rage work before MMICs became mainstream. And there is alot of learning potential because not everything is a "black box" and you can actually see whats going on.

[@great__success]

I remember e.g. great Shahriar's video on Vintage (1966) HP Frequency Meter, that were machined with such accuracy it is still extremely accurature to this day

[@Meow-hw5wi]

I was using a modified IDE cable to fix my E4421B. I feel like it may be helpful if you just made some extension cables for those cards :p

[@D__x]

One of bypass caps of U314 is gone?

[@funtechu]

Never thought I'd see the day when basic chips like the AD8180 are hard to get, but here we are

[@SidneyCritic]

Not that I know what I'm talking about - lol -, but I wonder if you can use the cheap transistor tester trick of finding one blown VRM in a row of VRMs on a GPU. Identical circuits should give the same reading, but the faulty one will be different, ie, it will see open or shorted where as the others will see a semiconductor.

[@frosty129]

how can a mux die like that in the middle of that chaotic board? And nobody else dies?

[@t1d100]

Excellent

[@douro20]

I don't think I've ever seen so many op amps on a single board before.

[@sbelectronicaindustrial6652]

More than fantastic... 👍👍👏👏

[@elektrophysiklk]

👍

[@ogitakasi3030]

good stuff

[@aqib2000]

What an absolute steal for just $100