TSP #86 - Teardown & Repair of an Agilent 8164A Lightwave Measurement System

[@mikeselectricstuff]

5R resistor will be to provide a minimum load - multi-output PSUs often need this for stability

[@Thesignalpath]

Thanks Mike. Before I checked the datasheet it was not clear that it was connected across the 5V supply. I was surprised to see that they burned 2A for minimum load!

[@Hatimoooo]

Don't you think that this is a very low value for such purpose?

[@JJayzX]

that laser module is beautiful

[@OctavMandru]

honestly, it feels like a cat playing with a mouse. very calm, tactical approach, typical for a cat :) Your videos are timeless, I enjoy it today, I will definitely enjoy it in 20 years

[@OneBiOzZ]

I have never seen 90s equipment look so white and not bromideded (totally a word) it is actually quite comforting! loved the video as i do all your videos (especially because i specialize in power supplies)

[@ThePapa41]

Shahriar, you did an outstanding job diagnosing the problem with this unit, without a schematic yet !!!! Love your video's. Lou S

[@mikeselectricstuff]

That diode was measuring 0R as it would be across the transformer secondary

[@Thesignalpath]

Yes, I traced it to the transformer somewhere in the video.

[@dentakuweb]

I think he figured that out at around 43:00

[@yaghiyahbrenner8902]

Doctor Mike.

[@paulmoir4452]

That was the one time I felt I had one over on Shahriar. Didn't last long dammit. :)

[@Hatimoooo]

I don't know why, but I always feel happy when you repair something : )

[@campbellmorrison8540]

Wow you are brave fixing a switching supply on live video :) Yes I look forward to seeing the internals of that laser module. Great video as always thanks

[@CreativeCheats]

Definitly interested in a video about the laser module. Great video by the way.

[@AlexTaradov]

Yes, do the optics part.

[@rafaellarios3707]

It always feels great to get something back up and running again, doesn't it?

[@JackZimmermann]

Great stuff! I found myself shouting "Yay!" when it booted. My wife thinks I'm crazy. Your earlier videos inspired me to actually buy broken equipment on eBay, so thanks to you I now have a Tektronix analog oscilloscope (it goes down to 500 µV so great for noise measurement) that I bought for next to nothing and repaired. Thanks man! Good to see some new videos.

[@machiasaudiowerks]

I for one, would love to see more demonstration and calibration of the LightWave multimeter! Kind of a funny coincidence, I just PM'd you regarding my LightWave multimeter issues, yesterday!

[@zaprodk]

It's good to see a new video, and in the new lab :D Usually there is never a full bridge rectifier on SMPS secondaries. One double diode (higher power) is strung between both ends of a winding with center tap. Another lower power, it's perfectly normal to parallel both diodes in the case to lower dissipation. The board in the lid is visually 'active PFC', so without it, nothing works. The sound of a "pumping" SMPS already tells me that PFC and primary are ok, and you should look on secondary for the short. The diodes are on two different secondaries, so no need for "matching". I am 100% sure that the optical power module broke down, overloaded one of your 15V supplies and killed the diode. A classic situation. A shorted diode would shut down the SMPS immediately. - Keep up the good work, and please do a video on the laser-module!

[@NSKChaitanya]

I was waiting to see who would realise that the short that he measured was an intentional paralleling of the anodes of the MURs :-D

[@todayonthebench]

Yes, the signs were there and it isn't a too complicated beast per say. It is just a lot of smaller parts linked one after the other making the final product. The classical ticking sound it made were a good sign that high side were okay. But I can also understand that if one isn't used to fiddling with SMPS systems that one will mostly scratch one's head wondering what all the stuff is even doing! Not that the active power factor correction board is helpful in its design, since it appears to have a small high side regulated power supply on it as well.

[@MrCarlsonsLab]

Great troubleshooting! Just so you know, and you may not know this unless you are working around this a lot..... Those "PF" series capacitors are notorious for physically leaking around the leads. They puddle on the board and erode traces. If they are not leaking yet...... *It's just a matter of time.* For some funny reason, power supply manufacturers love to put FB traces right between the legs of these caps. If you plan on keeping this unit for awhile, I strongly suggest you replace them.

[@Thesignalpath]

Thank you Mr. Carlson. As soon as I get a chance, I will replace them.

[@paulz4394]

Ххх

[@dentakuweb]

It's great to see you back making videos.

[@jimadams2473]

Okay, this was one of your best tear-down/repair videos I have seen. I thought maybe it had you stumped for a minute but you saved the day by salvaging that old part. First time I heard you cuss too, which was hilarious!!! Anyway, another EXCELLENT video. Keep up the GREAT work...

[@CKOD]

Keep it up with the tearing down and and fixing equipment, always intersting to watch. I was feeling inspired and picked up a broken 3000W inverter that appeared to be good build quality. (Not a giant mess inside like some) bit of poking about even without the schematic, figuring out the architecture (dual push pull converters to get the + and - HV rails and an H bridge for the output) 2 output mosfets gone short and 2 burnt out resistors on the gate driver card replaced and it was back and running like new. An output inductor and common mode choke and X caps to tame any transients in the future, and it was better than new.

[@mikeselectricstuff]

If it has power factor correction, that DC cap could easily be 340VDC, which WILL hurt!

[@Thesignalpath]

The cap was on the rectifier output. Charged to 170VDC roughly.

[@SuperZylar]

mike, the machine is an yank, sqrrt2*115=163V DC

[@mrlazda]

+SuperZylar This power supply use boost PFC (uc3854) so capacitor should be on output of PFC and voltage on it should be higher then peak value of maximum input voltage (in US that is 180V for 127V RMS) which is theoretical minimum below which this type of PFC will not work but usually that voltage is much higher. For universal input PSU (and most US power supply I saw) that voltage is usually near 400V (>380V us3854 suggest 385V for universal input) and do not change with input voltage. If voltage on capacitor is less then 180V that could be indication that PFC is not working or it is really bad design (and this PSU do not look that is designed by people who have no clue about PSU design).

[@stefantrethan]

Looks to me like the PFC was on the top PCB (the one with the large toroid choke). Would make sense since the plug connector was splicing it in right between the AC input filter and bulk capacitor. On power supplies of that vintage the PFC is often a separate module that looks a bit bodged in. Possibly because they continued to recycle existing designs for the power supply. Modern ones often have the PFC more tightly integrated. It is a 400V rated bulk cap, and universal input, so my money is on Mike for this one (>350VDC at the cap).

[@mrlazda]

stefantrethan Top PCB is not doubt PFC . If it is universal input than DC voltage on capacitor is >380V (due permitted tolerances maximum input RMS voltage is 267V)

[@todayonthebench]

The top board in the power supply is most likely doing power factor correction. This would be placed after the mains rectifier and input filtering, but the bulk capacitance would be placed after the power factor correction unit. Thereby it is no surprise that the board to boar interconnect is placed between the input filter and the bulk capacitance. The large inductor on the top board is what makes this logical. The top board also seems to have a small power supply on it, that seems to be having high side regulation, I would guess that this second power supply is used for standby power for the switch on/of functionality. This power supply can be seen in the lower right section at 20:40 And a small note. Switch mode power supplies secondary side should not have a full bridge rectifier. Since the principal of operation of an SPMS is from a practical standpoint just a coupled inductor. And yes, this means that we would expect a lot of diodes connected in parallel on the secondary side. Otherwise a good video.

[@davidgustafik7968]

Thanks for the video! I'd love to see the innards of the free space optics!

[@yaghiyahbrenner8902]

Doctor Shahriar. request for a second video. good video.

[@MustafaHammood]

Awesome! would definitely be interested in a video showing the TLS free space optics :)

[@AndreaThing]

I would really love to see a follow up of this video, with the repair of the main laser source! Just to let you know :)

[@SatyajitRoy2048]

It couldn't be more interesting than to to see a video with 0 dislike after 2K views. Double thumbs up for all of your good work.

[@catt87]

I would definitely like to see a repair attempt of that laser unit. That would be very enjoyable for me

[@cj-de2eb]

Another great video, thank you! I kept saying "maybe the diodes are paralleled!" but you didn't seem to hear me :-). I definitely would be interested to see the explanation and repair of the big laser module. There are plenty of power supply teardowns and repairs out there but you're the only one I know of that takes the time and has the knowledge to go through the more advanced stuff. Also look forward to seeing your lab tour. You do such a good job on things that I'm sure it will be inspirational for mine :-). I imagine finishing it out will include some soft wall treatments to improve audio a bit......

[@gkasprow]

the upper PCB seem to be an active power factor correction module. The big coil, 2 MSOFETs and the diode creates boost converter

[@widecast]

That type of rectifier is common in SMPS, and of course you were measuring across the transformer. I haven’t seen a diode bridge used on the secondary supplies of a switched mode power supply. They are usually a full wave rectifier with a centre taped winding and the 2 diode package.

[@MTTT-bl2uo]

The top board with the large inductor on it contains the active power factor correction circuitry. Other than the large inductor, the other giveaway is the UC3854 IC which is a power factor pre-regulator controller. As for the burned up power sensor module, if I had to guess I'd say it probably suffered a classic case of tantalum capacitor failure. One possible failure mode of tantalum capacitors is a short circuit, and if the power supply is able to deliver sufficient current into the failed capacitor, it can get so hot the capacitor burns up and emits quite a lot of soot (also smells very bad). Before writing off the module I'd clean off the soot and see if there wasn't too much damage done to the traces. If it was indeed a tantalum capacitor failure, there isn't going to be much left of it other than the solder pads it once sat on.

[@b.6603]

Thanks so much for the great video! Please, do the laser module video! This video also made me realize I should learn about how switching power supplies work but I suppose someone has already made a video on that and there are lots of great books. So a theory video on power supplies would be awesome but maybe there is no point in asking?

[@ciprianwinerElectronicManiac]

When I first heard that ticking I said to myself: shorted output diode(s)! Glad that you brought it to life :)

[@babybearkill1]

here's a little trick for fixing smps. get the data sheet on the smps controller and power the high side controller up with a single rail lab PSU, then poke around and make sure the gate drive is doing what its supposed to be doing. measure dead time and fall time. gate hysteresis diode failure shows up with this test. once you are happy items are behaving themselves throw around 50 or 60v on the main bridge rectifier and measure all other functions associated with the high voltage side. one of the killers of IGBT choppers is lazy rise and fall times causing both gates to be turned on momentarily. I do this every day and it saves hours of measuring individual components to find a failure.. current limit drive circuits to around 100mA and work up from there. works well with SG352X, TL494, international Rectumfinder parts and so on

[@gkasprow]

two cathodes of the diode are shorted by the transformer secondary winding. It is half bridge converter with central tap secondary rectifier such damage happens quite often.

[@AF6LJSue]

Please do show the walk through of the tuneable laser module, and repair if possible. Good Stuff.

[@nixxonnor]

Always Nice to see how you troubleshoot :D

[@kirkpennock2997]

I vote for a video on the huge module fix.

[@Juggle4Evr]

Very entertaining as always!

[@FamilienSoelberg]

Always love your videos!!!

[@stephenemanouil6745]

I enjoy watching most of your videos and usually learn something new, and this one was terrific. Have you repaired the large, tunable LASER plug-in at the bottom of the unit yet? I searched for a repair or tear down of the tunable LASER and did not find one. Thank You.

[@davidmuriungi4122]

Please do a video on the fiber optic module !!!

[@1hdsquad]

Would it have been possible to measure all output rails of the closed PSU with an oscilloscope simultaneously to see if one rail couldnt start to narrow down the problem? Id love to see the second video!

[@cccyberfamilydk]

Thanks. Can't wait for the next video. :)

[@colt4547]

Great Video. Thank you for sharing.

[@gfrias06]

Awesome video as always, thank you!

[@kissingfrogs]

Another great video. Very enjoyable. Thanks.

[@henkvandeleur6872]

Thanks for this video. Pity the burned module is beyond saving. Perhaps there is a loose available? Greetings from the Netherlands. Looking forward to next time. '73

[@mogwopjr]

Good lord that tunable laser module is huge. This is a great example of how communication equipment is getting smaller every year. Our tunable laser modules are XFP's. Though one XFP carries about the same price tag as the actual test meter we use it in. :/

[@MichaelLloyd]

Great video. Very informative!

[@shaxjl]

Thank you very much!

[@Nermash]

My algorhytm: IF supply ticks AND primary fuse ok THEN check secondary rectifiers :)

[@jaa93997]

☺ a simple cheap rectifier trolling a multi thousand dollar piece of fine instrumentation, poetic !

[@andreykorenev1433]

Thank you for the video! You are really lucky guy. When I see something broken on Ebay I always imagine something unrepairable inside. And please make a video about this optical stuff (you may add some theory as well).

[@FriedrichWinkler]

love the pooch cameos

[@vaualbus]

I would try to repair the burn module at least clean it up to see the real demage. It would an interesting process I think.

[@zaprodk]

Repair is not viable - unless your time is free. Burned PCB-material is highly conductive, so you would have to mill out a lot of the board with a dremel and then reconstruct the missing piece. Not worth the time imho.

[@javierpallalorden]

Nice!

[@FennecTECH]

I will be witing for part two :D

[@mikeissweet]

Bravo!

[@PsRohrbaugh]

If you haven't done so, PLEASE back up the proprietary software on that mechanical HDD. There are usb hard drive adapters for a few dollars, and you can use Linux or speciality programs to simply create an image of the entire disk - boot sector, partitions, etc. That way if the drive dies, your system doesn't become useless.

[@TheAlfieobanz]

Great vid. You got a sub. Love the cat.

[@ZaphodHarkonnen]

FDD: Floppy Disk Drive, HDD: Hard Disk Drive. All those messages were basically about the computer booting up.

[@phuang3]

Perhaps it just a faulty tantalum capacitor on the module. You can fix it.

[@BersekViking]

Show us the laser voodoo! :)

[@PsRohrbaugh]

Also, I watch on my cell phone - but the connector from the power supply looks a LOT like a standard ATX connector. If it is, check the pinout. You might be able to get spare power supplies extremely cheap.

[@AureliusR]

I'd say put up some acoustic dampening panels in the lab. Way too echo-y.

[@widecast]

A power supply shutting on and off like that is said to be hiccuping. One comes across this type of effect in solid state TV sets quite a lot.

[@gglovato]

"I'm not a patient man".... Proceeds to put out a 60min video hahahahaha

[@vaualbus]

Still plain to repair the bottom module??

[@SuperZylar]

whohoe, my first patreon pledge movie! for who didn't know : www.Patreon.com/TheSignalPath

[@daveblane6442]

WOW!

[@dtiydr]

56:08 Holy shit that cost some money..

[@NourMuhammad]

It is usually an over-current protection when you hear this click in the power supply and most likely the output stage has the problem around the rectifier circuit!

[@pufero1]

Try to start it whit out the hd i fear these old PATA 2,5" hd the can dead at 60.000 hours of use or less soo a clone backup is a must. You forget the on board anodes of diodes on the pcb shorted on the scondary of transformer one coil comon/ground and another coil. the laser module was the fpga DC to DC converters what died may be a solid cap shorted and killed they, theses old orange used for do that.

[@jimgeorge8956]

Agreed, PATA HDDs can be replaced with CompactFlash cards for low-demand applications. Also, while the burn marks on the laser module are impressive, I'd guess that it was likely an over-volted tantalum capacitor. They put on quite a show before they burn out. If the traces are still ok, the board may be salvageable.

[@FennecTECH]

Lol the cat

[@dtiydr]

31:55 xD

[@DavidMG99]

👍👍

[@wither8]

FDD passed = floppy disk drive

[@NotSleepy]

lasers!!!

[@chevsev]

if you would be interested in selling the exploded unit for parts send me a msg so we can talk.

[@teemoinvietnam1356]

this is my fetish...

[@geekchenxiang]

wow. I am the first one again.. Thanks again for your educational video. Great job!